Semiconductor Application Worldwide Dataquest

Semiconductor Application Worldwide Dataquest

Dataquest

Semiconductor Application

Markets Worldwide

Dataquest Incorporated

1290 Rldder Park Drive

San Jose, CA 95131-2398

(408) 437-8000

Telex: 171973

Fax: (408) 437-0292

United Kingdom

Dataquest UK Limited

Roussel House,

Broadwater Park

Denham, Nr Uxbridge,

Middx UB9 5HP

England

0895-835050

Telex: 266195

Fax: 0895 835260-1-2

Japan

Dataquest Japan Limited

Shinkawa Sankd Building 2 Fl

1-3-17 Shinkawa

Chuo-kuTokyo 104

Japan

011-81-3-5566-0411

Telex: 781-32768

Fax: 011-81-3-5566-0425

France

Dataquest Europe SA

Tour Galli6ni 2

36, avenue du G6n6ral-de-Gaulle

93175 Bagnolet Cedex

France

(1)48 97 31 00

Telex: 233 263

Fax: (01)48 97 34 00

Korea

Dataquest Korea

Dacheung Building Room 1105

648-23 Yoiksam-dong

Kangnam-gu, Seotil 135-80

Korea

011-82-2-552-2332

Fax: 011-82-2-552-2661

Germany

Dataquest GmbH

Kronstadter Stiasse 9

8000 Munich 80

West Germany

Oil 49 89 93 09 09 0

Fax: Oil 49 89 930 3277

Dataquest Incorporated

Lec^eway/Dataquest

The Corpwrate Center

550 Cochituate Road

Fiamingham, MA 01701

(508) 370-5555

Fax: (508) 370-6262

The content of this rejxjrt represents our interpretation and analysis of Infonnatlon generally available to the public or released by responsible individuals in the subject companies, but is not guaranteed as to accuracy or completeness. It does not contain material provided to us In confidence by our clients.

This information is not furnished in connection with a sale or offer to sell securities, or in connection with the solicitation of an offer to buy securities. This firm and Its parent and/or their officers, stockholders, or members of the families may, firom time to time, have long or short position

In the securities mentioned and may sell or buy such securities.

Printed in the United States of America. All rights reserved. No part of this publication may be reproduced, stored in retrieval systems, or transmitted, in any form or by any means—^mechanical, electronic, photocopying, duplicating, microfilming, videotape, or otherwise—without the prior permission of the publisher.

01991 Dataquest Incorporated

Welcome to Dataquest

Semiconductor Application Markets Worldwide

You are in the

Dataquest Perspective

binder

A series of multitopic publications that provide analysis on worldwide semiconductor application markets trends and issues and semiconductor news and views are contained in this binder.

Other Semiconductor Application Markets Worldwide service binders:

Source: Dataquest

An annually updated collection of reference documents for the Semiconductor Application Markets

Worldwide service. Worldwide and North American market statistics; Company Backgrounders; and

several guides such as How to Use Dataquest, Dataquest Research Methodology, and Dataquest

High-Technology GuideSegmentation and Glossary.

DataQuest

v w n acompanyoF

JUJD ThcDun&BradsticctCorpoiation

Dataquest

Perspective

Semiconductor

Application Markets

Worldwide

Index

October-December 1991 January 31, 1992

How to Use This Index

This is a cumulative index of key industry terms, companies, and products for all 1991 issues of Dataquest

Perspective. Entries are followed by the date of publication and the page numbeiCs). Produa names are listed under the company that manufactures/publishes the produa. General information about a company itself is foimd under the full company name. Each citation indicates only the beginning page of a discussion of a topic (the range of page numbers is not cited). A Table of Contents for all 1991 issues of Dataquest Perspective—listhig each issue number, date, and article title—is included at the end of the index! lOBASB-T LAN market, (Dec l6):3

ABS (antilock braking system), (Dec 16):7

Advanced Mobile Traffic Information and Communication System (AMTICS), (Dec l6):8

Aerospace applications. See Military/aerospace applications

Air bags for vehicles, (Dec l6):7

Airbus commercial airliner deliveries (1981-1995),

(Pa 7): 14

Airliner deliveries commerical (1981-1995), ( O a 7):14

AMTICS (Advanced Mobile Traffic Information and

Communication System), (Dec 16):8

Analog devices semiconduaor consumption for military/aerospace applications (1990-1995), ( O a 7):15

Antilock braking system (ABS), (Dec l6):7

Application-specific standard products (ASSPs) computer digital video opportunities for,

(Nov 18):5

T-lA"-3 mixed-signal chip sets, (Nov 18):13

Asia/Pacific-Rest of World OlO"W) automotive semiconduaor market in (1990-1995),

(Dec 16):8

LAN/PDDI chip set market in (1990-1995),

(Dec 1©:2

ASICs gate arrays, computer digital video market for

(1991-1995), (Nov 18):8

PLDs, computer digital video market for

(1991-1995), (Nov 18):8

ASICs (continued)

T-carrier cores, (Nov 18):14

ASSPs. See Application-specific standard products

(ASSPs)

AT&T (American Telephone and Telegraph Co.) communications applications semiconduaor revenue worldwide (1990), (Dec 1(S):10

Automobiles electric, (Dec l6):6

Automotive applications. See Transportation applications

Automotive electronics market analysis of, (Dec 16):5

Aviation and space electronics market civilian, ( O a 7):12

B

Batteries for electric cars, (Dec l6):6

Boeing commercial airliner deliveries (1981-1995),

( O a 7):14

California Air Resources Board (CARB) emission requirements for vehicles, (Dec l6):6

CAN protocol, (Dec 16):8

CARB. See California Air Resources Board (CARB)

CDR (constant density recording) for disk drives,

(Nov 18):12

©1992 Dataquest Incoipoiaied / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012735

Semiconductor Application Markets Worldwide

Cellular telephones equipment market, U.S. (1990-1995), ( O a 7):9

Pan-European GSM standard for, (Oct 7):10 chip sets

FDDI-II, (Dec l6):5

FDDI/CDDI, trends for, (Dec l6):4

LAN, trends for, (Dec 1 0 : 3 lAN/FDDI, market analysis of, (Dec l6):2

PCN, (Pa 7):10

T-l/T-3 application-specific standard p r o d u a

Qvilian aviation and space electronics market,

( O a 7):12

Commercial airliner deliveries (1981-1995), (Oct 7):l4

Communications applications control electronics market for (1991-1995),

( O a 7):8 semiconduaor suppliers for, (Dec l6):10

Compression. See Data compression

Computer digital video hardware vendors for, (Nov 18):7 market analysis of, (Nov 18):2P system block diagrams, (Nov 18):3, 6

Conferences and exhibitions

Telecommunications Industry Conference, ( O a 7):9

Constant density recording (CDR) for disk drives,

(Nov 18):12

Consumer applications control electronics market for (1991-1995),

( O a 7):8

Control applications

32-bit, candidates for, (Dec l 6 ) : l l market analysis of, ( O a 7):6

Controllers rigid disk drive (RDD), (Nov 18): 10

Control system block diagram, ( O a 7):7

DRAM (continued) memory cards using, ( O a 7):2

DVI (digital video interactive), (Nov 18):2

EEPROM

Electric vehicles, (Dec 16):6

Electronics equipment production military/aerospace applications worldwide

(1990-1995), ( O a 7):13 -^

Embedded control worldwide market (1991-1995), ( O a 7):8

Emission controls for vehicles, (Dec l6):6

EPROM computer digital video market for (1991-1995),

(Nov 18):8 memory cards using, ( O a 7):2

Ethernet chip sets, worldwide market for (1990-1995),

(Dec 16):2

FDDI vs., (Dec 16):3

Europe automotive electronics market phase in, (Dec l6):5 automotive semiconduaor market in (1990-1995),

(Dec 16):8

LAN/FDDI chip set market in (1990-1995),

(Dec 16):2

PCNs in, ( O a 7): 10 semiconduaor products market for military/ aerospace applications (1991-1995), ( O a 7):l6 eXecute-in-Place (XIP), ( O a 7):5

Exhibitions. See Conferences and exhibitions

D

Data communications semiconduaors for, ( O a 7):10

Data compression

ICs for computer digital video market (1991-1995),

(Nov 18):8 electronic photography market, ( O a 7):5 open standards for, (Nov 18):2

Defense electronics opportunities, ( O a 7): 13

Digital video interactive (DVI), (Nov 18):2

Digitizer ICs computer digital video market for (1991-1995),

(Nov 18):8

Disk arrays, (Nov 18): 11

Disk drives. See Rigid disk drives (RDDs) computer digital video market for (1991-1995),

(Nov 18):8

Fairs. See Conferences and exhibitions

FDDI. See Fiber-distributed data interface (FDDI)

FDDI-II chip sets, (Dec 1©:5

FDDI/CDDI chip sets trends for, (Dec l6):4 worldwide market for (1990-1995), (Dec 1©:2

Fiber-distributed data interfece (FDDI) chip sets for. See LAN/FDDI chip sets

Ethernet vs., (Dec 1©:3

Flash memory computer digital video market for (1991-1995),

(Nov 18):8 as lowest cost storage, ( O a 7):5 memory cards using, ( O a 7):2

Fujitsu company. See Fujitsu Ltd. memory card offerings of, ( O a 7):5 communications applications semiconduaor revenue worldwide (1990), (Dec 16):10

©1992 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose,

CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012735

Semiconductor Application Markets Worldwide

Gate anays. See under ASICs

Global Positioning System (GPS), (Dec 1©:8

GM

I m p a a test vehicle, (Dec 16):7

GPS (Global Positioning System), (Dec l6):8

GSM cellular telephone standard, ( O a 7): 10

J1850 protocol, CDec 1©:8

Japan

Advanced Mobile Traffic Information and Communication System (AlVfllCS), (Dec 16):8 automotive electronics market phase in, (Dec 10:5 automotive semiconductor market in (1990-1995),

(Dec 16):8

LAN/FDDI chip set market in (1990-1995),

(Dec 16):2

Joint Photographic Experts Group (JPEG)

JPEG compression standard, (Nov 18):2

JPEG. See Joint Photographic Experts Group (JPEG)

H

H.261 (Px64) compression standard, (Nov 18):2

Hand-held computers memory card market for (1991-1995), ( O a 7):4

Hard disks. See Rigid disk drives (RDDs)

Highways intelligent, (Dec 16):8

Hitachi Ltd. communications applications semiconduaor revenue worldwide (1990), (Dec l6):10 market share, automotive and truck semiconduaors, (Dec l6):9

LAN chip sets trends for, (Dec l6):3

LAN/FDDI chip sets market analysis of, (Dec l6):2

LANs wireless, semiconduaors for, ( O a 7):10

Laptops, notebooks, and portables memory card market for, ( O a 7):4

Logic semiconduaor consumption for military/aerospace applications (1990-1995), ( O a 7):15

IBM company. See IBM Corp.

Token-Ring products for UTP media, (Dec 10:3

IBM Corp.

Media Control Interface (MCD specification,

(Nov 18):2

Integrated circuits (ICs) compression computer digital video market (1991-1995),

(Nov 18):8 electronic photography market, ( O a 7):5 digitizer, computer digital video market

(1991-1995), (Nov 18):8 rigid disk drive (RDD) controller, (Nov 18):10 semiconduaor consumption for military/aerospace applications (1990-1995), ( O a 7):15

Intel

28F001BX 1Mb flash memory, ( O a 7):5

8051 microcontroller, (Nov 18):12

ActionMedia II board, (Nov 18):4

DVI standard, (Nov 18):4

Exchangeable Card Architecture (ExCA), ( O a 7):3 memory card offerings of, ( O a 7): 5

Intelligent control woridwide market (1991-1995), ( O a 7):8

Intelligent highways, (Dec 1 0 : 8

Intelligent Vehicle Highway System (IVHS), (Dec 1 0 : 8 rVHS anielligent Vehicle Highway System), (Dec 10:8

M

Market analysis automotive electronics, (Dec 10:5 computer digital video, (Nov 18):2 control applications, ( O a 7):6

LAN/FDDI chip sets, (Dec 10:2 memory cards, ( O a 7):2 military/aerospace semiconduaor opportunities,

( O a 7):12 rigid disk drives (RDDs), (Nov 18):9

T-carrier equipment, (Nov 18):13 telecommunications semiconduaor opportunities,

( O a 7):9

Mass storage applications control electronics market for (1991-1995),

( O a 7):8

McDonnell Douglas commercial airliner deliveries (1981-1995),

( O a 7):14

Memory cards defined, ( O a 7):2 market analysis of, ( O a 7):2

Memory products semiconduaor consumption for military/aerospace applications (1990-1995), ( O a 7):15

©1992 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012735

Semiconductor Application Markets Worldwide

Microcomponents semiconduaor consumption for military/aerospace applications (1990-1995), ( O a 7):15

Microsoft Corp.

Media Control Interface (MQ) specification,

(Nov 18):2

Military/aerospace applications electronics equipment production worldwide

(1990-1995), ( O a 7):13 market analysis of, semiconductor opportunities,

( O a 7):12 semiconduaor consumption for (1990-1995),

( O a 7):15

Mirroring for disk drives, (Nov 18):11 . . _

Mitsubishi memory card offerings of, ( O a 7):5

Motion Pictures Experts Group (MPEG)

MPEGI compression standard, (Nov 18):2

MPEGn compression standard, (Nov 18):2

MPEGin compression standard, (Nov 18):5

Motorola

68HC11 microcontroller, (Nov 18):12 company. See Motorola Inc.

Motorola Inc. communications applications semiconduaor revenue worldwide (1990), (Dec l6):10 market share, automotive and truck semiconduaors, (Dec 16):9

MPEG. See Motion Pictures Experts Group (MPEG)

Multimedia systems computer digital video for, (Nov 18):2

Multiplexers automotive, (Dec l6):8

SONET-based fiber-optic, (Nov 18):13

T-1, (Nov 18):13

T-2, (Nov 18):13

T-3, (Nov 18):13

N

National Semiconduaor company. See National Semiconduaor Corp.

HPC-Plus microcontroller, (Nov 18):12

National Semiconduaor Corp. communications applications semiconduaor revenue worldwide (1990), (Dec 1©:10 market share, automotive and truck semiconduaors, (Dec l6):9

Navigation on highways, (Dec l6):8

NEC Corp. communications applications semiconduaor revenue worldwide (1990), (Dec l6):10 market share, automotive and truck semiconduaors, (Dec 16):9

NexCom Technology Inc. solid-state disk (SSD) technology at, ( O a 7):6

Nissan

FEV technology, (Dec l6):6

North America automotive electronics market phase in, (Dec l6):5

North America (continued) automotive semiconduaor market in (1990-1995),

LAN/PDDI chip set market in (1990-1995), point-of-sale CPOS) system revenue (1990-1995), semiconduaor products market for military/

T-1 standard in, (Nov 18):13

T-carrier semiconduaor market in (1987-1995),

See also United States

Notebooks. See Laptops, notebooks, and portables

O

(Dec 16):8

(Dec 16):2

(Dec 16): 11 aerospace applications (1991-1995), ( O a 7):l6

(Nov 18): 14

Oki Electric Industries Co. Ltd. market share, automotive and truck semiconduaors, (Dec l6):9

Optoelectronic devices semiconduaor consumption for military/aerospace applications (1990-1995), ( O a 7):15

OTP ROM memory cards using, ( O a 7):2

Pacific. See Asia/Pacific-Rest of "World (ROW)

Palmtop computers. See Hand-held computers

Pan-European GSM cellular telephone standard,

( O a 7): 10

PBX applications wireless, semiconduaors for, ( O a 7):10

PCMCIA. See Personal Computer Memory Card

Industry Association (PCMCIA)

PCNs. See Personal communications networks (PCNs)

PCs. See Personal computers (PCs)

Pen-based computers memory card market for (1991-1995), ( O a 7):4

Peripheral applications control electronics market for (1991-1995),

( O a 7):8

Personal communications devices functional comparison of, ( O a 7):9

Personal communications networks (PCNs) chip sets for, ( O a 7):10 in Europe, ( O a 7):10 semiconduaors for, ( O a 7):9

U.S., ( O a 7):9

Personal Computer Memory Card Industry Association

(PCMCIA) memory card standard

Revision 1.0, ( O a 7):3

Revision 2.0, ( O a 7):3

Personal computers (PCs) memory card market for (1991-1995), ( O a 7):4

©1992 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012735

Semiconductor Application Markets Worldwide

Philips communications applications semiconductor revenue worldwide (1990), (Dec 16):10 market share, automotive and truck semiconductors, (Dec l6):9

PLDs (programmable logic devices). See under ASICs

Point-of-sale (PCS) system revenue in North America (1990-1995), (Dec l 6 ) : H

Portables. See Laptops, notebooks, and portables

POS (point-of-sale) system revenue in North America (1990-1995), (Dec l 6 ) : l l

Programmable logic devices (PLDs). See under ASICs

Prometheus vehicle technology effort, (Dec 16):8

Protocols

CAN, (Dec 16):8

J1850, (Dec 1©:8

T-1, (Nov 18):13

Px64 (H.26I) compression standard, (Nov 18):2

Smart transmissions for vehicles, (Dec l6):7

Solid-state disk (SSD) replacement, ( O a 7):6

SONET-based fiber-optic multiplesers, (Nov 18):13

SRAM computer digital video market for (1991-1995),

(Nov 18):8 memory cards using, ( O a 7):2

SSD (solid-state disk) replacement, ( O a 7):6

Standards data compression, open, (Nov 18): 2

Pan-European GSM cellular telephone standard,

( O a 7):10

T-1, in North America, (Nov 18):13

SunDisk

2.5/5/lOMb SSD plug-and-play subsystems,

( O a 7):6 company. See SunDisk Inc.

SunDisk Inc. solid-state disk (SSD) technology at, ( O a 7):6

SynOptics

Token-Ring products for UTP media, (Dec l6):3

R

RDDs. See Rigid disk drives (RDDs)

Rest of World (ROW). See Asia/Pacific-Rest of World

(ROW)

Rigid disk drives (RDDs) market analysis of, (Nov 18):9 memory cards as replacements for, ( O a 7):3

ROM memory cards using, ( O a 7): 2

ROW (Rest of World). See Asia/Pacific-Rest of World

(ROW)

Semiconductors computer digital video opportunities for,

(Nov 18):5 consumption of automotive market worldwide (1990-1995),

(Dec 16):8 military/aerospace applications (1990-1995),

( O a 7):15 data communications use of, ( O a 7):10 military/aerospace applications for, ( O a 7): 12

PCN use of, ( O a 7):9 rigid disk drive (RDD) content of, (Nov 18):10

T-carrier equipment content of, (Nov 18):13 telecommunications opportunities for, ( O a 7):9 wireless PBX use of, ( O a 7):10

SGS-Thomson Microelectronics B.V. communications applications semiconduaor revenue worldwide (1990), (Dec l6):10 market share, automotive and truck semiconduaors, (Dec l6):9

Shows. See Conferences and exhibitions

Siemens AG market share, automotive and truck semiconduaors, (Dec l6):9

T-1 multiplexers, (Nov 18):13

T-1 protocol, (Nov 18):13

T-1 standard in North America, (Nov 18):13

T-lA'-3 application-specific standard p r o d u a (ASSP) chip sets, (Nov 18):13

T-2 multiplexers, (Nov 18):13

T-3 multiplexers, (Nov 18):13

Taiwan automotive semiconduaor market in (1990-1995),

(Dec 16):8

T-carrier equipment market analysis of, (Nov 18):13

Telecommunications applications semiconduaor opportunites in, ( O a 7):9

Telecommunications Industry Conference, ( O a 7):9

Texas Instruments Inc. communications applications semiconduaor revenue worldwide (1990), (Dec 10:10 market share, automotive and truck semiconduaors, (Dec 16):9 products. See TI

n . _ _ _

company. See Texas Instruments Inc. memory card offerings of, ( O a 7):5

TMS320 DSP, (Nov 18):12

Token-Ring chip sets, worldwide market for (1990-1995),

(Dec 16):2

Toshiba

4Mb flash solid-state disk (SSD), ( O a 7):6 company. See Toshiba Corp.

Toshiba Corp. communications applications semiconduaor revenue worldwide (1990), (Dec l6):10 market share, automotive and truck semiconduaors, (Dec 16):9

©1992 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012735

Semiconductor Application Markets Worldwide

Trade shows. See Ckjiiferences and exhibitions

Transportation applications control electronics market for (1991-1995),

(Qa 7):8

See also Automotive electronics

U

United States cellular telephone equipment market (1990-1995),

( O a 7):9

Intelligent Vehicle Highway System (IVHS),

(Dec 1©:8

POfs in, ( O a 7):9 wireless LAN/WAN equipment markets (1991-1995),

(Oct 7):10 wireless PBX equipment shipments (1991-1995),

(Oct 7):10

See also North America

Unshielded twisted pair (UTP) media, (Dec l6):3

U.S. Advanced Battery Consortium, (Dec 16):6

UTP (unshielded twisted pair) media, (Dec l6):3

ZDR (zoned density recording) for disk drives,

(Nov 18):12

Zoned density recording (ZDR) for disk drives,

(Nov 18):12

Vehicle electronics. See Automotive electronics

Vehicles electric, (Dec l6):6

A^deo editing systems, (Nov 18):3

VRAM computer digital video market for (1991-1995),

(Nov 18):8 w

WANs senaiconduaors for, ( O a 7): 10

Wireless LAN/WAN equipment markets

U.S. (1991-1995), ( O a 7):10

Wireless PBX equipment shipments, U.S. (1991-1995), (Oa 7):10 semiconduaors for, ( O a 7):10

XIP (eXecute-in-Place), ( O a 7):5

©1992 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012735

Semiconductor Application Markets Worldwide

Dataquest Perspective issues covered in this index:

VoL 1, No. 1: October 7, 1991

Memory Cards: An Emerging and Potentially

Explosive Market, 2

Control ^ p l i c a t i o n s : The Big Part of the

Iceberg, 6

Semiconductor Opportunities in

Tdecommunications, 9

Mil/Aero Outlook: Positioning for New Needs, 12

VoL 1, No. 2: November 18, 1991

Computer Digital \^deo: A Multimedia

Opportunity, 2

Rigid Disk Drives: A Case for Integration, 9

T-Carrier Market Offers Mixed-Signal Semiconductor

Opportunities, 13

VoL 1, No. 3: December l 6 , 1991

LAN/FDDI Applications: Excellent <3iip Set

Opportunity, 2

Automotive Applications: More Controls Offset

•Vicious Economics, 5

Semiconductor Application Markets Inquiry

Highlights, 10

©1992 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA. 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012735

Dataqyest

n m a company of

MMM» The Dun &Bradslrccl Corporation

Dataquest

Perspective

Semiconductor

Application lUiarlcets

Worldwide

Vol. 1, No. 1 October 7, 1991

M a r k e t Analysis

Memory Cards: An Emerging and Potentiatty Explosive Market

The memory card market is poised for rapid growth as portable computing, electronic photography, and other applications incorporate the memory card as an enabling technology.

By Nicolas Samaras Page 2

Control Applications: The Big Part of the Iceberg

Control applications comprise an estimated 70 percent of the semiconductor market. This article examines several high-volume applications and profiles opportunities for embedded

MPUs.

By Gregory Sheppard Page 6

Semiconductor Opportunities in Telecommunications

This article summarizes implications for the semiconductor industry on key market trends discussed at Dataquest's recent conference titled "Personal & Wireless Communications: The

Next Frontier."

By Krishna Shankar Page 9

Mil/Aero Outlook: Positioning for New Needs

As the Cold War ends, where are the semiconductor opportunities in defense and civilian aerospace electronics? This article examines the applications, the OEMs, and the changing requirements that impaa this market for chips.

By Gregory Sheppard Page 12

©1991 Daaquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

Semiconductor Application Markets Worldwide

Market Analysis

Memory Cards: An

Emerging and Potentially Explosive Market

w h a t Are M e m o r y Cards?

A memory card is a portable semiconductor storage device that contains memory ICs. It resembles a thick credit card (3.3mm) with an edge connector at one end (see Figure 1).

Figure 2

Memory Card Usage In a Portable PC

Memory cards jDerform a function similar to that of a floppy disk. They store binary data.

As program or data storage media, memory cards are not new; they have been around for at least 20 years. They have been used in computer games, point-of-sale (POS) systems, photocopiers, and laser printers. More recently, electronic organizers such as the Casio BOSS and the Sharp Wizard along with palmtop PCs such as the Poquet and the HP 95LX have

Figure 1

Example of Memory Card

Source: Adlron Corporation

begun using memory cards for data storage.

Figure 2 shows their application in portable

PCs.

The memory card form factor has not changed much over time, but the type of edge connector and the electrical/mechanical interface have.

The edge connector of a memory card is the conduit that allows data to move to and from the card's memory ICs. It defines the card's capabilities. To date, we have seen cards with a variety of connectors including 38-, 40-, 50-, and 60-pin.

Memory Card Varieties

Memory cards contain mostly semiconductor memory ICs that belong to one of the following famUies: mask ROM, EPROM, OTP, SRAM,

DRAM, EEPROM, and flash. DRAM memory cards are relative newcomers and are meant to be used as "extended/expanded" memory with no need for battery backup. SRAM cards with battery backup have been used as solid state

"floppies" in the current generation of electronic organizers. Until recently, SRAM cards were the

Source: Intel Corporation

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

Semiconductor Application Markets Worldwide

Table 1

M e m o r y Card Alternatives

Type

ROM

EPROM/OTP

DRAM

SRAM

EEPROM

Flash

Source: Dataquest (October 1991)

Density

32KB—16MB

32KB—2MB

64KB—12MB

32KB—2MB

32KB—256KB

32KB—4MB only nonvolatile memory cards. Flash memory

ICs today provide a promising alternative. Items such as language translating software a n d dictionaries typically come in mask ROM cards, as they are t h e most dense and least expensive.

Functionally, they are huge look-up data tables that n e e d n o change. Table 1 lists t h e various memory card alternatives.

M e m o r y Card Applications

Memory card applications include the following:

• Personal computers

• Factory automation

• Instrumentation and testing

• Avionics

• POS terminals

• Musical equipment

• Medical instrumentation

On Standards

What inhibited memory card growth in the past w a s the lack of standards. In J u n e 1989, the Personal Computer Memory Card Industry

Association (PCMCIA) was formed in the United

States, with a broad-based membership that included semiconductor c o m p a n i e s along w^ith software and hardware vendors. The PCMCIA's originally stated goal w a s to establish a standard for memory cards used with DOS-based

PCs. It succeeded rather quickly as standards go. The first revision of a memory card standard w a s published in August 1990.

Revision 1.0 of die PCMOA/JEIDA standard defined the following:

• The form factor—a device the size of a credit card, 3-3mm thick with a 68-pin socket connector

• The interface—parallel type bus, 8-bit/l6-bit

• The address space—64Mb

T h e PCMCIA w o r k e d closely with t h e J a p a n

Electronic Industry Device Association (JEIDA) a n d JEDEC. This close cooperation e n a b l e d t h e p r o m p t international acceptance of the standard.

Revision 2.0, as a n n o u n c e d in September, addresses XIP (eXecute-In-Place) and I / O functions such as m o d e m s and LANs for PCMCIA bus cards. Intel also announced the Exchangeable Card Architecture (ExCA), a hardware a n d software implementation of the PCMCIA Revision 2.0 system interface. It is Intel's stated intention to make ExCA an industry standard so that different types of cards (memory, LAN, m o d e m , and wireless commuriications) from different manufacturers will be interoperable.

Do Memory Cards Replace Hard

Disks?

Stricdy speaking, memory cards are not hard disk replacements. Rotating media has not b e e n terribly successful w^ith removable hard disks. A n u m b e r of companies have tried that approach, but technology and costs kept it out of t h e mainstream. Thus, after 20 years of PCs, w e are conditioned to think of hard disks as storage devices that belong inside the PC enclosure.

This idea is a technology-dependent perception, and there is n o reason w h y it should b e so.

O n the other hand, memory cards, being a solid-state storage medium, are removable a n d portable. At a density of, say, 20Mb, is a memory card acting like a "removable hard disk"?

We believe that it is.

Memory cards have the follow^ing advantages over floppy/hard disks:

• Faster access a n d transfer rates

• Space, power, a n d weight reduction

• More ruggedness

However, they d o have the following disadvantages:

• They are exjjensive.

• They have lower capacity.

The Cost IssueHow Important

Is It?

In 1991, the average selling price (ASP) of a

2.5-inch 40MB hard disk drive is $250.00, which translates t o $6.25 p e r megabyte. The

3-5-inch floppy cost is close to $1.00 f>er megabyte. By comparison, a 1MB flash card costs

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

Semiconductor Application Markets Worldwide approximately $300.00 or $300.00 per megabyte—a substantial disparity! Semiconductor memory certainly costs more.

The question is, "Can you put a floppy disk drive in a palmtop PC to take advantage of that cost disparity?" The answer is, "No." There is not enough power (or space). The issue,

then, is not cost. Here the removable storage medium dictates the product's capabilities and its success or failure in the marketplace. Without a memory card, a palmtop is nothing more than an electronic organizer. It is the memory card that transforms a palmtop into a fullfledged personal computer.

The Memory Card Market

As with any emerging technology, market size projections are difficult at best. The following assumptions may be used to gauge a portion of the total available market:

• The majority of palmtop and pen-based PCs will use memory cards (80 to 95 percent).

• The ratio of 1 to 3 cards per system for

1991 will increase to 3 to 5 by 1995-

• A portion of notebook PCs will use memory cards (10 to 20 percent over the same period of time).

Figure 3 provides some useful boundary conditions. Dataquest expects worldwide shipments of pen-based PCs to grow at a compound annual growth rate «IAGR) of 174 percent, from

96,000 units in 1991 to nearly 5.5 million units in 1995. At the same time, hand-held PC shipments will grow at a 108 percent CAGR from

503,000 units in 1991 to approximately 9.4 million in 1995. Together they amount to approximately 600,000 units in 1991, growing to almost

15 million by 1995. Some simple assumptions on memory card average selling prices indicate that this could easily become a billion dollar market by 1995.

Memory cards used in non-PC applications

(which may account for as high as 90 percent of total memory card shipments in 1991 and 40 to 60 percent by 1995) are not included in this discussion. Electronic still photography alone may provide an explosive market for memory cards.

Figure 3

PCs Employing Memory Cards

Millions of Units

12-

10

K X I Pen-Basad

Hand-Held

8^

9.365

2-

.503

,096

1991

Source: Dataquest (October 1991)

1.559

.353

1992

1993

3.087

3.498

5.588

5.422

1994

1995

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

Semiconductor Application Markets Worldwide

What Are the Key Developments

Needed for Memory Cards to

Succeed?

Three developments are necessary for the success of memory cards. These developments and the applications where they are needed are as follows:

• Cost reduction—all applications

Development of data-compression ICs—

PCs

XIP—^palmtop PCs

Cost Reduction

Flash memory cards hold the promise for becoming the least expensive form of solid-state storage. From a cell standpoint, flash rivals that of DRAM. Unlike DRAM or SRAM, it is nonvolatile, which means there is no need for battery backup. The need for bulk erasing of currentgeneration flash ICs creates a problem that requires clever solutions. With SRAM or DRAM cards, a single byte can be erased; EPROMderived flash most often can be erased at the chip level (i.e., the whole chip). Recently, some vendors have announced products that allow erasure of particular memory segments. A prime example is the Intel 28F001BX 1Mb flash memory, which is segmented into areas of one 8KB, two 4KB, and one 112KB—all of which can be independently erased and programmed.

EEPROM-derived flash is far more flexible at a cost premium (larger die). Flash EEPROM cells are larger than flash EPROM. Mask ROM memory cards will be the least exjjensive for the foreseeable future.

Data C o m p r e s s i o n ICs

Data compression ICs represent a key development for the electronic photography market and, to a lesser extent, for palmtop and penbased PCs. Data compression ICs will be the subenabling technology devices. Without them, the future of electronic photography is in doubt. Thirty-six exposures (pictures) can be stored in a 2MB flash memory card in compressed form. If no compression were used,

40MB would be needed!

eXecute-in-Place (XIP)

Simply stated, XIP allows a memory card to

"plug-and-play." That is, once the card is plugged into the PC, program execution begins much in the way a program runs after one types in the program name and hits carriage return. That procedure is in contrast with currentgeneration PC architectures that need to copy the program code from secondary storage (hard disk) to main memory (DRAM) before execution. A palmtop PC with XIP capability needs just a single copy of a program, usually stored in a mask ROM memory card.

The PlayersMemory Cards

Table 2 lists some of the companies active in the memory card market and their products.

Table 2

Memory Card Offerings

Intel

Fujitsu

Mitsubishi

Texas Instruments

1MB flash cards

4MB flash cards

256KB to 1MB flash cards

512KB to 16MB mask ROM memory cards

128KB to 1MB DTP memory cards

256KB to 1MB EPROM memory cards

64KB to 2MB SRAM memory cards

256KB to 2MB flash cards

512KB to 16MB mask ROM memory cards

128KB to 192KB EEPROM memory cards

64KB to 512KB SRAM memory cards

512KB to 12MB DRAM cards

1MB to 4MB DRAM cards

64KB to 512KB OTP cards

Source: Dataquest (October 1991)

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

Semiconductor Application Markets Worldwide

Other companies include Datakey, DuPont,

Epson, Fujisoku, ITT-Cannon, Maxell, and Oki.

The PlayersSolid-State Disks

A number of companies are working on solidstate disk (SSD) replacement—a challenging task, to say the least. SunDisk Incorporated, located in Santa Clara, California, chose to focus primarily on hard disk replacement (solid-state disk) with a proprietary flash memory technology and architecture. With backing from AT&T and Western Digital, the venture-capital-funded start-up launched three SSD products recently, all aimed at pen-based and palmtop PCs. The

2.5/5/lOMB SSD "plug-and-play" subsystems come with an IDE industry-standard interface.

SunDisk is working with Apple, GRID, IBM, and Polaroid. The company claims a 40MB capacity on two SSD cards. Memory cards are next. Another semiconductor start-up company that uses flash memory and a proprietary serial architecture optimized for solid-state disk replacement is NexCom Technology Inc., located in Santa Clara, California. Toshiba is thought to be working on a 4Mb flash SSD as well.

Some Thoughts o n the Future of

Memory Cards and PCs

In the past, the computer was the exf>ensive component and the storage medium (floppy disk), the inexp>ensive one. We've become accustomed to that oddity and do not seem to question it. However, the computer is just a machine that manipulates information. It is the information that is important and valuable, not the machine that manipulates it. So perhaps it is fitting that the information carrier, a memory card, may cost more than the computer it is attached to. In the future, we will be using platforms (palmtop PCs) that cost much less than the storage media (memory cards) they use. Imagine a $50.00 PC attached to a $100.00 memory card! At least losing the PC will not be a problem anymore!

Dataquest Perspective

Dataquest believes that memory cards represent an important enabling technology. They have the potential to transform still photography and to make the 35mm film and cameras that use it obsolete. In the process, they will change that industry and provide tremendous opportunities for growth in the consumer electronics market.

Memory cards will not eliminate rotating magnetic media any time soon. Instead, they will selectively replace them only when and where it makes sense. The bulk of the memory card growth will not come at the expense of rotating media. Growth will come from the creation of new markets. This should be good news for the semiconduaor memory industry.

Ultimately, we believe, memory cards may revolutionize portable PCs by enabling them to become smaller, more rugged, lighter, faster, and perhaps user friendly in a way that appeals to the vast majority of people who at present have no use for them. In doing so, memory cards may be the enabling technology that will make the PC of the future a true consumer item. •

By Nicolas Samaras

Control Applications:

The Big Part of the

Iceberg

An estimated 70 percent of all semiconduaors are used for control, or non-user-programmable applications. Items ranging from computer peripherals, to public telephone switching, to the family car represent such usage of semiconductors. If the semiconductor market was charaaerized as an iceberg floating at sea, the high profile PC and computer system markets would be the tip above the waterline and control applications would be the large portion below the water. The purpose of this article is to characterize some of the larger-volume control application opportunities.

Analog World Differentiates?

There are literally thousands of different types of control applications. Almost all have to do with the management or automation of analog world events. A laser printer generating a printout; the engine control unit managing ftjel, air, and spark; and even the precise movement of welding robots are examples of these events.

Control applications vary depending on what kind of data, signals (sensor or actuator), or power they manage (see Figure 1). Generally, modern control applications utilize an array of functionalized and general-purpose analog circuitry for input and output and microcomponents, ASICs, ASSPs, and memory to implement the control intelligence.

High-Volume Opportunities

Most control applications never wimess more than a few thousand units of production (particularly many instrumentation, factory control, and

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

Semiconductor A p p l i c a t i o n Markets W o r l d w i d e

Figure 1

Generic Control System

['••^••••••••••••l A n a l o g o r M i x e d S i g n a l

I i Digital

BUSS

Inter Face

Sensor

Antenna

Switch

Keyboard

Microphone

Amplifier

Signal Processing

Receivers

A/D :

ASSP

inputs

Data I/O

MPU/MCU

_DSP.__

A ^ P

Oirtputs

Control Memory

(ROM, EPROM...)

Scratch Pad/Buffer

(SRAM/DRAM)

Amplifier

Driver

D/A

ASSP

Display

Printer

Solenoid

Switch

Motor

Source: Dataquest (October 1991) mil/aero applications). Often these types are designed around architectural backbone standards like VME, AT, or PI bus. Some applications such as motor control (often counted as part of other applications, e.g., disk drives) involve tremendous unit potential but remain very fragmented.

Table 1 list some high-volume control applications. Cost sensitivity coupled vv^ith system feature enhancement is paramount for most of these applications. Almost all of these applications lend themselves to ASSP development where the NRE can be amortized over the volume. ASSPs help reduce board chip count and increase functional density. ASICs are often used in the first generation of a product or to support proprietary product differentiation.

Computing Power: Embedded

MPUs—A Growing Alternative?

As can be concluded from Table 1, control applications can be a real fab filler. Most consumer electronic designs use either custom or

4-bit MCUs today. The trend is toward incorporating 8-bit MCUs in some consumer electronics as more features are added for product differentiation. The 8-bit MCUs are finding increased usage in audio system and various automotive control modules like the antilock braking system (ABS). Mass storage designs are beginning to move out of 8-bit MCUs into l6-bit MCUs and DSPs. Image handling

(printers, digital video, HDTV, etc.), performance graphics, networking, cellular communications, and auto engine management applications are moving to 32- or 64-bit levels if they are not there already.

At the 16- and 32-bit performance levels, l6-bit

MCUs, embedded l6-bit MPUs and 32-bit RISC

MPUs, as well as l6- and 32-bit DSP units are competing for the designer's choice. Such l6-bit versions as the 80186, 68(X)0, 32010, and Z8000 have been found to be extremely popular in many (embedded) control applications (see

Table 2). Expect tailored variants of 32-bit RISC processors to become more popular as suppliers offer development help in the form of tools and real time software that takes advantage of

RISC performance. Digital signal processors are finding greater acceptance as a new generation of user-friendly development tools (such as code debuggers) emerge.

Dataquest Perspective

Control applications represent a broad market opportunity and should be treated as such. As noted in Table 1, the high-volume applications can be very fragmented and economically difficult to target. Dataquest recommends focus as the necessary discipline when taigeting one of these markets. Focus might be easy to say, but it is difficult to do. Without writing a management book, a simple recommendation is for clients to utilize their companies' relationships to identify and jointly develop compatible

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

8 Semiconductor Application Markets Worldwide

Table 1

High-Volume Control Applications Worldwide

Estimated

Unit Volume

(1991-1995)

(Millions) Application

Mass Storage

Rigid Disk Drives

Flexible Disk Drives

Tape Drives

Optical Disk Drives

Peripheral

Laser Printer

Fax Subsystem

Fax Machines

Graphics Subsystem

X Windows Terminals

Copiers

Communications

CO Switch (lines)

LAN NICs

Telephone Handsets

Cellular Handsets

Consumer/Auto

VCRs

Color TVs

Camcorders

Portable Audio

Appliances (w/controls)

Auto Engine Management

Auto ABS

Auto Stereo

Estimated weighted average in 1991

^CD-ROM to rewritable

Add-in card or on motbcboard

Source: Data^est (October 1991)

165

290

75

295

1,600 l60

65

240

225

230

12

10

40

55

45

105

4

22

235

55 l60

35

Semiconductor

Content

(Dollars)

48

14

42

15-80

50-200

20-150

30-200

40-300

80-200

25-250

3-15

40-100

2-15

30-150

20-50

18-60

60-100

2-35

1-15

15-30

15-30

7-25

Key Design

Considerations

Cost reduaion

Cost reduction

Ck)st reduaion

Cost/performance

Differentiation/cost

Differentiatiori/cost

Cost reduction

Performance/cost

Differentiation/cost

Differentiation/cost

Performance/cost

Differentiatiorj/cost

Differentiatiori/cost

Differentiation/cost

Differentiation/cost

Differentiation/cost

Differentiation/cost

Differentiation/cost

Differentiation/cost

Cost

Cost

Differentiation/cost

Table 2

Estimated Intelligent Control Market—Worldwide (Millions of Units)

32/64-Bit l6/32-Bit

16-Bit

8-Bit

4-Bit

Total

Source: Dataquest (October 1991)

1991

14.4

16.8

49.7

681.0

822.0

1,573.9

1995

44.0

33.0

268.0

1,175.0

1,200.0

2,720.0

CAGR (%)

1 9 9 1 - 1 9 9 5

78

18

52

15

10

15

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

Semiconductor Application Markets Worldwide opportunities early. Then they should prioritize funding of produa development accordingly.

High-volume control applications like the ones discussed in this article offer good opportunities for products such as application-specific standard products (ASSPs). Dataquest believes that digital, analog, and mixed forms of ASSPs are an excellent way to add value and margin. • user-friendly network that will enable "anytime, anywhere, any person" communications.

Table 1 shows a comparison of cost/functionality for the various personal communications system tyjjes, ranging from low-cost cordless telephone systems to digital PCN systems.

By Gregory Sheppard

Semiconduct€)r

Opportunities in

Telecommunications

Dataquest held its 1991 Telecommunications

Industry conference in August. The theme of the conference (Personal & Wireless Communications: The Next Frontier) aptly describes the focus of the conference speeches and panel sessions. This article provides an analysis and summary of key semiconductor trends for telecommunications applications as gleaned from the conference.

The A p p l i c a t i o n s

Semiconductors for Personal

Communications Networks

The ultimate goal of a personal communications network (PCN) system is to provide every person with a low^-cost pocket telephone (less than

$100) that is connected to a ubiquitous digital network using lo-w-cost microcell-based base stations. Such a pervasive digital PCN system will be compatible with existing and plarmed future long distance digital telecommunications networks. The goal is to provide a seamless,

PCN i n t h e U n i t e d States

The FCC has granted experimental licenses to

BellSouth, Graphic Scanning, Motorola, NYNEX, and PCN America (Millicom subsidiary). Experimental licenses are pending for American Personal Communications, Ameritech, GTE, McCaw, and several others. The goal of these FCC trials is to test the technology, cost, and userfriendliness feasibility of PCN systems based on microcell-based, spread-spectrum (2-GHz) transmission using Code-Division Multiple Access algorithms. Regulatory/licensing, standards, frequency allocation, and industry structure issues are exp)ected to be resolved for the U.S. market in the 1992/1993 time frame.

Meanwhile, the U.S. cellular telephone market is entering the analog-to-digital market transition period. Figure 1 illustrates Dataquest's estimates for the U.S. cellular telephone equipment market between 1990 and 1995. The market is expected to grow at a healthy compound annual growth rate (CAGR) of 19-4 percent, from

$1.8 biUion in 1991 to $3.6 billion by 1995.

Dataquest expects a gradual transition from a high analog content to a high digital content in the cellular telephone equipment market by

1995. The actual digital PCN telephone market is not expected to blossom into a large, mainstream market until 1995. Beyond 1995, the digital PCN equipment market is forecast to grow rapidly to $5.4 billion by the year 2000.

Table 1

Functional Comparison of Personal Communications Devices

Function

Communications Range

Mobility

Terminal Cost

Terminal Size

Battery Life

Base Station Cost

Source: Dataquest (October 1991)

Originate

200m

CV2

Limited; no handoff

Low ($100)

Small

High

Low

Paging

Receive

Metro area

High

Low ($100)

Small

High

Medium

CeUular

Originate/receive

>2 Miles

Automobile

High ($400-$700)

Medium/large

Low

Very high

PCN

Originate/receive

200M

Pedestrian

Low ($100)

Small

High

Low

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

10

Semiconductor Application Markets Worldwide

Figure 1

Estimated U.S. Market for Cellular Telephones

Millions of Units

10

Units

Revenue

CAGR

1991-1995

24.9%

19.4%

6.9

^vc^xv

Billions of Dollars

10

2.9

- . ^ s-'i

•V '<.

t ,

$1.8

i.i

7"!^^, m

1990

Source: Dataquest (October 1991)

PCN i n E u r o p e

The Pan-European GSM digital cellular phone standard is the most well documented of the wireless PCN standards. European semiconductor companies are racing to implement the GSM standard into application-specific standard products (ASSPs) using high-density gate array/ standard cells and full-custom designs. Many of these chip sets incorporate on-chip digital signal processing (DSP) cores to enable rapid A/D and D/A voice conversion.

PCN Developments

Many telecom IC companies are using application-specific integrated circuits (ASICs)

(gate arrays and standard cells) in combination with programmable logic array building blocks in order to hasten time to market for the firstgeneration digital cellular telephone design.

Dataquest anticipates that future PCN chip sets will be optimized as ASSPs using core telecom standard cells and DSP building blocks. Submicron high-performance CMOS and BiCMOS technology will enable shrinking chip counts, smaller sizes, lower power dissipation, and better performance in future-generation digital PCN systems.

The semiconductor industry believes that it can offer single-chip 0.5-micron technology

CMOS/BiCMOS PCN solutions using 500K-type embedded gate arrays with optimized embedded telecom macro cells such as DSP cores,

1991 1995 microcontrollers, ADCs, DACs, filters, cache

RAM, and Codecs.

Semiconductors for Premise Telecom

PBX Application

Dataquest expects wireless PBX technology to inject some life into the mature office premise PBX equipment market. Figure 2 shows Dataquest's estimate for the growth of the U.S. wireless PBX equipment market between 1991 and 1995. Wireless PBX equipment revenue is projected to grow at an astounding CAGR of 75 percent, from $36 million in 1991 to $340 million by 1995- The semiconductor market for wireless PBX pocket phones, base terminals, and network access interface cards should grow rapidly in response to the wireless PBX system growth projections.

Semiconductors for Data

Communication

Data communication appears to be going wireless—analogous to voice communication.

Much attention is focused on wireless local area networks (LANs) using RF and infrared transmisson technologies. Wireless LANs need to provide fiigh flexibility in office configuration, low change costs, and compatibility with hardwired backbone-wired LANs and long distance wide area networks (WANs). Figure 3 shows the growth of the U.S. wireless LAN and wireless

WAN markets between 1991 and 1995. The U.S. wireless LAN market, in particular, is expected

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

Semiconductor Application Markets W o r l d w i d e

Figure 2

Estimated U.S. W i r e l e s s FBX S h i p m e n t s

K'V"^X''"X\\N

^^c^^^^NN

'Xx\.\.v\.\>,.

^ j X

VvXSJ^X

W ' v s . '

K ' ^ ' X V V S

% ' ' \ W \ '''-•, ''''••- '•'-;•- ' •

^ms:m

1995

^ ' ^ . W X X X - --

Millions of Dollars

400

Thousands of Lines

400

360

320

280

E 3 Lines

Revenue

240

200

CAGR

1991-1995

77.6%

75.3%

363

H Vv. % Vv " v "V "'r- •>'• -I':

W v ^ v ™

•:>ss-'::s:-%r::"

v S ^ ' v V W V X

160

120

80

40

0

36,5

1991

S36

Estimated U.S. W i r e l e s s LAN/WAN E q u i p m e n t Markets

1995

Source: Dataquest (October 1991)

Figure 3

150.

120-

90-

60-

30

0

Millions of Dollars

300-

270

240

210

U.S. Wireless LAN

U.S. Wireless WAN

CAGR

1991-1995

123.6%

20.5%

ISO. t ^ . - ' - ' . ^ ' ^ -

1 9 9 1

Source: Dataquest (October 1991) t o grow explosively at a CJVGR of 124 percent, from $10 million in 1991 to $250 million by

1995. Numerous proprietary standards and protocols are emerging for wireless LAN applications using 920-MHz spread spectrum transmisson/conventional bipolar technology ICs, 18-GHz microwave transmission using high-jDerformance gallium arsenide technology ICs, and infrared point-to-point transmission using conventional

RF technology ICs.

Choice of Semiconductor

Technology for Telecom

Applications

Historically, telecom applications have lagged behind data-processing applications in their use of increasing very large scale integration systemon-a-chip architectures. The mixed-signal nature of telecom applications (RF, microwave, analog

I/O and amplification, and digital switching) has

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-6000 / Fax (408) 437-0292

0011333

11

12 Semiconductor Application Markets Worldwide traditionally implied relatively low levels of integration and performance. Ho'wever, with the recent trend toward digital cellular networks, high-speed digital data networks, and wireless transmission, a gradual segmentation of telecom semiconductor applications is occurring. Highperformance CMOS technology is being universally embraced for the data compression, digital signal processing, and switching applications.

Mixed-signal ASICs and ASSPs incorporating telecom core cells are being implemented in

CMOS as well as BiCMOS technology, which combines the high gain/frequency sensitivity of bipolar process technologies with the integration and low-power attributes of CMOS technology.

The adoption of microwave transmission standards in direa broadcast satellite communications, global positioning systems, and satellitebased global cellular telephone networks has spurred the acceptance of GaAs technology.

Dataquest Perspective

Telecommunications is shaping up to be a key semiconductor applications driver for the 1990s.

The emergence of cost-effective ultralarge scale integration-level submicron chip technologies, open network systems architecture, digital cellular networks, and advanced networking software is revolutionizing the telecommunications industry. Dataquest predicts the rapid emergence of a high-volume, cost-driven, highly competitive telecommunications chip set industry that will cater to a competitive open standards-based voice, data, still-image, and interactive fullmotion video communications market. Semiconductor companies that develop strong applications expertise in conjunction with influential telecommunications hardware/service companies can exploit the near-term emergence of comprehensive open standards for digital voice, data, and video communications. •

By Krishna Shankar

Mil/Aero Outlook:

Positioning for New

Needs

Defense Spending Impact

In spite of the war in the Persian Gulf and instability in the USSR, spending on military hardware is expected to decline globally for the foreseeable future. Although the war did a lot to highlight the effectiveness of electronics as the eyes, ears, and brains of weapon systems, vanishing super-power tensions are affecting political thinldng the most. It is most likely that defense sp>ending in the NATO countries w^ill stay flat in current dollar terms as that sector shrinks to be a smaller part of the respective economies. Aside from some short-term business with NATO-friendly countries like Saudi Arabia, the export of military systems to other countries is expected to be less allowed as a result of the Iraq example.

Figure 1 presents Dataquest's forecast of military and civil aerospace elearonics production. We expect military electronics production to remain roughly flat over the forecast period while the smaller civilian sector continues to have doubledigit growth as a result of expanding space and aviation demands.

Pockets Of Growth

In spite of an overall flat envelope, numerous growth opportunities exist in defense electronics. Many of these opportunities w^ill come from upgrades of existing platforms (aircraft, ships, and ground vehicles) as the production of new platforms is either slowed or discontinued.

Table 1 lists some of these opf)ortunities and some of the OEMs skilled in these technologies.

A Civil Market?

It is Dataquest's belief that the civilian space and aviation electronics sectors will witness solid growth for the bulk of the decade.

Although doubt remains regarding the viability of the space station program (and its worldwide elements) there remains a multitude of government scientific and commercial communications satellite projects. The electronic content of satellite development can run as high as 60 percent, posing a significant opportunity for radiationresistant components.

Spending by NASA (United States), ESA (Europe and Canada), and the two Japanese space agencies is expected to continue climbing significantly as earth resources and environmental monitoring programs expand. As many as 51 commercial communication satellites are exfjected to be delivered over the next three years, up from 39 in the previous three-year period. The need is being driven by increased demand for communication and direa-broadcast transponders. The Iridium cellular communication program promoted by Motorola could expand that number greatly, if it becomes politically feasible. The United States is expected to maintain nearly 57 percent of the satellite production, while France and the

©1991 Dataquest Incorporated / 1290 Bidder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

Semiconductor Application Markets W o r l d w i d e

Figure 1

Worldwide Military/Aerospace Electronic Equipment Production

Billions of Dollars

160-

140

Civil rr%1 Military

120

lUQ^

80^

60-

40-

20^

87.6

0

1990

Source: Dataquest (October 1991)

83.7

i^

\ V \ X \ '

;vv\V

1991

CAGR 1990-1995

IVIIIitary - 0 . 2 %

Civil 12.8%

Total 2.3%

SS.2

m

19S2

% #

•$::m

\ W ''•••

.. \ ''I: '••;,'

1993

95-4

w^Msp-m

1994

99.7

1995

Table 1

Defense Electronics Opportunities

G r o w t h A p p l i c a t i o n

Rugged computers/peripherals

K e y OEMs

Miltope, Rugged Digital, Codar,

Magnavox, DY-4

Space avionics/payloads GE, Hughes, Matra, TRW, Lockheed,

NEC, IBM, Honeywell, Loral

TI, Smitlis Industries, Hughes U.S. army vehicle electronics

<SAVA)

Avionic processor a n d data c o m m . m o d u l e s

Hughes, TI, Westinghouse, IBM,

Harris, GEC, ESD

Antisubmarine warfare sensing systems

Microwave/millimeter-wave front-ends

IR tracking/imaging

U n m a n n e d aerial vehicles

Source: Dataquest (October 1991)

AT&T, IBM, GE, Thomson

TRW, Westinghouse, GE, Lockheed,

GEC, Thomson, Raytheon, H u g h e s

Martin Marrieta, TI, Hughes, Loral,

GE, HAR

McDonnell Douglas, lAI, TRW,

Teledyne

N o t e s

VME b u s dominates, Futurebus+ for U.S. Navy

N e e d rad-hard KDs

Based o n 68020 a n d 1553 data comm,

SEM-E size, i960 a n d

R3000-based CPUs, PI bus, high-sfjeed data bus, 1553

Quieter subs m a k e this a priority

Modules for radar, comm. elec. warfare, missiles

Gulf w a r s h o w e d effectiveness

Market t o d o u b l e by 1999

13

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-«000 / Fax (408) 437-0292

0011333

14 Semiconductor Application Markets W o r l d w i d e

Figure 2

Commercial Airliner Deliveries

Units

1000

900 flOO

700-

600-

500-

400

300

200

E21 Airbus

^ ^ McDonnell Oougias

L X I Boeing

I I Oiher

430

100-

0-

^ ^

1961

L

Source: Prudential/Dataquest (October 1991)

United Kingdom will acquire 17 and 10 percent, respectively. France, Japan, Canada, Italy, Israel, and India are cutting into the U.S. dominance of space electronics. Companies such as MBB

(Germany), Alenia (Italy), Mitsubishi and NEC

Oapan), and lAI (Israel) are becoming forces in space electronics.

In civil aviation electronics, substantial opportunities remain in both avionics and ground-based air traffic control systems. Figure 2 shows the dramatic growth in aircraft deliveries over the past decade. A substantial growth of 5 percent per year in passenger air miles, especially for the Pacific region, is one of the main drivers of growth. Companies such as Honeywell,

Rockwell-Collins, Allied-Signal, Sextant Avionique, and GEC are the principal civilian avionics suppliers in the world.

Semiconductor Market: Slower,

Maybe Profitable

The worldwide market for semiconductors used in military and dvil aerospace electronics is expected to remain flat in 1991 but grow at a 5.8 percent rate from 1990 to 1995 (see Figure 3). Spending on replacement equipment for

Operation Desert Storm and orders by friendly

Persian Gulf states—^approximately $20 billion, mostly in the United States—is generating some one-time orders mainly for mature products.

Semiconductor consumption rates are expected to continue exceeding the equipment growth rate as new^ equipment and upgrades have

687

1990

764

>1.

\ . '•,. \ "-r \r ';- \ V ^

1995 more value absorbed by complex microprocessor units (MPUs), application-specific standard products (ASSPs), application-specific integrated circuits (ASICs), microwave monolithic ICs

(MMICs), and memory ICs.

The dramatic cuts to the 1991 equipment procurement accounts in the United States, the

United Kingdom, and France are cause for a bleak 1991. Some recovery is expected in 1992 as platform upgrades gain priority in spending budgets. Over the coming decade, expect the

Japan/ROW market to grow laiger as Japan continues developing its civilian aerospace electronics industry and countries like Taiwan and

South Korea produce more of their own defense electronics.

Semiconductor Market Qualities

Earlier in this article we noted which mil/aero equipment categories are the most lucrative for semiconductor marketers and which principal

OEMs are producing that equipment. In the coming years mil/aero applications will make the following requirements of semiconductor technology:

• Reduced power consumption and space, increased reliability and maintainability—

These are the assumed drivers behind any upgrade or new design. Most often, this is translated into emphasis on CMOS/BiCMOS/

GaAs VLSI, mixed-signal, and hybrid conversion to monolithic or multichip modules.

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

Semiconductor Application Markets Worldwide

Figure 3

Worldwide Military/Aerospace Semiconductor Consumption by Major Product Category

MOS

Mtcrocomponent

10%

IS

MOS

Microcomponent

13%

1990 = $2.8 Billion

Source: Dataquest (October 1991)

Emphasis on minimizing adverse effects of radiation—Total dose, dose rate, neutron emission, and soft memory array errors lead a long list of parameters being considered by

OEM designers. Enhanced bulk CMOS will continue vying with silicon on sapphire and silicon on isolator for designs.

• Testability, designability, and replaceability—

These issues affect every design decision.

- Already in the United States, VHSIC hardw^are definition language (VHDL) design capture is required on all new^ projects as platform life cycles head for multiple decades and semiconductor life cycles shrink to just a few years. A major reason is to reduce the effort in sourcing semiconductor products years after the original product has been phased out.

- Test features such as JTAG and the TM bus, which affect chip/package design, are being required of an increasing number of programs as well,

• More digitally based, real-time control requiring extensive use of high-performance (CISC or RISC) MPUs, digital signal processors, memory, and real time software (principally

Ada)

• More modularity and commonality—Emphasis is on standardizing on data communication standards (1553, HSDB), backplane buses (PI and Futurebus+), common modules (SEM-E or multichip module upgrade), and high-level language software portability.

1995 = $3.8 Billion

• Greater bandwidth and sensitivity for frontend (analog envirorunent) processing—This requirement translates into higher frequencies

(exceeding 50 GHz) for RF components like

MMICs and finer-resolution infrared.

Products In Demand

Table 2 lists some products that Dataquest believes will be in strong demand over the next five years. Each of these product areas can trace its demand from one or more of the above factors. It is important to note that mature standard logic, linear ICs, and MPUs usually enjoy a long phaseout period, and the opportunity for aftermarket suppliers and emulating solutions is proving to be attractive.

Dataquest Perspective

As the tide of this article implies, to be a successful (profitable) supplier to the military and aerospace community, focus is recommended.

We recommend that semiconductor companies with "design-win" class products like ASSPs,

ASKDs, and MPUs select a manageable list of accounts to target. As OEMs continue consohdating and reduce the length of their supplier lists, it becomes paramount for semiconductor suppliers to be responsive if they wish to remain on the list. We suggest studying which OEMs are the best positioned for growth opportunities in their resf>ective markets. Table 1 provides one such list.

A side benefit of maintaining a strategic account focus is that a good relationship will drive the

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

16

Semiconductor Application Markets Worldwide

Table 2

K e y S e m i c o n d u c t o r Products f o r Mil/Aero A p p l i c a t i o n s i n North America a n d E u r o p e

Product Area

Bus interface (all technologies)

CMOS/BiCMOS gate arrays

CMOS/BiCMOS cell-based ICs

(Mixed signal)

CMOS PLDs

(FPGAs)

Data comm. ICs (1553/SCC/HSDB, etc.)

Digital signal processors

8-bit MCU l6-bit MPU

32-bit MPU

SRAM

EEPROM

Flash memory

GaAs MMIC

Amplifiers

Data conversion

Power MOSFET

Rad-hard/level-S ICs

Source: Daiaquest (October 1991) decisions regarding -which commercial portfolio products should be militarized a n d -when that should happen. It is important t o note commercialization trends like the conversion to standard military drawing parts or even a rugged specification; the adoption of the qualified manufacturers list should make it easier (less costly) for semiconductor companies to serve the market.

However, OEM procurement oiganizations are increasingly wary of interlopers in t h e market, a n d sourcing preference ultimately will reside with those companies thoroughly committed to serving the unique military quality a n d contracting needs. •

By Gregory Sheppard

In Future Issues

1991-1995

Market Size ($M)

1,010

390

45

580

740

670

355

790

1,305

680

(120)

385

(110)

305

170

140

245

155

1,755

The following topics will b e featured in future issues of Semiconductor Application Markets

Worldwide Dataquest Perspective:

• Opportunities in digital video

• Trends in rigid disk drive applications

• T-Carrier semiconductor market opportunities

For More Information . . .

O n the topics in this issue Semiconducor Application Markets (408) 437-8261

About other Dataquest publications Sales (408) 437-8250

About upcoming Dataquest conferences Conferences (408) 437-8245

About your subscription Customer Service (408) 437-8402

Via fax request Fax (408) 437-0292

The content of this report represents our interpretation and analysis of information generally available to the public or released by responsible individuals in the subject con^Huiies, but is not guaranteed as to accuracy or completeness. It does not contain material provided to us In confidence by our clients. Individual companies reported on and analyzed by

Dataquest may be clients of this and^or other Dataquest services. This information is not furnished in connection with a sale or offer to sell securities or in connection with the solicitation of an offer to buy securities. This firm and Its parent and/or their officeis, stockholders, or members of their Amities may. from time to time, have a long or short position in the securities mentioned and may sell or buy such securities.

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011333

Dataoyest

a company (^

The Dun & Bradsticct Corporation

Dataquest

Perspective

Special Edition

October 28, 1991

Market Analvsis

Worldwide Semiconductor Industry Forecast: Fourth Quarter 1991

Dataquest expects recovery of the U.S. economy to stimulate spending on electronics systems, spurring the worldwide semiconductor market to grow 13.5 percent in 1992, up from 9.3 percent growth in 1991, and to grow 15.7 percent in 1993.

By Terrance A. Birkholz Page 2

The Downside and Upside to Our '92 Forecast

Semiconductor end-use markets are currently giving mixed signals concerning a recovery.

Our 1992 worldwide semiconductor forecast assumes a moderate recovery in the end-use markets. However, there is both a downside and an upside to this assumption.

By Mark FttzGerald Page 9

©1991 Dataquest Incorporated / 1290 Bidder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

Special

Edition

Market Analysis

Worldwide Semicondtictor Industry Forecast:

Fourth Quarter 1991

S u m m a r y

Dataquest expeas the worldwide semiconductor market to grow 13.5 percent in 1992, up from

9.3 percent growth in 1991, and to further grow

15.7 percent in 1993 (see Figure 1). Recovery

oi the U.S. economy will stimulate worldwide systems production, which in turn will stimulate semiconduaor consumption. In the short term, the cyclical upturn of the data processing market will help boost MOS memories' contribution to overall growth and help firm the foundation of microcomponent growth. In the long term, semiconductor market growth will be driven by networking the stock of data processing capabil-

ity, computer-based graphics, and image-based processing, placing new demands on processing power and the associated complement of memory capacity.

Dataquest's Semiconductor

Forecast Methodology

Dataquest's semiconductor forecast methodology leverages the resources of its parent, The Dun &

Bradstreet Corporation, as well as the considerable internal resources of Dataquest.

Dun & Bradstreet information is used to develop the macroeconomic forecasts for the world's major economies. This forecast identifies trends in the economic health of the world's leading consumers and producers of electronic equipment. Using this forecast in conjunction with input from Dataquest's regional offices,

Dataquest identifies the likelihood of whether a particular region or country will increase or decrease its consumption of electronic equipment.

Dataquest's Semiconductor Applications Mariset group, along with Dataquest's various electronics systems groups, provides a long-range outlook for the overall growth of the electronic equipment market. Semiconductor content ratios are developed by region to reflect the growing penetration of semiconductors into electronic equipment. This establishes a five-year compound annual growth rate (CAGR) for total semiconductors for a five-year period from a demand-side perspective.

Dataquest's worldwide Semiconductor service and its Semiconductor Equipment, Materials, and

Manufacturing service, in conjunction with its various regional offices, collaborate to formulate ejq^ectations of semiconductor market shortrange fluctuations around the long-range trend.

Tactical market issues and anticipated semiconductor materials demand significantly impact the

Figure 1

Worldwide Semiconductor Consumption (Factory Revenue, Dollar-Based Annual Growth)

Percentage

40

30 -

20 -

10 -

0

-10 H

-20

\ \ \

1985 1986 1987

1988 1989 1990 1991

1992 1993 1994 1995

Source: Dataquest (October 1991)

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

Special Edition short-range forecast out to 12 months. Semiconduaor equipment purchases and semiconductor device trends drive the forecast in the 12- to

24-month time frame. Semiconductor fab facilities and long-term semiconductor device trends have the greatest impact on the forecast period covering two to five years.

The final step in the forecast process is to reconcile expected fluctuations in the electronics market and trends in the semiconductor industry so that the fluctuations do not inexplicably diverge from semiconductor industry trends.

Dataquest anticipates that, in the absence of shocks to the market, market fluctuations converge toward the long-term trend.

Forecast Assumptions

The worldwide economic climate is expeaed to improve in 1992. The Dun & Bradstreet Corporation forecasts the following outlook for the

Group of Seven CG7) countries (see Figure 2):

• The U.S., C:anada, and U.K. economies will register negative real economic growth in

1991 but recover at rates of 2.8, 4.0, and

1.8 percent real gross national product/gross domestic produa (GNP/GDP), respectively, in 1992.

Figure 2

G7 Countries' Estimated Economic Outlook

Real GNP/GDP Growth, Local Currencies

Percentage

6

Real GNP/GDP growth is expected to accelerate in France and Italy during 1992, from 1.3 percent in 1991 to 2.4 percent and firom 1.4 percent in 1991 to 2.5 percent, respectively.

• Real GNP/GDP growth is expected to decelerate in Germany and Japan during

1992, from 3.0 percent in 1991 to 2.0 percent and from 4.5 percent in 1991 to 3.2 percent, respectively. TTie cost burden of Germany's reunification and the rise in Japan's cost of capital are moderating these countries' shortterm growth prospects. Both economies are expected to reaccelerate in 1993-

Growth in the G7 economies is expected to converge toward the countries' respective steady-state rates through 1994.

The improved economic prospects bode well for the semiconductor industry outlook, given that computers and related electronic gear represent a significant share of the G7 economies' business fixed investment.

Acceleration of worldwide systems production growth—to 9-0 percent in 1992 from 5.4 percent in 1991—will be accompanied by the

-2 -

-3 -

United States ' Canada ' Japan ' France ' Germany ' Italy ' United

Kingdom

Country

1989 1990 H 1991e ^ 1992e g 1993e Q 1994e

Source: The Dun & Bradstreet Q>iporation

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

Special Edition resumption of economic growth (see Figure 3), as shown by the following factors:

• Business conditions in the data processing and consumer markets are expeaed to show significant improvement as businesses and households begin to relax their budget constraints

• Data processing up 10.3 percent in 1992 from 5.8 percent in 1991

O Consumer up 9-8 percent in 1992 versus

6.8 percent in 1991

• Transportation electronics production growth is expected to more than double—to

12.6 percent in 1992 from 5-7 percent in

1991—spurred by increased consumer spending, combined with increasing share of electronic systems' added value to new vehicles.

• Communications and industrial electronics growth are expected to remain positive and stable. Spending on medical electronics and analytical instruments helped bolster the industrial segment from recession-induced decreased spending on measuring and controlling electronics.

• MiKtary/dvilian aerospace electronics was hit hard by "Washington budget cuts in 1991,

Figure 3

Worldwide Electronics Production

(Factory Revenue, DoUar-Based Annual Growth)

but this segment is expected to resume modest growth (at a permanently lower dollar level) as western defense agencies upgrade existing systems with more sophisticated electronics.

Semiconductor Outlook: Overview

Dataquest expects the worldwide semiconductor market to grow 9.3 percent in 1991 to

$63.6 billion, u p from $58.2 billion in 1990, and 12.1 percent in 1992 to $72.2 billion (see

Table 1). (Note that Table 1 expresses the value and growth of the Japan and Europe markets in local currencies' terms. In addition to valuing the worldwide market assuming current exchange rates, the worldwide market is valued in U.S. dollars, assuming constant 1990 exchange rates, which removes the effects of exchange rate variation on growth.)

Our Oaober 1991 forecast represents a downward revision to our May 1991 forecast when

•we forecast the market to grow 13.7 percent in

1991 and 16.6 percent in 1992. Approximately

65 percent of the reiHsion in 1991 and 50 percent of the revision in 1992 is accounted for by appreciation of the US. dollar against the

Japanese yen and major European currencies since the May forecast.

Percentage

15

10 -

5 -

Data ' Communications < Industrial

Processing

Consumer

Equipment Market

1989

Mil/Aero ' Transportation

1990 H 1991e ^ 1992e j ^ 1993e Q 1994e ^ 1995e

Source: The Dun & Bradstreet Corporation

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

Special Edition

Table 1

Worldwide Semiconductor Consumption b y Region—^1990-1995

(Factory Revenue in U^. Dollars and Local Currencies)

North America ($M)

Annual Growth (%)

Japan ($M)

Annual Growth (%)

Japan (¥B)

Annual Growth (%)

Exchange Rate: ¥ per U.S.$1

Europe ($M)

Annual Growth (%)

Europe (EcuM)

Annual Growth (%)

Exchange Rate: Ecu per U.S.$1

Asia/Pacific-ROW ($M)

Annual Growth (%)

Worldwide ($M)

Annual Growth (%)

Worldwide ($M in 1990 U.S.$1

Exchange Rates)

Annual Growth (%)

Source: Dataquest (October 1991)

1990

17,386

-3.1

22,508

-2.1

3,241

2.1

144.00

10,661

9.3

8,380

-6.0

0.786

7,670

17.6

58,225

1.8

58,225

0.7

1991

18,483

6.3

25,544

13.5

3,501

8.0

137.06

10,828

1.6

8,890

6.1

0.821

8,792

14.6

63,647

9.3

62,899

8.0

1993

23,888

15.2

33,341

12.9

4,601

12.9

138.00

13,777

19.2

11,683

19.2

0.848

12,532

20.4

83,538

15.7

83,235

15.8

10,405

18.3

72,213

13.5

71,894

14.3

1992

20,728

12.1

29,524

15.6

4,074

16.4

138.00

11,556

6.7

9,799

10.2

0.848

15,335

11.3

13,004

11.3

0.848

14,486

15.6

93,787

12.3

93,446

12.3

1994

26,758

12.0

37,208

11.6

5,135

11.6

138.00

1995

28,816

7.7

CAGR (%)

1990-1995

10.6

40,232

8.1

5,552

8.1

138.00

16,368

6.7

13,880

6.7

0.848

16,246

12.1

101,662

8.4

12.3

11.4

9.0

10.6

16.2

11.8

101,276

8.4

11.7

Although 1991 is shaping up as a modestgrowth year—worldwide market growth averaged 19.1 percent per annum in the 1985 through 1990 period—^it is nonetheless a rebound over last year's 1.8 percent growth.

Growth in 1991 was hampered by the following three factors:

• Deeper- and broader-than-expected U.S.-led economic recession

• The recession's growth-an-esting affect on computer spending

• Military spending cuts

Growth is forecast to accelerate through 1993 but will be constrained by the relatively moderate rate of overall economic recovery and the effects of saturation and maturity in the relatively developed markets.

Semiconductor Outlook: Regions

North America

The North America systems and semiconductor n^rkets were hit hard by the econonwc recession of 1991. Both 1992 and 1993 are expeaed to be years of accelerating growth as businesses resume computer and related equipment spending in an environment of renewed vigor in fixed investment. The following three factors will tend to restrain semiconductor growth below the peak rates experienced in the last decade:

• Two-thirds of desktops have computers on them—^After 30 years of innovation and booming sales, there are inevitably fewer opportunities for investment.

• The computer market's share of U.S. capital investment more than doubled, from less than 3 percent in 1977 to about 7 percent in

©1991 Daiaquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

Special Edition the mid-1980s, but has remained unchanged since then.

• Previously, new systems—those without dose substitutes—enabled the computer industry to increase its share of capital spending faster than overall investment fell.

These factors should not be construed to mean that opportunities for further semiconductor penetration are absent through the forecast horizon. Indeed, the next round of computer and computer-related equipment spending will involve connectivity/networking and higher-level graphics and image-based processing. Both of these areas represent the new frontier for microcomponents and the associated memory complement and for analog and mixed-signal

ASIC.

Japan

Japan's growth was hit hard in the first quarter of 1991 by the combined effea of a recession that was already under way in the United States and complicated by the Gulf war. We expect

Japan's market to revive in 1992 in response to the resumption of chip and systemis export growth to the United States and Europe.

Renewed vigor in the computer arena will help firm MOS memories, while advances in camcorders, large-screen TVs, wireless and car telephones, and robot systems designs will boost microcontroller unit (MCU), MOS logic, and analog device growth.

Japanese manufacturers will use the remainder of 1991 to position themeelves to take full advantage of the market's upturn in 1992.

Europe

A recession in the United Kingdom plus the reunification-induced drag on the German economy restrained Europe semiconductor market growth in 1991. However, appreciation of the

U.S. dollar against the major European currencies masks the true situation of the market: In dollar terms, the market is expected to decelerate to 1.6 percent growth in 1991 from 9-3 percent growth in 1990. In European currency unit

(Ecu) terms—a good proxy for a weighted basket of European currencies—^market growth is expected to resume expansion at a rate of

6.1 percent in 1991 from a 6.0 percent shrink-

age in 1990 and to accelerate to 10.2 percent growth in 1992.

Improved overall business conditions will help firm up indigenous PC production and consumption in 1992, which will translate into improved prospects for ASICs, microcomponents, and, in particular, MOS memories. In the long term, however, the ASIC market will be fraught with severe average selling price (ASP) pressure stemming from increasing integration and smaller production volumes per design.

Asia/Paciftc-Rest of World

Growth in Asia/Padfic-Rest of World (ROW) is and will continue to be fueled by domestic companies' investment, but more importantly from foreign direct investment. The inflow of foreign capital, combined with the relative immaturity of the industry, shields the semiconductor business from the wide swings in activity that tend to rock the other, more established regions. Even so, memory and microcomponent consumption have been severely hurt by the softness of PC business, while analog consumption has felt the pinch of households' curtailed consumer electronics purchases.

Dataquest exf>ects semiconductor consumption growdi to accelerate in 1992 and 1993 as the western export markets stimulate data processing and consumer electronics production.

Semiconductor Outlook: Devices

Table 2 presents worldwide detail of the semiconductor device forecast. Volatile pridng make

MOS memories the swing factor accounting for year-to-year changes in overall market growth.

Microcomponents provide more stable growth in both the short and long term. Overall, worldwide revenue growth is expected to accelerate in 1992 following recovery of the systems markets and peak in 1993. We expect growth to moderate in 1994 and 1995.

Bipolar Digital

The bipolar logic market was hit hard in 1991 by the recession: Businesses postponed major purchases of mainframe and high-end computer equipment, the largest users of these devices.

Standard logic, as a share of total bipolar logic, declined at a faster-than-expected rate, also in response to slower-than-expected market conditions.

Compounding the recession's cyclical effects are important structural and technological dynamics:

Through 1995, bipolar logic will continue to be replaced by CMOS, BiOIOS, and GaAs IC:s as these devices become more cost competitive.

Also, as chip functionality and integration increase, unit volumes of ASIC designs will decrease; that is, ASIC manufacturers face the

©1991 Dataquest Incorporated / 1290 Bidder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

Special Edition

Table 2

Worldwide Semiconductor Consumption b y Device

-1990-1995

(Factory Revenue in Millions of U.S. Dollars)

Total Semiconductor

Annual Growth (%)

Total IC

Annual Growth (%)

Bipolar Digital

Annual Growth (%)

Bipolar Memory

Annual Growth (%)

Bipolar Logic

Annual Growth (%)

MOS Digital

Annual Growth (%)

MOS Memory

Annual Growth (%)

MOS Microcomponent

Annual Growth (%)

MOS Logic

Annual Growth (%)

Analog

Annual Growth (%)

Total Discrete

Annual Growth (%)

Total Optoelectronic

Annual Growth (%)

Source: Dataquest (October 1991)

1990

58,225

1.8

13,091

-20.0

10,068

22.8

9,133

7.9

10,571

12.6

8,235

7.5

2,687

2.3

47,303

0.8

4,440

-1.6

459

-15.0

3,981

0.3

32,292

-2.2

1991

63,648

9.3

51,863

9.6

4,095

-7.8

414

-9.8

3,681

-7.5

35,926

11.3

13,418

2.5

12,063

19.8

10,445

14.4

11,842

12.0

8,777

3,008

11.9

1992

72,211

13.5

59,672

15.1

3,966

-3.2

407

-1.7

12,044

15.3

13,210

11.6

9,241

5.3

3,298

9.6

3,559

-3.3

42,496

18.3

15,958

18.9

14,494

20.2

3,436

-3.5

50,980

20.0

19,378

21.4

17,465

20.5

14,137

17.4

1993

83,537

15.7

69,840

17.0

3,843

-3.1

407

0.0

15,017

13.7

10,040

8.6

3,657

10.9

1995

101,661

8.4

86,141

8.9

3.390

-6.8

352

-6.9

3,038

-6.8

64,546

10.0

24,447

8.3

22,216

11.2

17,883

11.1

18,205

8.3

11,172

4.8

4,348

8.1

3,259

-5.2

58,661

15.1

22,583

16.5

19,982

14.4

16,096

13-9

16,808

11.9

10,656

6.1

4,024

10.0

1994

93,786

12.3

79,106

13.3

3,637

-5.4

378

-7.1

CAGR (o/o)

1990-1995

11.8

12.7

-5.3

-5.2

-5.3

14.9

13.3

17.2

14.4

11.5

6.3

10.1 prospea of increasingly complex chips and smaller volume production runs.

Bipolar logic's remaining life cycle will be driven by the quick-processing and switching requirements of centralized, high-end computer systems.

MOS Memory

Slow DRAM bit growth in 1991, which in turn added to ASP softness, combined to make 1991 revenue only marginally improved over 1990.

Weak market conditions have also permitted users to extend the 1Mb life cycle until higher density per-bit prices fall to appropriate levels.

The anticipated recovery of computer production in addition to the emeigence of memoryintensive PC applications—^including more powerful operating systems, user-friendly graphical user interfaces, and digital video—^will help drive DRAM bit growth in 1992 and beyond.

The emerging generation of laptop, hand-held, and pen-based PCs is also expeaed to give renewed vigor to the DRAM market.

Softness in the PC market, vendors in oversupply, and customers selling off inventory have combined to make for very slow SRAM bit growth and rapidly falling ASPs. Slowing bit growth and Korean/Taiwanese manufacturers

"buying" market share will constrain revenue growth. Actual future revenue growth may be further constrained as manufacturers follow through with plans to switch fab capacity to

SRAM devices, exacerbating an existing overcapacity situation. On the positive side, expected growth will be bolstered by further application of caches in PCs, and slow SRAM bit growth will be fueled by new applications in consumer markets.

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

8 Special Edition

The nonvolatile memory market will be bolstered by continued penetration of flash memories but at the expense of EEPROM market growth. We expect flash growth to accelerate in the forecast period, fueled by consumer and data processing applications. Acceptance of palm-top and pen-based computers and the substitution of memory cards for disk drives for the task of mass storage will be critical to flash's future growth. In the long term, consumer acceptance of electronic photography will be the wild card that adds a superktive increment to growth.

MOS Microcomponents

Notwithstanding the slow^down in PC shipments, microcomponents is expected to be the fastestgrowing device family in 1991, 1992, and, on average, through 1995- Two factors contribute to this situation. Intel's proprietary position in the 80486 MPUs places a floor underneath prices and MCUs are steadily penetrating consumer electronics and telecommunications.

Furthermore, Dataquest expects microcomponent growth to be fueled by the trend toward higher-performance PCs that include multimedia and networking funaions, which in turn will require a higher level of dedicated processing power for implementation.

Market revenue will be boosted by Intel's proprietary edge in the MPU market with its 80486 chip. Helping to constrain revenue growth, however, will be ASP pressure originating from competitive alternate sources to an Intel-based

PC (for example, the AMD-led price pressure in

80386 MPU applications).

MOS Logic

•Workstations, laptop PCs, and telecom applications are the driving forces behind today's MOS logic growth, although the lackluster showing in the PC arena at large tends to drag unit and revenue growth below^ what it would be otherwise. The recession has spelled lower unit volumes per ASIC design, putting a further squeeze on manufacturers' profit margins. We look to field-programmable gate arrays, MOS gate arrays, CBICs, and application-specific standard products to drive future device growth and to MOS ftall-custom chips to restrain growth.

Analog

Dataquest's analog forecast remains essentially unchanged from the May forecast. As 1991 draws to a close, Dataquest will be looking to consumer confidence to improve, forming a firm foundation for 1992 growth. Beyond peak growth in 1993, analog as a product family faces the prospect of decelerating growth resulting from product maturity in laige segments of the market plus decelerating growth in some

(mature) end markets. Integration of analog functions to MPU and digital signal processing chips, however, will provide continued vitality to analog technology. We expect telecomspecific applications and computer-related massstorage and graphics applications to be the areas driving incremental growth.

Dataquest Perspective

Dataquest expects the 1990 through 1995 period to be characterized by relatively moderate market growth: Average growth in the 1990 through

1995 period is forecast to be 11.7 percent per annum versus 19.1 percent per annum in the

1985 through 1990 period. Part of this growth deceleration is a result of the moderation of the major world economies' growth prospects vis-avis the decade of the 1980s. More important, however, are the combined effects of the maturing end-use markets on the demand-side and the increasing incremental costs associated with marginal changes in manufacturing technology and system/chip performance.

In the 1980s, the workplace and households in the world's major industrialized economies could be characterized as a vacuum waiting to be (ftirther) filled by the breath of solid-state technology. The void was filled with desktop processing systems and VCRs, systems that were unrivaled by close product substitutes.

The task of the 1990s will be to continue to

add to the stock of electronic gear but also

(and at least as important) to enhance the stock's productivity through, for example, networking and image-based processing. Both these areas are new and fruitful ground for cost-competitive, iimovative, and technologyoriented semiconductor companies. But because of the relative complexity of these systems, their investment profiles will likely be more smooth—^less "peaked" than, for instance, the booming PC market of the 1980s.

Semiconductor manufacturers are advised not to miss out on the plodding progress the workplace segment is making toward connectivity and image-based processing while waiting for the next PC boom. •

By Terrance A. Birkholz

©1991 Dataquest Incorporated / 1290 Eidder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

Special Edition

The Downside and

Upside to Our

'92 Forecast

Dataquest's semiconductor forecast for 1992 calls for 13.5 percent growth in worldwide device sales, u p from 9.3 percent growth in

1991. A critical assumption of the forecast is an improvement in the semiconductor end-use markets—data processing, consumer, communications, industrial, military/aerospace, and transportation. Our forecast assumes that a moderate recovery in the major end-use markets will be well under way by the first quarter of 1992.

Table 1

1992 Semiconductor Forecast Probability

Distributioii

ProbabiUty Annual Growth

Rate (%)

<9

>9 but <12

>12 but <15

>15

Total

Somce: Dataquest (October 1991)

0.15

0.25

0.50

0.10

1.00

This assumption has a downside. Currently, there are few signs of a recovery in most of the semiconductor end-use segments. If the major end-use markets, i.e., data processing or consumer, fail to turn u p soon, as we have assumed in our forecast, then our estimated growth for the w^orldwide semiconductor industry in 1992 may be too high.

On the other hand, there is also an upside to our forecast. The semiconductor end-use markets have historically seen strong growth as the electronic equipment industry pulled out of a recession. If history repeats itself and there is a strong recovery in end-use markets rather than the moderate growth assumed in our forecast, then our estimated growth for the worldwide semiconduaor industry in 1992 may be too low.

Dataquest believes that business conditions in data processing will show improvement in 1992 as businesses begin to relax their budgets. Data processing is forecast to grow 10.3 percent in

1992 versus 5.8 percent in 1991.

To achieve our estimated growth in 1992, the computer equipment cycle must begin to turn up in the fourth quarter of 1991. However,

August data from the U.S. Department of

Commerce (DOC) on office and computing equipment are still giving mixed signals (see

Table 2). Orders were up 9-0 percent in August

1991 versus monthly orders a year ago. Yet, last year's data were very weak because of the

Mideast crisis, so 9-0 percent growth over an

August 1990 base cannot be viewed as a strong positive signal.

Figure 1

Semiconductor End Use b y Application

Segment

Perhaps the best methodology for bracketing the upside and downside of our forecast is to consider different outcomes in terms of probabilities (see Table 1). In order to better understand the qualitative arguments for assigning these probabilities, a more careful review of the factors driving individual end-use markets must be considered.

Communication —.

14.1% 1

Applications

Data Processing

Data processing applications accounted for

45.3 percent of the semiconductors shipped in

1990 (see Figure 1). This segment includes mainframe computers, minicomputers, workstations, personal computers, and peripheral equipment. It is quite obvious from the size of this end-use market that the health of the semiconductor industry is tied very closely to the fortunes of the data processing equipment industry.

/ Consumer \ / JT^^^

1 20.6% \l^''-'''''^

\ Data Processing

\ 45.3%

Percentage

Transportation

4.7%

Source: Dauquest (October 1991)

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

10

Special Edition

Table 2

0£Eice and Computing Equipment Data

U.S. Department of Conunerce

Monthly 1991 Growth Rate versus Same Month in 1990

Orders

Shipments

Backlog

Inventory

Production

Source: U.S. Depaitment of Commeice

June 1991

4

-3

1

-16

-5

July 1991

-13

1

-4

-14

-1

An optimistic note in the DOC data is the inventory cycle. Inventories were depleted at an 11 percent clip, while shipments declined by 1 percent. At some point, we anticipate that computer companies will be forced to begin ramping production in order to replenish their inventories. Assuming that inventory levels are very lean, a strong recovery in data processing will cause semiconductor demand to snap back, and growth could well surpass the 13.5 percent we have forecast.

But the bottom line is that orders for data processing equipment have been weak through the third quarter of 1991. And although it is a little early to be an alarmist, if orders continue to run at current levels through the fourth quarter, we expect our forecast to be optimistic. Needless to say, any delay in an upturn for data processing wiU only push the semiconductor industry recovery out further.

Consumer

Consumer applications accounted for 20.6 percent of the semiconductors shipped in 1990

(see Figure 1). Consumer electronics is forecast to grow 9.8 percent in 1992 according to Dataquest. To achieve this growth, U.S. consumers will have to increase their spending within the next several quarters.

August 1991

9

-1

-1

-16

-3 released on October 3 show little improvement in the U.S. unemployment rate, which is stuck in the 6.7 percent area.

In Japan and Germany, consumers have maintained a strong level of spending through 1991, although their continued spending into 1992 is questionable considering that, according to The

Dun & Bradstreet Corporation, both economies are decelerating. In Japan, the growth in gross national product (GNP) is expected to fall from

4.5 percent in 1991 to 3.2 percent in 1992; in

Germany, growth in GNP is forecast to fall from 3.0 to 2.0 percent.

It can be aigued that an uneven recovery in the United States and the weakening economic climate in Japan and Germany will delay consumer electronic equipment purchases. However, the consumer is getting help. Monetary policymakers in both the United States and Japan are loosening the reins. Interest rates have fallen to a 20-year low in the United States and are creeping lower in Japan. Stock market activity in both countries also seems to be pointing to better times: The U.S. market is reaching an alltime high, and the Japanese market has stabilized and is moving higher. This factor bodes well for consumer confidence; the demand for consumer electronics could well surpass our expectations. If this happened, our 1992 forecast would err on the conservative side.

Yet the Conference Board, a private business research firm, reported that the level of consumer confidence in the United States continued to deteriorate in September (see

Figure 2). The survey showed that, compared with a month ago, consumers are a good deal less positive in their assessment of prevailing conditions and also somewhat less optimistic in their expectations for the months ahead. The consumer confidence index is well below its level of just prior to the beginning of the

Mideast crisis. Moreover, employment data

Comtnunication

Communication applications accounted for

14.1 percent of the semiconductors shipped in

1990 (see Figure 1). This end market, albeit small, remains a bright spot in tenns of drivers for the semiconductor industry. The largest segment of communications equipment is telecommunications. Because of the weak global economic climate, there has been a slowdown in the ordering patterns of long distance and

©1991 Dataquest Incoipoiated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

Special Edition

Figure 2

Consumer Confidence Index

11

Cuirent

1 1 1 1 1 1 1 1 1 1 1 \ 1 1 i 1

May Jun. Jul. Aug. Sep. Oct. Nov. Dec. Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep.

1990

1991

Source: The Conference Board

cellular companies in the industrialized countries. It appears that companies are delaying purchases of switching equipment at this time and are settling for stripped-down versions of some switching equipment until volumes pick up.

On a positive note, the fastest-growing regional markets for telecommunications equipment are the less-developed countries, and there has been no slowdown in this segment. Many of developing countries are quickly upgrading their antiquated analog systems with digital lines.

Smaller segments of the communications equipment market—^i.e., LANs and personal communication—are also experiencing strong growth. of the semiconductors shipped in 1990 (see

Figure 1). The mil/aero segment will provide little growth for semiconductor demand any time soon. U.S. President George Bush's recent announcement concerning changes in the U.S. government's nuclear strategy is expected to put several programs in immediate jeopardy of losing funding. The rail-mobile MX missile program, Boeing's Short Range Attack Missile, the

U.S. Navy's nuclear-armed Tomahawk cruise missile, and perhaps the B-2 bomber are all expected to suffer when Congress evaluates the defense budget.

Industrial

Industrial applications accounted for 10.2 percent of the semiconductors shipped in 1990

(see Figure 1). The industrial segment is expected to show marginal growth in 1992.

Spending on medical electronics and analytical instruments helped bolster the industrial segment through 1991 and should perform well through 1992. The measure and control electronics segment is in a recession, and Dataquest expects litde increased spending in this segment in 1992.

Automotive

Automotive applications accounted for 4.7 percent of the semiconductors shipped in 1990

(see Figure 1). The automotive segment is expected to improve during the next several months. U.S. domestic auto sales are forecast to increase from the depressed level of 6.0 million in August to 6.5 million in October and

November according to Morgan Stanley, a New

York investment bank. The big problem in the automotive segment is the consumer with his disastrous real disposable income and lack of consumer confidence.

Mil/Aero

Mil/aero applications accounted for 5.2 percent

On a more positive note for the semiconductor industry, there was a 20-year low in auto inventories at the end of model year 1991.

According to industry estimates, in early September the 1992 models had only a 50-day

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

12 Special Edition

supply when 65 days is nonnal. Therefore, a small increase in demand by the consumer is expected to cause auto manufacturers to ramp their production, increasing the demand for automotive electronics.

Dataquest Perspective

The growth in semiconductor demand is forecast to accelerate in 1992. But, in order to achieve the forecast growth rates, the major semiconductor end-use markets—the data processing and consumer segments—^need to begin showing more life soon.

The fourth quarter of 1991 will be pivotal. If the U.S. economy pulls itself out of recession and if Japan and Germany experience only a moderate deceleration of their economies, then our 1992 forecast is very reasonable. In fact,

Dataquest's forecast may be conservative if the major end-use markets perform better than our expectations. •

By Mark FitzGerald

For More Information . . .

On the topics in this issue Semiconductor Group (408) 437-8196

About on-line access On-Line Service (408) 437-8576

About other Dataquest publications Sales (408) 437-8246

About upcoming Dataquest conferences Conferences (408) 437-8245

About your subscription Customer Service (408) 437-8402

Via fax request Fax (408) 437-0292

The content of this report represents our inierpreution and anal^is of information generally available to tbe public or released by responsible individuals In the subjea companies, but is not guaranteed as to accuracy or completeness. It does not contain material provided to us in confidence by our dtents. Individual con:q>anies reported on and analyzed by Dataquest may be clients of this and/or other Dataquest services. This information is not furnished in connection with a sale or offer to sell securities or in connection with the solicitation of an offer to buy securities. This firm and its parent and/or their officers, stockholders, or members of their timtlies may. from time to time, have a long or short position in the securities mentioned and may sell CM* buy such securities.

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0011610

Dataoyest

n n a company of

A i n The Dun &Braclslrcct Corporation

Dataquest

Perspective

Semiconductor

Application Marlcets

Worldwide

Vol. 1, No. 2 November 18, 1991

Market Analysis

Computer Digital Video: A Multimedia Opportunity

Most of the standards issues are setded, and application software is rolling out along with the hardware. This article examines semiconductor opportunities in this developing market.

By Gregory Sheppard Page 2

Rigid Disk Drives: A Case for Integration

The disk drive market is very competitive, with product differentiation and cost control the industry watchwords. The integration of data and control electronics onto as few as one IC should be an attractive opportunity.

By Nicolas Samaras Page 9

T-Carrier Market Offers Mixed-Signal Semiconductor Opportunities

In an era of growing digital network commmunication, T-carrier equipment and line cards offer an attractive application for semiconductors. In particular, merchant mixed-signal ASSPs and ASICs will be in increasing demand.

By Krishna Shankar Page 13

©1991 Dataquest Incorporated / 1290 Eidder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

Semiconductor Application Markets—Worldwide

Market Analysis

Computer Digital

Video: A Multimedia

Opportunity

Is It Really an Opportunity?

Widiin the realm of multimedia computing are emerging, real hardware—and therefore IC;— opportunities for image and sound processing.

For purveyors of hardware, it is a badly needed opportunity to add value in a market driven by commodity economics. The fundamental user benefit is a more natural and effective use of

PCs and workstations. Central to market acceptance of computer digital video and its derivative chip demand are the following two assumptions:

• Hardware and software standards will firm, and interoperable and widely available products will emerge.

• Producers and viewers of multimedia source material (e.g., an interactive mulitmedia training manual) will find this technology easy to use and helpful in their jobs.

Software standards are now available for multimedia extensions to Windows (DOS) and OS/2 for IBM-compatible PCs. IBM Corporation and

Microsoft Corporation have agreed to what is known as the Media Control Interface (MCI) specification and have set in motion 40-plus independent software vendors (ISVs) producing multimedia tools and applications. This effort complements an already robust effort for Apple

Computer Inc. platforms.

Furthermore, at least 30 hardware vendors, including many platform makers, are beginning to buUd a substantial market for digital video add-in cards and embedded functions. Compression is one of the enablers of digital video because of the overwhelming amount of data involved in storing and transmitting it. The Joint

Photographic Experts Group (JPEG) and JPEGcompatible digital video interactive (DVI) is the emerging image data compression standard of choice for most PCs and workstations. The

Motion Picture Experts Group (MPEG) compression standard for motion applications is probably only a year away. To further market acceptance of MPEG, consumer multimedia (principally CD-I) players are moving to the MPEG standard as well.

As for the demand side of the equation, such early adopters as companies that produce ad copy with photographs are already users. According to Dataquest surveys, we can expect a substantial percentage of the corporate, educational, and government users to begin following in the coming two to three years as titles and applications appear.

This article assesses the growing hardware and

IC market for the capture, processing, and compression of digital video images by PCi and workstations.

What Is It?

There will not be a homogeneous use for multimedia computers. Some will be on the desktops of producers of publications, training and educational programs, and sales presentations.

These installations will most likely be a minority of hardware sales but crucial for the market to develop. The bulk of users of multimedia will require only the playback of imageembedded material. A more complete listing of computer digital video uses includes the following:

• Photo processing (advertising, publishing)

• Interactive training

• Interactive education (K-12)

• Interactive presentations (personal sales, point-of-sale)

• Interactive information displays (Kiosks)

• Image processing and analysis (medical, military, resource, inspection)

• Video/image library (e.g., multiple listings)

• Desktop publishing

• Video communications (video phone, e-maii)

• Video entertainment/games

Figure 1 illustrates the hardware dimensions of digital video. It spans the range of such grayscale and color-image capture systems as cameras, video phones, and scanners to output systems such as printers to storage systems and networking. It is important to note that color peripherals require high bandwidths and are candidates for compression ICs as well.

Standards and Alliances

what about market-enabling factors such as operating system support, application program interfaces (APIs), and, ultimately, application

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

Semiconductor Application Markets—^Worldwide

Figure 1

Computer Digital Video System

Videophone Line

Video Camera

DOt

z^

JPEG, Px64

VCR

CD-ROM

Laser Disk Player

JPEG, DVI

Eg^^g^^^^™

MPEG I

^ ^ ^

JPEG

Color Printer

(^^m

JPEG

:7^ a ^ 3

^

MPEG I, MPEG II

JPEG, Px64

Network

^ ^

Color Scanner

Source: IIT, Dataquest (November 1991) software? A tidal wave of support is building with many key players. Three key developments are the introductions of Apple's Quick-

Time multimedia extension to its System 7.0 operating system, Microsoft's multimedia extension 1.0 for Windows 3.0, and IBM's OS/2 2.0 multimedia extensions. A group led by

Microsoft and Tandy has created a baseline specification called Multimedia PC (MFC), which guarantees the user interoperability with other vendors' systems.

The Apple Computer platform sports easily a dozen ISVs producing software editors and presentation creators. IBM and Microsoft have agreed to adopt a multimedia programming interface (MPI) and data specifications that address ISV needs for common APIs and data formats across DOS, OS/2, and Windows. The

MPC group claims that at least 60 multimedia applications are available from 40 companies.

Likewise, IBM claims that at least seven companies are working on ISV toolkits for its products, with availability starting early next year.

Other activities in multimedia include Kaleida, the new Apple/IBM joint venture that is planning product availability within two to three years. The result of this effort could be embedded multimedia on the motherboard. For the educational market, publishers such as McGraw-

Hill are developing interactive subject software.

Early results show that images and the interactive nature of this style of learning yield better retention results than regular computer programs.

Video Editing

One of the key enablers of the computer video market will be the availability and ease of use of economical editing hardware and software for the creation of source material. This aspect is probably what will limit the early proliferation of digital video the most. Video editing systems priced from $10,000 to $60,000 are entering the market, the products of such companies as Avid Technology (Burlington, Massachusetts), Digital F/X (Mountain View, California), and Editing Machines (Washington, D.C.).

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

Semiconductor Application Markets—Worldwide

These systems are PC centered and involve add-in boards, attachable boxes, and programcontrolled VCRs. These systems generally have graphical user interface software interfaces with multiple windows for viewing multiple images.

One of the key functions of these systems is adding time codes (SMPTE) to each video frame along with sound synchronization. The recent MFC and IBM announcements highlight more focus on "layman" editing; we should expea more economical and easy-to-use editing packages throughout 1992.

Compression Is E x p a n d i n g

One reason that digital image handling had not been incorporated into mainstream PCs is that the data bandwidth and storage required had made it economically impractical. A goodquality color photograph can require upward of

20 Mbytes to store it uncompressed. likewise, moving 20 Mbytes to and from a VHS tape or over a LAN, for example, w^ould require an inordinate amount of time. Compound the problem by manipulating 30 frames a second for motion video; then the need for compression becomes a prerequisite.

Compression takes advantage of the fact that much of the information in a picture does not change dramatically over a smaU area and therefore is redundant. For motion images, there is the additional observation that most of the image does not change from frame to frame, and this is redundant as well. Compression algorithms take advantage of these two factors.

To address the need for compression, several standards have been proposed. Some are open, and some are proprietary. The algorithms employed in image compression are inherently

Table 1

Open Compression Standards

Feature

Color Still Image

Motion Video

Real-Time Video

Capture/Playback

Broadcast Motion

Compression Ratio"

Bandwidth"

Standard

JPEG

X

X*

X

to 80:1

*Requires bigber bandwidth

**VaTies with resolution, frame rate, and image complexity

Source: Dataquest (November 1991)

MPEG I

X

X

X to 200:1 to 1.5Mbps

"lossy," meaning that the decompressed image is not exactly the same. Ho^vever, the losses are generally not noticeable to the human eye.

Table 1 presents the various open standards and their features such as maximum compression ratios. The JPEG standard is principally applied to full-color still images while the

MPEG standard applies to full-motion color video and sound. Both standards are open and managed by committees of the International

Standards Oiganization (ISOVIEC. It is important to note that these proposed standards can be implemented in either hardware or software. In fact, both Apple and Microsoft have JVIPEG algorithms built into their multimedia extensions providing an interim alternative—albeit, slow^— for users until the hardware is available.

Another open standard optimized for teleconferencing and video phones is known as H.261

(or Px64) and managed by the International

Telegraph and Telephone Consultative Committee (CCnT). The DVI standard from Intel Corporation involves proprietary coding schemes but has been modified to support the open standards as well. Initially, DVI was asymmetrical; however, with the release of its i750B product on its ActionMedia n board (with IBM),

Intel can support real-time video capture.

In a recent survey, Dataquest determined that the majority of OEMs designing digital video systems are using JPEG compression because of its virtue as an open, multisourceable standard.

The symmetry feature or ability to create source material (real-time and compressed) on the desktop is also an advantage. When finalized next year (planned), the MPEG I wiU of¥er users an ability to compress motion images even further (up to 200:1) while keeping the bandwidth at 1.5 Mbps (CD-ROM and T-1

MPEG n

X

X

X to 100:1 to 10Mbps

Px64

X

X to 2000:1

64Kbps-2Mbps

©1991 Dataquest Incorporated / 1290 Kidder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

Semiconductor Application Markets—^Worldwide rates). Real-time MPEG encoding currently is difficult to implement economically. A variation known as the JVC extension offers less lossy compression with a 50:1 ratio but trades off by generating a 6-Mbps bit stream. MPEG III is targeted at HDTV resolution (2 million pixels) with

60-Mbps bandwidth.

Compression Players: Getting

Compressed

Both large and small players have entered the video compression chip set market. Some offerings are optimized toward one or more of the specs. Some were also the result of projects with particular customers. The leading participants currently include C-Cube Microsystems

(JPEG/MPEG), Intel (DVI/JPEG/MPEG), LSI Logic

(Px64/JPEG/MPEG), SGS-Thomson QPEG), and

IIT 0PEG/MPEG/Px64). Brooktree Corporation,

Cypress Semiconductor Corporation, Motorola

Incorporated, and Texas Instruments Inc. are rumored to be working on MPEG chip sets.

Intel is reportedly working with PictureTel to create a 30-frame-per-second MPEG encoder capable of 1 billion operations per second.

Computer platforms are only one of many markets for these compression ICs. The advent of color and its need for high communications bandwidth is forcing fax machines, scanners, and printers to consider JPEG. The video conferencing and phone markets are of course the home for Px64 ICs. Another very interesting market for compression is home delivery of more channels and higher resolution TV by broadcast (terrestrial or satellite) or cable companies. In this case, a decoder would be needed for TV sets to receive the improvements. Additionally the regional Bell telephone companies in the United States are moving closer to being allowed to deliver TV and TVbased services over regular phone lines in competition with the cable companies. Image compression will be necessary for this as well as a means to deliver video bandwidth to residential areas, which are mosdy not on optical fiber.

The Hardware

Initially, add-in cards for PC^ and workstations will be the principal means of adding digital video to those platforms. Currentiy, different cards perform the functions of image capture and compression. There are also cards for TV tuning, for capturing broadcast or cable signals, and for audio processing. The audio cards can offer CD-quality sound capabilities and can interface to musical instruments through the musical instrument digital interface. Substantial peripheral hardware will be required for most multimedia setups including such storage devices as digitally controlled VCRs, laser disk players, and CD-ROM (including CD-I and CD-

ROM/XA). Ctertainiy rewriteable optical disks will become very important to image storage as well. Other equipment such as digital still or motion cameras and color scanners will be required for image capture.

There is a good chance that by 1995 the image capture and compression functions will appear on the motherboard coupled with the other graphics subsystem functions like the graphics controller, buffer memory, and RAMDAC. The motivation for this migration is the continual search for differentiated products by hardware vendors.

Figure 2 presents a generic, full-featured computer digital video system. Video signal input and output are managed by a set of digitization functions that digitize and restore TV signals like NTSC or PAL. The input signal can either be composite or come with separate luminance and chrominance signals as with S-VHS. Generally, the output is in a digital form of the YUV color space. The pixel stream created by this function is then converted into the RGB color space and corrected (inverse gamma). Then the data can be stored in a pixel frame for the compression function to take over. This series of steps is generally what comprises the capture function so far.

The compression function takes the buffered data and creates pixel blocks out of the rasterized data so that compression functions of discrete cosine transformation iDCT) and Huffman coding (for JPEG/MPEG) can operate. The compressed images can be output in standardized file formats (for example, TIFF and PICTO and then be stored or transmitted. To decompress for display, the above mentioned steps are reversed. The embedded graphics hardware can be shared by employing a video mixer. This arrangement also allows "windowing," which combines resident graphics with video stills or moving images.

Chip Opportunities

As shown in Figure 2, there are several semiconductor opportunities in digital video. Many of them are for application-specific standard products (ASSPs). The data conversion ICs are mixed-signal opportunities while the ASSPs are principally digital CMOS. As market economies develop, certainly many of these functions are subject to integration. Most of the systems

©1991 Dataquest Incoiporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

Semiconductor Application Markets—^Worldwide

Figure 2

Generic Computer Digital Video

Host Bus

Audio

DSP

VCR

Laser Disk

Camera

NTSC

PAL

SECAM

VCR-

Digitization

A to D/D to A

Decode/Encode

Digital Controls

Color Display-*-

Grapfiics

Subsystem

Controller*

Display Buffer*

RAMDAC*

Video Mixer

Pixel * "

Processing

' Color Space

Conversion

' Raster/Block

Conversion

32-Blt*

MPU

Control

Logic

SRAM

PLD/FPGA

-^ 16-24

Bits

EPROM/Flash

Clock

Pixel Buffer

2-8MB

DRAM/VRAM

T

Host MPU

Main Mennory

Hard Disk

Optical Disk

Compression

Decompression

DCT

Huffman

Coding

a

Host Bus

*Can be part of motherboard

**Can be part of compression IC

Source: Dataquest (November 1991) handle <352 x 240 pixel resolution at 30 frames per second. The trend is to push this to 640 x

480 in the coming years. Examples include the following:

• D to A, A to D (video rates, 8-bit)

• NTSC/PAL/SECAM encoder/decoder

• Audio signal processor (l6-bit)

• Video mixer

• RAM-palette DAC (video, RGB)

• Color space converter (YUV/RGB) implemented on integrated compression ICs

• Raster/block conversion ICs, implemented on integrated compression ICs

• JPEG/MPEG/Px64 compression ICs

• Pixel buffer (2MB to 8MB of DRAM/VRAM,

80ns, x8)

• SRAM buffer/scratch pad (1/4MB, <45ns, x8)

• EPROM/Flash control code (1/4MB)

• PLD/FPGA (2K to 3K gates)

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

Semiconductor Application Markets—^Worldwide

Hardware Vendors

Table 2 lists s o m e of t h e a n n o u n c e d suppliers of digital v i d e o hardware. The majority of the offerings c o m e in the form of add-in cards for p o p u l a r b u s e s like NuBus, S Bus (Sun), ISA/

EISA, a n d MCA. Functions performed include image capture, compression, display m a n a g e ment, and TV tuning. We e x p e a platform makers t o join t h e list over t h e coming y e a r as t h e y scramble for product differentiation. As can b e seen, the list is long a n d undoubtedly s o m e shakeouts will occur. It is possible that this market will b e h a v e as the VGA card market did because the v i d e o function could m o v e t o the motherboard.

Forecast: A Scenario

Figure 3 presents a forecast of h o w digital video functionality could pentrate t h e worldw i d e installed base. Penetration should not

Table 2

Digital V i d e o H a r d w a r e V e n d o r s

Company

Aitech

Af>ple Computer

Commodore

CompuAdd

DesignTech

Digital F/X

E-Machines

IBM lEV International

Intel

Magni Systems

Mass Microsystems

NCR

Sonice: Dataquest (November 1991)

Location

Fort 'Worth, Texas

Cujjertino, California

West Chester, Pennsylvania

Austin, Texas

San Jose, California

Mountain View, California

Beaverton, Oregon

White Plains, New York

Salt Lake Qty, Utah

Princeton, New Jersey

Beaverton, Oregon

Sunnyvale, California

Dayton, Ohio

Company

NEC Technologies

New Media Graphics

NewTek

New Video

Philips Consumer

Radius

RasterOps

Rapid Technology

Sun Microsystems

Tandy

True Vision

Video Logic

Zenith Data Systems

Figure 3

C o m p u t e r Digital V i d e o Opportunity*

Thousands of Units

5000-

Units Worldwide

4500

Average Selling Price

4000

3500

3000

2500-

2000-

1500*

1000

500

0

1990 1991 1992

* Add-In boards and embedded functions

Source: Dataquest (November 1991)

1993 1994

Location

Tokyo, Japan

BiUerca, Massachusetts

Topeka, Kansas

Venice, California

Eindoven, Netherlands

San Jose, California

Santa Clara, California

Amherst, New York

Mountain View, C:alifomia

Ft. Worth, Texas

Indianapolis, Indiana

Cambridge, Massachusetts

Glenview, Illinois

V ^

I

I

1995

Doiiars

5000

- 4500

- 4 0 0 0

- 3 5 0 0

- 3000

2500

- 2000

_ 1500

1000

I- 500

0

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

s

Semiconductor Application Markets—^Worldwide

Figure 4

Computer Digital Video Semiconductor Opportunity

SRAM 3%

EPROM/Flash 2%

PLD/FPGA 1 %

Buffer/

Interface/

Clock 1 %

SRAM 5%

EPROM/Flash 3%

PLD/FPGA 3%

Buffer/

Interface/

Clock 2%

1 9 9 1 = $ 1 2 . 1 Million 1 9 9 5 = $ 1 , 2 1 6 . 4 Million

Source: Dataquest (Novembei 1991) exceed 1 percent before 1994. As of today, it should take another year to setde the standards and generate a critical mass of application software and titles. It will then take another one to two years for the early adopters and producers

(authors) of source material to move into action. This sets the stage for the market to take off by 1995. Tlie worldwide market for boards (1 set per platform) could reach 4.8 million units by 1995. Most of the production of these boards will initially be in the United

States and some in Europe; then production will migrate to Japan, Taiwan, and South Korea as the market matures.

Based on the above scenario, the semiconductor total available market presented by add-in boards and motherboard functions will be a billion dollar world market by 1995. About

55 percent of the opportunity is in VRAMs,

DRAMs, and, to a lesser degree, Flash memory and SRAMs (with bit price assumptions) as shown in Figure 4. Another 20 percent of the opportunity is for ASSPs for NTSC/PAL/SECAM digitization and control. Another 20 percent of the opportunity is in the compression chip ICs.

The remaining 5 percent is for control logic, dock ICs, and bus transceiver/buffers. The

RAMDAC, graphic controller, SCSI ICs, and CD-

ROM controller opportunities are not counted in this analysis but will be positively impacted by digital video growth.

Dataquest Perspective

Clearly, any oudook on computer-based multimedia and digital video is governed by the assumptions regarding the development of a critical mass of users wanting this functionality and suppliers agreeing to provide useful, multisourced, and economical systems. It appears that these factors are coming tpgether and that a substantial market is emeiging. As we have said, it will take the market a few years to develop, as producers and users of image information go down the learning curve and rediscover the axiom of a picture saying a thousand words.

The opportunity to semiconductor companies as modeled from the above scenario is prindpaUy in two areas. One is certainly for commodity products like DRAM, VRAM, SRAM, EPROM/

Flash, PLD logic, and interface logic. The other is for complex ASSPs for the digitization of TV signals and compression. This is an opportunity to produce value-added products. It is very likely that digital video applications will become price sensitive quickly in order to develop a mass market. Dataquest believes that the challenge will once again be placed on the semiconduaor industry to drive out costs through superior design and manufacturing.

By Gregory Sheppard

©1991 Dataquest Incorporated / 1290 Kidder Paric Drive, San Jose, CA. 95131-2398 / (408) 437-«000 / Fax (408) 437-0292

0012320

Semiconductor Application Markets—^Worldwide

Rigid Disk Drives: A

Case for Integratifm

The State of the Rigid Disk Drive

Industiy

This has not been a good year for personal computers. While PC sales were flat, the rigid disk drive (KDD) industry was gearing up in anticipation of a rapidly growing portable PC market only to find itself with an oversupply of rigid disk drives, vendor consolidation, and depressed drive average selling prices (ASPs).

However, as tradition has it, down markets allow equipment manufacturers to focus heavily on new product development. So in the past year w e have seen a myriad of new produa announcements that include gigabyte-class

3.5-inch RDDs, 2.5-inch drives that offer 130MB, and the newest 1.8-inch RDDs at 42MB. There is no sign of slowing down the pace of iimovation, and the race to produce faster, smaller, and less expensive drives at ever-increasing capacities is picking up.

According to Dataquest's Computer Storage group, the 1991 overall RDD increase in unit shipments over 1990 was a modest 13.4 percent

(see Table 1). Dataquest expects a total unit volume increase of 19.6 percent in 1992. In particular, the 2.5-inch rigid disk drives should outpace the pack with a 115 percent increase in 1992, followed by a 23 percent increase for the 3.5-inch form factor. The 3.5-inch RDDs will be the preferred drive for desktop PCs, for workstations, and generally for systems that need higher capacities and data throughput. The

2.5-inch drives are the form faaor of choice for notebook and a portion of the pen-based PCs.

The 1.8-inch, a newcomer, should reach a volume perhaps as high as 1 million units in 1992.

As expected the more mature form factors of

5.25, 8, and 14 inches will be in decline. The rigid disk drive ASPs should remain relatively stable (see Table 2).

The worldwide RDD market will grow at a modest 5 percent (revenue) rate in 1992, reflecting heavy pricing pressures (see Table 3).

Dataquest's worldwide factory revenue for rigid disk drives shows the 3.5-inch drive revenue growing at a compound annual growth rate

(CAGR) of 17.8 percent, from about $6.2 billion in 1991 to about $11.9 billion in 1995. The

2.5-inch form factor revenue will grow at a

CAGR of almost 43 percent over the same period to approximately $3 billion by 1995.

Table 1

Worldwide Rigid Disk Drive Unit Projections (in Thousands)

2.5-Inch

3.5-Inch

5.25-Inch

1990

714

21,742

6,877

8-10-Inch

14-Inch

377

237

Total

29,947

Souice: Dataquest (November 1991)

1991

2,888

23,727

6,750

417

167

33,948

Table 2

Rigid Disk Drive Factory ASPs (in Dollars)

1992

6,192

29,262

4,684

318

134

40,591

1993

9,960

31,516

3,734

282

95

45.587

1994

12,968

35,304

4,384

127

44

52,828

1995

12,836

34,955

6,396

85

33

54,305

CAGR (%)

1991-1995

45.2

10.2

-1.3

-32.8

-33.3

12.5

2.5-Inch

3.5-Inch

5.25-Inch

8-10-Inch

14-Inch

1990

263

258

474

7,580

21,934

Source: Dataquest (November 1991)

1991

247

260

535

9,485

24,523

1992

244

275

777

9,705

23,643

1993

280

297

926

9,455

23,184

1994

284

338

866

8,230

20,892

1995

231

339

915

6,794

19,697

CAGR (%)

1991-1995

-1.6

6.9

14.3

-8.0

-5.3

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

10

Semiconductor Application Markets—^Worldwide

Table 3

Worldwide Rigid Disk Drives Factory Revenue (Millions of Dollars)

1990

2.5-Inch

3.5-Inch

5.25-Inch

187

5,612

3,262

8-10-Inch

14-Inch

2,855

5,207

Total

17,123

Source: Dataqaest (November 1991)

1991

714

6,162

3,613

3,951

4,098

18,537

1992

1,513

8,047

3,641

3,085

3,178

19,464

1993

2,786

9,373

3,457

2,664

2,198

20,478

1994

3,678

11,915

3,798

1,045

928

21,363

Table 4

Worldwide Rigid Disk Drive Semiconductor Market (Millions of Dollars)

1995

2,970

11,864

5,851

577

650

21,913

CAGR (%)

1991-1995

42.8

17.8

12.8

-38.2

-36.9

4.3

2.5-Inch

3.5-Inch

5.25-Inch

8-10-Inch

14-Inch

Total

1990

34

990

323

113

83

1,543

Source: Dataquest (November 1991)

1991

124

994

301

131

61

1,612

1992

248

1,165

217

105

52

1,787

Table 4 represents the worldwide semiconductor market for RDDs. The 2.5-inch drives will enjoy the highest growth as far as semiconduaor content is concerned. This is primarily due to the need for substantial integration, dictated by limited space, and power availability.

RDD Controllers—^How Many ICs

Are Needed?

The movement toward an ever-shrinking form factor has created a need for higher levels of integration. Today's 3-5-inch RDD controller electronics is designed around one or two dozen ICs. The trend is for 2.5-inch drives to use just 10 to 12 ICs, with 1.8-inch drives now using six very large scale integration devices.

1993

369

1,204

190

98

38

1,899

Most likely, the market for RDDs tvill expand as both ASPs and form factors shrink. New opportunities ought to proinde for RDD market grotvth.

The day of the single RDD controller IC is approaching, along with its significant implications for semiconductor vendors. When the chip

1994

493

1,362

257

46

19

2,178

1995

512

1,403

431

33

15

2,394

CAGR (%)

1991-1995

42.5

9.0

9.4

-29.4

-30.0

10.4 count drops from a dozen ICs to two or three, it is natural to assume that revenue loss for semiconductor vendors will follow. This assumption is partially true; the vendor base serving the RDD market will shrink. To be a player in the future, a semiconductor company will need to understand all aspects of the RDD design, including mechanical. The only other option would be to form a very close relationship with the drive manufacturer acting perhaps as a foundry. The price of single chip drive controller will not be $42 (the cost of the current average semiconductor content of a small drive). Most likely it will be somewhere between $8 and $25, allowing for different controller versions. That should give birth to the

$100 drive with a low capacity of about 40MB to 100MB. The market is elastic and thus may provide for unprecedented growth into areas that never used RDDs. Thus the $5 to $25 single controller IC may become the "enabler" for new market growth.

T h e S e m i c o n d u c t o r Makeup o f a n RDD

The semiconductor content of a small rigid disk drive (5.25 inches or less) costs approximately

$42 in 1991 (see Figure 1). The cost breakdown

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

Semiconductor Application Markets—^Worldwide 11

Figure 1

Estimated Rigid Disk Drive Semiconductor Content (Cost per Drive)

1995 1991

Analog

$15

ASIC/ASSP

$15

Micro

$6

Memory

$6

$42

Source: Dataquest (November 1991)

- - - ^

Preamp

$2 - $4

Mixed-Signal, Single-Chip

RDD Controller

$5 - $25

Memory

$1 - $12

$8 - $41 is as follows: analog ($14.91), micro ($5-99); memory ($6.33), ASIC/ASSP ($15.17), and standard logic ($0.34). By 1995 almost all the functions may be integrated in a mixed-signal, single-chip controller. Then the drive electronics will most likely be reduced to three ICs: the controller, the buffer memory, and the head preamp. Under some conditions a small amount of buffer memory may be integrated into a low-end controller, resulting in a two-chip solution. The head preamp IC is and most likely will remain separate, a condition dictated by the necessary proximity to the disk drive's head(s) and needed noise isolation from the controller section of the system. drive makers are trying to address this problem by offering both logical and innovative solutions.

Disk arrays fall in the logical category. For example, four drives equipped with a SCSI interface can be connected together in parallel and serve a single-host computer. Data then can be striped across the four drives and provide for a substantial performance increase over single-drive systems.

Process Technology

Mixed-signal is a key technology. Semiconductor vendors with such capability and experience will do very well supplying components to the

RDD market.

I m p r o v i n g RDD P e r f o r m a a c e

Disk Arrays

Current-generation PCs are about 40 times faster than the original XT. By comparison, currentgeneration RDDs are only 10 times faster than the equivalent original XT drives. AS C3'U performance increases, so does the percentage of time that programs wait for disk I/O. Disk

As vendors attempt to remain competitive and offer ever-increasing capacities in ever-shrinking form factors, RDDs may for the first time be displacing semiconductor memories in some applications....

Another scheme called "mirroring" uses two drives per system to provide for redundancy.

TTie same data are written to both drives; if one drive fails, the data are still available on the other. Beyond that, the two disks can be used to improve the overall system performance by sharing the data storage and retrieval task.

In the innovative category, RDD companies are experimenting with the use of parallel data

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

12

Semiconduaor Application Markets—Worldwide

channel usage on 3-5-inch drives. Here 9 disk drive heads are synchronized to offer a "parallel" read/write configuration. Spindle drive synchronization of up to 10 drives is also believed possible.

The Quest for Higher Densities

Attempts to increase disk drive capacity include a move to offer Constant Density Recording

(CDR) or a subset of CDR, Zoned Density

Recording (2DR). CDR/ZDR increases the amount of data stored on each track by keeping a constant bit density on every track through an increased data rate for tracks of increasing radius. CDR can increase a drive's density by nearly 50 percent. ZDR is simpler to implement and offers density increases near those of CDR. However, implementing CDR or

ZDR complicates the servo system and the read/write channel electronics.

Buffer/Cache Sizes

The preferred size of the buffer/cache memory in the small form factor RDDs is 32KB (a single

256Kb SRAM). Buffer sizes should increase to between 128KB and 512KB by 1995 (a single

1MB/4MB device that may be either SRAM or

DRAM). Only the higher-performance drives seem to implement data caching schemes. A small amount of memory, perhaps 8KB, may be integrated with low-end disk drive controllers.

High-end systems may keep the memory separate to provide flexibility in system design.

The Drive to Dominate: Microcontroller versus DSP

The Motorola 68HC11 and Intel 8051 are beginning to be replaced by l6-bit microcontrollers in RDD control functions. Digital signal processors (DSPs) such as the Texas Instruments Inc.

TMS320 or high-end l6-bit microcontrollers that incorporate DSP functions (multiplieraccumulator) such as the National Semiconductor Corporation HPC-Plus are increasingly being used. Tiiis trend is seen as necessary to achieve higher capacities in ever-shrinking drives.

Interface Preferences

So far the RDD interface has remained a mixture of IDE and SCSI or SCSI-2. The RDD vendors appear to be flexible and most are offering both interfaces as they wait for the marketplace to point the direction. In the long run it appears that SCSI may dominate as the interface of choice, because it offers the most flexibility and highest data throughput rates.

Forces Driving RDD D e v e l o p m e n t

Downsizing

Smaller form faaor rigid disk drives fit into notebook and pen-based PC^s as well as other equipment where space is at a premium and where semiconductor memories are typically found. Oddly enough, this is perhaps the most significant technological development for RDDs.

A 1.25-inch drive creates immense opportunities for market growth.

Lower Power Consumption

Lower power consumption is necessary if RDDs are to be used in future (ever-shrinking) and power-stingy portable PCs. The 2.5-inch and

1.8-inch RDDs consume approximately 2W in the active mode and about 0.4W in standby;

1.8-inch drives offer a 0.015W sleep mode. Both

2.5-inch and 1.8-inch drives operate from a single 5V supply; 3V versions should be introduced shordy. Some 3.5-inch drives still require a 12V supply in addition to the 5V.

Cost Reduction

Beyond the obvious (widespread usage of

RDDs in portable PCs), lower costs are a necessary condition if new markets are to be created. Along with size and power consumption reductions, lower ASPs will enable penetration into new applications where mass storage was not used at all, such as television sets, YCRs, and automobiles.

Dataquest Perspective

Although it is masked by the current turmoil in the PC market (inevitable consolidation of RDD manufacturers and the myriad of new product announcements), something interesting is taking place within the rigid disk drive market. As vendors attempt to remain competitive and offer ever-increasing capacities in ever-shrinking form factors, RDDs may for the first time be displacing semiconductor memories in some applications—an interesting turn of events.

RDD vendors would differentiate their products hosed on performance or cost.

At a future 1.25-inch or 1-inch form factor,

RDDs will occupy the same board space as a couple of IC packages while offering megabytes

©1991 Dataquest Incorporated / 1290 Bidder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

Semiconductor Application Markets—^Worldwide

13

of nonvolatile storage. It is true that the information stored cannot be accessed as easily as with semiconductor memories, but this is acceptable for many applications. Laser printer fonts are an example. There are other applications where serial data storage is a natural fit:

TV and VCR video buffers, for example.

Does this mean that RDDs will steal significant market share from semiconductor memories? We do not believe that this will happen. Most likely, the market for RDDs will expand as both

ASPs and form factors shrink. New opportunities ought to provide for RDD market growth.

Will the semiconductor market for RDDs essentially dwindle to nothing? Not likely. As we move toward the single-IC RDD controller the number of suppliers may drop to less than a dozen. In a future scenario similar to today's

DRAM market, the "drive controller" may become a high-volume commodity item. For the suppliers (semiconductor companies) that remain, the game may prove very profitable.

The future rigid disk drive market may be 2 to 10 times its current size and far less dependent on the personal computer market.

Alternately a variety of single-chip drive controllers offering a spectrum of capabilities may become the norm. RDD vendors would differentiate their products based on performance or cost. In either case the single-chip controller will enable the RDD market to expand and penetrate new applications, something not achievable before because of space or cost constraints.

The future rigid disk drive market may be 2 to

10 times its current size and far less dependent on the personal computer market. •

By Nicolas Samaras

T-Carrier Equipment

Market Offers Mixed-

Signal Sefnicanductor

OppoftuniUes

T-Carrier Overview

T-l protocol is a point-to-point digital communications facility that can cany 24 voice-frequency

(VF) channels over a single line. Typical T-l equipment users include Fortune 2500 companies, universities, medical facilities, financial institutions, local telephone companies, interexchange carriers, and government agencies.

The North American T-l standard allows the multiplexing of 24 VF chaimels i(A Kbps) over a single line at a combined rate of 1.544 Mbps.

T-l service can be provided by span lines, satellite, digital microwave, coaxial cable, and fiber optics. Tlie major trends in the North American

T-l market include the following:

• Significant T-l equipment price erosion due to functional very large scale integration (VLSI) and vigorous competition

• Incorporation of encryption and packet switching features

• Enhanced network management and control features

It should be noted that an increasing amount of carrier voice transmission is now based on higher-capacity T-2 multiplexers (96 VF channels per line) as well as fiber-based T-3 multiplexers

(672 VF channels per fiber) and broadband fiber-based SONET multiplexers. The emergence of extensive fiber-optic communication networks, together with the evolution of Integrated

Services Digital Network (ISDN) standards over the next decade, will accelerate the trend toward SONET-based fiber-optic multiplexers.

T-Carrier Semiconductor Content

Dataquest observes the emergence of highly iategrated, mixed-signal T-l/T-3 applicationspecific standard product (ASSP) chip sets that address the digital and analog functions

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

14 Semicondud:or Application Markets—^Worldwide

Figure 1

Estimated North American T-Carrier Semiconductor Market (Millions of Dollars)

Millions of Dollars

2QD'

I I Premise T-Carrier

180-

^ ^ Public Networl< T-Carrier

160-

140-

120-

100

80

60

40-

20-i

0

59.4

63.5

1987 1989

Source: Dataquest (November 1991)

127.5

65.0

1991

139.0

69.0

1993

142,8

6 9 . 0

1995 required to implement the transmission portion of asynchronous 1.544 Mbps T-carrier line card.

Earlier-generation T-carrier cards used separate bipolar analog line interface chips and digital pulse-code modulation multiplexing chips in order to interface to the T-carrier pipeline emeiging from the premise. Such discrete eariy implementations of T-1 using bipolar analog interface chips, phase-lock loops, and digital

MOS multiplex logic suffered from voltage drift and temperature drift problems that affected the performance of the T-1 line. The emergence of mixed-signal design techniques based on CMOS and BiCMOS process technologies has given impetus to the single-chip integration efforts for

T-carrier line card applications.

Many telecom semiconductor companies offer T-carrier ASIC cores in addition to ASSP products.

T-carrier transceiver ASSPs support various framing standards such as super frame, extended super frame, SLC-96 (AT&T digital loop carrier standard), T-1 data multiplexer (TIDM), and

ISDN primary rate applications. Performance monitoring is supported through features such as on-chip error detection and correction algorithms and data-link controllers.

In addition to the T-carrier ASSP transceiver chip set, typical T-carrier line-cards will contain an 8- or l6-bit CMOS microcontroller with a parallel bus interface to the transceiver. The microcontroller allows distributed control and programming of the T-carrier operation. Lowercost T-carrier line-cards, alternatively, have limited hard-wired program instruction sets directly embedded on the ASSP. In such cases, the ASSP transceiver bus can be directly accessed to relay execute instructions. In addition to the transceiver ASSP and the microcontroller, the T-carrier line-card may have some buffer memory (l6 to 64K SRAM), EPROMS

( 2 5 6 K )

for program storage, and some programmable logic devices such as programmable logic arrays or field-programmable gate arrays.

Many telecom semiconductor companies offer

T-carrier ASIC cores in addition to ASSP products. Such core ASIC libraries include

T-carrier line performance monitors, transceiver cores, and line-interface circuits. T-carrier equipment companies use these ASIC cores to customize their system with regard to features such as backplane interface, network monitoring and testing, and digital cross-connect features.

T-Carrier Semiconductor Market and Applications

Figure 1 shows Dataquest's market estimates for the history and future growth of the T-carrierrelated semiconductor device market. The total

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

Semiconductor Application Markets—^Worldwide 15

T-camer semiconductor market is expected to grow at a sedate compound annual growth rate

(C:AGR)

of 3.8 percent from $118 million in

1990 to $142 million by 1995. The market is essentially composed of mixed-signal ASIC and

ASSP devices. The merchant-supplied portion of this market should grow substantially faster than the captive portion. T-carrier semiconductors are used in both premise and public (central office and network) applications. trunk lines, are a high-growth factor in the public telecom T-carrier market.

In the Juture, even the smallest business office locations should be able to afford a T-1 pipe that can carry voice, data, image, video, and graphics directly to the interexchange carrier.

Semiconductor Forecast Assumptions

Dataquest assumed a tenfold multiplication factor to arrive at the combined T-l/T-3 node units based on the annual end-to-end T-l/T-3 installed lines. Such a multiplication factor accounts for the multiple number of T-canier line-cards in a typical end-to-end public network trunk line with multiple repeater and digital cross-cormea statiorw. Based on such an approach, Dataquest estimates that the number of public network T-carrier nodes will increase at a CA.GR of l6.1 percent from 0.777 million nodes in 1990 to 1.64 million nodes by 1995.

The semiconductor content per public T-carrier node is assumed to change from $70 per node in 1990 to $45 per node in 1990 by 1995.

For the premise portion of the T-carrier market,

Dataquest assumed an I/O ratio of 12 percent for the semiconductor content of premise

T-carrier systems. Dataquest believes that the premise T-carrier equipment market, which includes chaimel service units (CSUs), data service units (DSUs), and embedded T-1 line cards in PBX systems, will grow at a CAGR of 1.5 percent from $530 miUion in 1990 to $571 million by 1995. We have applied a 12 percent semiconductor I/O ratio to the premise T-carrier equipment forecast to arrive at the premise

T-carrier semiconductor market.

In 1991, the total T-carrier semiconductor market was almost equally divided between public and premise telecom applications. Dataquest expects the public telecom portion of T-carrier semiconductor applications to grow at a CAGR of 6.3 f>ercent from $63 million in 1991 to

$74 million by 1995. In contrast, the premise portion of the T-carrier semiconductor market is expected to be relatively flat, it will only grow at a CAGR of 1.5 percent from $65 million in

1991 to $69 million in 1995.

Digital cross-cormect systems that are used in public networks to flexibly configure, multiplex, and monitor fiber-optic SONET data streams, together with lower bandwidth T-1 and T-3

The premise T-carrier market has become relatively mature because of the high penetration of

T-1 lines into the premise telecom market.

Point-to-point trunk T-1 trunk lines and private network T-1 lines are facing increasing competition from broadband fiber-based SONET private network data service providers. Only the network access segment of the premise T-carrier market will exhibit growth between 1991 and

1995 because of the increasing affordability of

T-1 line-cards and their incorporation in lowend PBX systems. The new market of fractional

T-1, which allows for subrate usage and sharing of a trunk T-1 line, is expected to proliferate the applications of T-1 network access linecards in small office premises. In the future, even the smallest business office locations should be able to afford a T-1 pipe that can carry voice, data, image, video, and graphics directly to the interexchange carrier.

Dataquest Perspective

The past few years have seen a proliferation of focused semiconductor companies that address the telecom ASSP and ASIC market. Some of the key T-carrier semiconductor vendors include

AT&T, Base2, Crystal Semiconductor, Dallas

Semiconductor, Level One, Mitel, and VLSI

Technology. The deregulation of the longdistance network business, the emergence of broadband network standards, coupled with the availability of mijred-signal IC design tools and

VLSI process technology, has spurred the emergence of a merchant T-carrier semiconductor market that challenges the proprietary in-house expertize of vertically integrated telecom companies such as Alcatel, AT&T, Fujitsu, Northern

Telecom, NIT, and Siemens. Dataquest believes that broad-range telecom ASSP suppliers that cater to the emerging fiber-optic SONET networking market as well as the more mature

T-carrier market will be well positioned to exploit the hybrid nature of future long-distance networking solutions. •

By Krishna Shankar

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

16 Semiconductor Application Markets—Worldwide

In Future Issues

The following topics will be featured in future issues of Semiconductor Application Markets

Worldwide Dataquest Perspective:

• Opportunities in pen-based computers

• Opportunities in personal communications networks (PCNs)

For More Information . . .

On the topics in this issue Semiconduaor Applications Market Worldwide (408) 437-8261

About on-line access On-line Service (408) 437-8576

About other Dataquest publications Sales (408) 437-8246

About upcoming Dataquest conferences Conferences (408) 437-8245

About your subscription Customer Service (408) 437-8402

Via fax request Fax (408) 437-0292

The content of this report represents our interpretation and analysis of infbrmation generally available to die public or released by responsible individuals in the subject con^anies, but is not guaranteed as ^ accuracy or comiMeteness. It does not contain material provided to us in confidence by our cUente. Individual cc»npanies reported on and analyzed by

I>ataquest may be clients of this and/or other Dataquest services. This information is not furnished in connecticm with a sale or offer to sell securities or in connection with the solicitation of an offer to buy securities. This firm and its parent and/or their officers, stockholders, or members of their £unilies may, from time to time, have a long or short position in the securities mentioned and may sell or buy suc^ securities,

©1991 Dataquest Incorporated / 1290 RicJder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012320

DataQuest

n n 3u>m|»nyor

mUm ThcDun&BradsticctCorporation

Dataquest

Perspective

Semiconductor

Application {Markets

Worldwide

Vol. 1, No. 3 December 16, 1991

Market Annlvsis

LAN/FDDI Applications: Excellent Chip Set Opportunity

Desktop networking will continue to be a robust opportunity for chif>-set suppliers during the next several years. This article examines the worldwide market for LAN/FDDI chip sets as driven by various applications.

By Krishna Shankar Page 2

Automotive Applications: More Controls Offset Vicious Economics

The economics of 1991 was rough on the global automotive industry. However, the future for electronics and semiconductor suppliers to this industry is bright because of fastgrowing applications penetrating the bulk of the world's vehicles.

By Gregory Sheppard Page 5

Inquiry Summary

Semiconduct€tr Application Markets Inquiry Highlights

Dataquest's Semiconductor Application Markets inquiry summary is designed to inform our clients of commonly asked questions and Dataquest's respective answers. No confidential information provided by our clients is included in this material. The information contained in this publication is believed to be reliable, but it cannot be guaranteed to be correct or complete.

• Which are the leading semicondurtor suppliers to the global communications equipment marketi' Page 10

• What is the size of the North American point-of-sale system market' Page 11

• What are the leading application candidates for 32-bit control' Page 11

©1991 Dataquest Incoiporated / 1290 Bidder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012434

Semiconductor Application Markets Worldwide

Market Analysis

LAN/FDDI AppUciOians:

Excellent Chip Set

Opportunity

Worldwide LAN/FDDI Chip Set

Consumption Market Forecast

We estimate that the worldwide LAN/FDDI chip set market will grow at a meteoric compound annual growth rate (CIAGR) of 27 percent from

$350 million in 1990 to $1.15 billion by 1995.

Desktop connectivity continues to be the driving force behind the need to provide shared access to data and devices indefjendently of geographic and time boundaries. Applications to date that have impacted desktop connectivity include electronic mail, word processing, spreadsheets, desktop publishing, databases, and voice.

Dataquest expects networking applications for the 1990s to be multimedia in nature, including voice, data, graphics, still image, and full-motion video connectivity. High-speed (bandwidth), fault-tolerant, distributed-intelligence, seamless networks will be needed to cope with the increasingly demanding requirements of graphics and video comanunication. This article examines the IAN and fiber-distributed data interface

(FDDI) application chip set market trends.

The North American consumption market accounts for almost 70 percent of the current worldwide LAN/FDDI chip set market because of the region's leading position in networking equipment production. However, the North

American LAN/FDDI chip set consumption market is expected to decline to 50 percent of the

$1.15 billion market by 1995 as Asia/Pacific-Rest of Worid (ROW), European, and Japanese LAN equipment companies emeige to exploit the high-growth global networking market. Many

North American semiconductor companies will increase their shipments of low-end LAN chip sets to the Asia/Pacific region for incorporation in low-price, high-volume NIC add-in card applications.

Figure 1 shows the growth of different segments of the LAN/FDDI chip set market between 1990 and 1995. Ethernet chip sets accounted for the dominant portion (57 percent)

Figure 1

Estimated Worldwide LAN/FDDI Chip Set Maritet, by Segment

(Millions of DoUars)

Millions of Dollars

500

5 0 0 -

400^

3 0 0 -

900 —

1 0 0 -

0

^ ^ ^ 1 300

^ ^ ^ ^ E ^ / ' / ' / ' i ^ 2bO

200

^ ^ ^ H 75

1990

• Ethernet E3 Token-Ring

1

^ ^ ^ J y ^ ^ ^ ^ y y ^ ^ ^ ^ 100

mbma^

1995

^ FDDI/CDDI S Others

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012434

Semiconductor Application Markets Worldwide of the $350 million 1990 market. The affordability and increasing m o m e n t u m of unshielded twisted-pair (UTP) media, together with the early adoption of ethernet by network leaders such as Digital Equipment Corporation, h a s led t o ethernet's position as t h e largest LAN chip set market. Token-ring chip set consumption accounted for 21 percent of the 1990 market.

The LAN chip set market is segmenting into low-end, midrange, and high-end segments that address different price/performance points within end-user markets.

The token-ring LAN market has traditionally b e e n driven by IBM Corporation. The tokenring LAN market recently acquired significant m o m e n t u m because of t h e joint armouncement of 16-Mbps Token-Ring products for UTP media by IBM and SynOptics Communications Inc.

These p r o d u a s apparentiy successfully address traditional customer concerns about network signal distortion (jitter) and crosstalk problems that can occur with UTP-implemented Token-Ring solutions.

Dataquest believes that the worldwide ethernet

LAN chip set market will g r o w at a CAGR of

20 percent to $500 million by 1995. The emerg e n c e of UTP Token-Ring protocol is expected t o spur growth of the token-ring chip set market at a higher CAGR of 32 percent to

$300 million by 1995. Higher average selling prices for token-ring chip sets are also partially responsible for the higher token-ring chip set market CAGR, compared with t h e ethernet chip set market CAGR.

Dataquest believes that the FDDI/CDDI chip set market will grow rapidly at a CAGR of

59 percent from only $25 million in 1990 t o

$250 million by 1995. The e m e r g e n c e of global, well-defined FDDI standards, together with a low-cost copper-wire media implementation scheme, is expected to spur the adoption of high-bandwidth-oriented (100 M b p s a n d a b o v e ) desktop connectivity schemes for graphics a n d desktop video conMiunlcations applications.

Table 1 compares ethernet a n d FDDI LAN features. The tenfold iacrease in b a n d w i d t h from

10 Mbps to 100 Mbps, together with vastly increased network size, is expected to favor t h e ultimate extension of FDDI internetworking b a c k b o n e s to the desktop.

LAN Chip Set Market Trends

The LAN chip set market is segmenting into several distinct market segments, b a s e d o n applications and desired features.

Low-End Segment

The low-end segment includes t h e low-price, high-volume lOBASE-T LAN market that is increasingly being used as the mainstream office w o r k - g r o u p networking solution. T h e traditional network interface card is being s h r u n k t o t w o or even single-chip solutions that a r e migrating to the motherboard. Such core LAN chip sets are u s e d in high-end desktop PCs, portable

PCs, a n d entry-level workstations. Dataquest believes that high-end PC manufacturers will increasingly adopt built-in networking capability o n the motherboard in an attempt t o differentiate themselves in a commodity market.

Such high-volume, low-end LAN chip sets incorporate t h e Ethernet CSMA/CD protocol core h o o k e d to an 8-bit or l6-bit host data bus. A limited amount of on-board cache buffer m e m ory supports serial communications rates of

16 Mbps to 20 Mbps, together with capabilities for automatic retransmission without host processor intervention. Some LAN chip set vendors have successfully integrated t h e LAN

Table 1

FDDI v e r s u s Ethernet

Feature

Maximum Physical Bandwidth

Maximum Packet Size

Access Method

Topology

Maximum Length of Network

Maximum Length between Nodes

Source: Dataquest (December 1991)

100 Mbps

4,500 bytes

Token-passing

Dual ring/star/tree

100km

2km

FDDI

Ethernet

10 Mbps

1,500 bytes

CSMA/CD

Bus/star

2.5km

500m

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012434

Semiconductor Application Markets Worldwide controller function and the serial transceiver functions on a single chip.

Midrange Segment

Applications for the midrange LAN chip sets include midrange workstations, diskless workstations, and local intelligent hubs that provide limited support for mixed protocol environments. Chip sets in this segment offer users some programmability and customization options for product differentiation. Recently, mixed-protocol chip sets that claim to offer token-ring and ethernet support have emet^ed to address this market.

High-End Segment

Applications for high-end LAN chip sets include distributed enterprisewide intelligent hubs, bridges, intelligent routers, high-performance servers, and high-performance workstations in distributed computing applications. Such highperformance LAN/FDDI backbone environments always tend to support several LAN protocols.

High-end LAN chip set vendors are beginning to offer shrink-wrapped silicon and software solutions that are optimized to operate in a mixed-protocol environment.

Figure 2 shows a functional schematic block diagram of general partitioning trends for a high-end LAN chip set. In addition to the LAN controller and serial interface blocks, such chip

Figure 2

LAN Chip Set Implementation

sets are bundled together with RISC coprocessors and generous amounts of buffer SRAM or

DRAM memory in order to allow LAN equipment vendors to offer feature-rich, highperformance, customized solutions. These chip sets support multiple industry-standard network operating systems.

High-end LAN controllers will offer features such as buffer memory management, mixed protocol network performance monitoring and control, remote boot-up, 32-bit direct host bus interface and memory access, and external address filter capability. The LAN serial interface block will offer features such as programmable protocol selection, on-chip data/clock recovery, loopback self-diagnostics capability, and built-in lOBASE-T transceiver.

FDDI/CDDI Chip Set Market

Trends

The deterministic, dual counterrotating nature of

FDDI LANs, together with their high-bandwidth and enterprisewide inter-LAN capabilities, is considered to be very desirable in fault-tolerant, real-time, large-file transfer applications. The three key applications for FDDI/CDDI chip sets are as follows:

• Backbone LAN supporting LAN interconnections of departmental and work group LANs.

• Back-end networking supporting host-to-host and host-to-I/O connections.

Host

Communications

Processor

LAN Controller

Buffer Memory

LAN Serial

Interface

Media

Source: Dataquest (December 1991)

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012434

Semiconductor Application Markets Worldwide

• Front-end LANs supporting technical and graphics server-workstation graphics-intensive distributed computing applications.

FDDI equipment vendors are exploring several lower-cost copper-media alternatives to the traditional fiber-media solution. These vendors include Cabletron Systems, Chipcom Corporation, Digital, and SynOptics. Lower-cost FDDI

NIC produas will allow for rapid penetration of

FDDI from inter-LAN backbones to the highvolume, desktop IAN market. Dataquest believes that FDDI-II chip sets aimed at multimedia LAN applications will begin to arrive in the market during the 1992 to 1993 time frame.

FDDI-II will support multimedia applications by providing 6-Mbps increments of bandwidths

(carved out of the 100-Mbps ting) dynamically allocated to isochronous T-1 channels for voice and video. Dataquest believes that high-end graphics workstations and high-performance file servers will incorporate ULSI-level single-chip

FDDI-II on the motherboard in the 1994 to

1995 time frame.

Dataquest Perspective

Dataquest sees emergence of a global, highvalue-added, merchant LAN/FDDI chip set market that will grow phenomenally at a CAGR of 27 percent from $350 million in 1990 to

$1.15 billion by 1995. Ethernet and token-ring chip sets make up the bulk of the market today. However, because of the emerging need for high-bandwidth, fault-tolerant, enterprisewide networking for graphics and multimedia applications, Dataquest forecasts the high-speed FDDI/

CDDI chip set market to grow dramatically from $25 million in 1990 to $250 million by

1995.

Dataquest believes that semiconductor companies should forge timely, strategic OEM partnerships unth

LAN/FDDI networking equipment companies in order to take advantage of end-user experience and potentially big design wins.

The LAN chip set market is segmenting into low-end, midrange, and high-end segments that address different price/performance points within end-user markets. PC, workstation, and high-performance file server products will incorporate customized, feature-rich, high-end LAN chip sets on the motherboard in their quest for product differentiation in an increasingly commodity hardware market. Dataquest believes that semiconductor companies should foige timely, strategic OEM partnerships with LAN/FDDI networking equipment companies in order to take advantage of end-user experience and potentially big design wins.

By Krishna Shankar

Atitofnotive Applications: More Controls

Offset Vicious

Economics

Clearly 1991 was a year to forget if one depends on the global automotive and truck electronics market. The dominant markets for electronics-rich vehicles are the G7 economies, all seven of which were either in recession or slowing during the year. The outlook for 1992 is for a moderate recovery in consumer durables, which includes vehicles.

The continued penetration of new and more sophisticated controls and features into vehicles continues to buffer the market variability and to offer new opportunities. Figure 1 illustrates the various phases of the automotive electronics market. Cars and trucks for the North

American market have entered phase III, while vehicles in Japan are in the later stages of phase n and European vehicles are midway through phase II.

N e w Opportunities

There are dozens of emerging uses for elertronics in vehicles. Three forces generally push the increased use of electronics, as follows:

• Regulations for reduced air emissions, improved fuel economy, and safety

• Product differentiation

• Economics—cases where chips and printed circuit board (PCBs) are cheaper than electromechanical parts

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012434

6

Semiconductor Application Markets Worldwide

Figure 1

Road Map for Vehicle Electronics Penetration

, , . , r - - ^ : i : : j j i L

^ " ^ ^ ^ ^ ^

Phase 1

Legislation

Phase 11

Legislation

Phase III

Phase IV

Entertainment-

Analog

Entertainment-DSP

Power train 1

Source: Dataquest (Uecem ber

im

Emission Controls Shift to

Next Gear

Designing an engine that is both efficient at using hydrocarbon energy and able to simultaneously minimize air emissions is a techno-economic challenge. The California Air

Resources Board (CARB) continues to be the global pacesetting regulatory body for establishing emission requirements. CARB regulations are being adopted across the United States in the form of the Clean Air Act of 1990. Many European and Asian countries are also following

CARB guidelines. The principal impact on the semiconductor industry will be a new evolution of power train electronic controls.

In the United States and Japan, these controls are becoming 32-bit microcontroller unit (MCU)based, with multipoint electronic fuel injection

(EFI), integrated cruise control, and distributorless ignitions. This 32-bit real-time processing also allows integration of transmission control for optimal shifting.

Electric Vehicles

Another irmovation from CARB is a requirement that 10 percent of the new cars sold in California by 2003 (2 percent by 1998) have zero air emissions (that is, be electric). These vehicles would be battery operated using electric motors rather than engines. The concept is to tai:get commuting vehicles and let the local power

Power train II

Antilock braking

Electronic suspension

Electronic steering

Airbag/restraints

Multiplexina

Collision avoidance

Navigation

Intelligent highways plants (with strict emission controls) with coalderived electric energy power the commuters.

To support development of an electric car battery, an industry group known as the U.S.

Advanced Battery Consortium has been formed.

Its charter is to develop a quick-chaige, highpower-density battery that can run a vehicle more than 100 miles at freeway speeds before recharging. The group comprises Chrysler Corporation, Ford Motor Company, and General

Motors Corporation (GM); EPRI (an electric utility research group); and the U.S. national laboratories. The consortium plans to spend

$260 million over the next four years, with half of the funding coming from the U.S. government.

The principal impediments to the electric vehicle market are costs and standards.

Green movement pressure in Europe is experted to stimulate efforts for that market as well. The Japanese effort seems to be a waitand-see approach, although Nissan's FEY technology based on nickel-cadmium batteries is capable of a range of 156 miles at a maximum speed of 81 mph. In the United States, GM is

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-6000 / Fax (408) 437-0292

0012434

Semiconductor Application Markets Worldwide

the apparent leader In electric vehicle technology; a production variant of its Impact test vehicle is rumored to be ready by 1993.

The electronic content of an electric vehicle should be similar in doUar value to the combustion engine version, but different power train technology will be needed. The traditional power train electronic elements would be replaced by battery and electric motor controls.

A transmission might not be needed because direct-drive DC motors accommodate torque demands. Semiconductor applications for electric vehicles include the following:

• Battery charging (4 to 6 hours), management

Goading and distribution), and monitoring

(status and safety)

• DC propulsion motors and management

• Greater than 300V smart power ICs and transistors, switches, DC/DC conversion, voltage regulators, and protection circuits

• 16- to 32-bit MCUs for motor control

O Multiplexing and diagnostic interface ICs

The principal impediments to the electric vehicle market are costs and standards. Key costs include those of the vehicle and inclusive battery technology, maintenance (such as replacement batteries), and infrastructure changes

(for example, 220/440V battery recharging units for the home). On the standards front, the

U.S. National Highway Traffic Safety Administration estimates that it will take 18 months to establish safety standards (that is, 300V handling), service tools, and service training.

Assuming that other U.S. states and countries follow California's lead, the electric vehicle market could reach 1 million units a year by 2003.

32-Bit Controls Stimulated by

Smart Transmissions

Many vehicle manufacturers, including Ford,

GM, and Toyota, have announced their intention to use 32-bit MCUs in the power train.

Motorola Incorporated is the prime beneficiary of design-wins to date. The design goal is to integrate traditional engine control functions with those of transmission and cruise control management. The driving force behind this integration is the processing power the CARB regulations will require, along with a separate request for improved fuel economy. The U.S.

Congress is expected to soon approve some . form of increase in the corporate average fuel economy (CAFE) requirement from 27.5 mpg today to 34 mpg or higher by 2(X)1.

Car and truck semiconductor applications will continue to present an attractive opportunity through the forecast period.

It is interesting to note that Ford chose a variant of Motorola's 88K RISC technology as its controller. Ford cited small die size and the availability of high-level language programmability as some of the decision criteria. Other semiconductor opportunities for electronic transmission control include fuzzy logic (to augment the

MCU) for shifting decisions, power transistors

(either MOSFET or bipolar) to drive solenoids, and various sensors (accelerator pedal, fluid pressures, and torque).

Air Bags

Air bags continue to penetrate the available market as regulatory deadlines near. Dataquest estimates that more than 25 million units (including passenger-side air bags) will be shipped worldwide annually by 1995. Nissan, Siemens,

Toyota, and TRW Incorporated are some key air bag controller OEMs. These shipments translate into a $350 million annual semiconductor market for air bag control units (assuming a

$14 content). Next-generation bill-of-material requirements include the following:

• Voltage regulator, zener diode, DC-DC converter

• 8-bit MCU with 2-Kbit serial EEPROM, 10-bit

A/D at 18us

• Driver circuits (5-5)

• Signal conditioning circuits (from sensors)

• Acceleration sensors (2 solid-state)

Antilock Braking System (ABS)

Like air bags, antilock braking systems are being driven by the consumer desire for safety.

Dataquest projects that the world market for these systems could exceed 30 million units annually by 1995. The semiconductor content for a typical four-wheel ABS (courtesy of

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012434

8 Semiconductor Application Markets Worldwide

Dataquest's ESAM service) runs at $24 including two MCUs (8- or l6-bit), voltage regulators, drivers, signal conditioners, a diagnostic IC, and an EEPROM. This translates into a $450 million semiconductor market annually by 1995. Key

OEMs include Bosch, Delco, ITT-Teves, Kelsey-

Hayes, Nippon ABS, and Nippondenso.

Multiplexing

Efforts at creating standardized data communication inside and outside of the car continue. The key issue inside the vehicle involves settling on and implementing a protocol. The leading candidates are the SAE J1850 and the Bosch/Intel

CAN. J1850 appears to be the leader for diagnostic or class B applications as it has the broadest base of support. It operates between

10 and 100 Kbps over a single wire or a differential pair.

Remaining issues are related to implementation and will most likely take another year or two to settle. The total available worldwide market for standalone J1850 or (IAN communications

ICs, assuming that the price is right ($2 to

$4 per node), could reach 120 million units annually by 2000 (assuming 10 nodes per vehicle penetrating 25 percent of all vehicles).

A substantial fraction of this function could move to the MCU communications block, however.

Navigation and Intelligent

Highways

The need to navigate optimally through heavy urban traffic or over seldom-traveled country roads continues to be a draw for vehicle navigation systems, especially for mobile jobs such as trucking, repair services, and emergency crews. Irutial versions used dead-reckoning technology with CD-ROM-based maps. The latest versions are using the U.S. Air Force's Global

Positioning System (GPS) to add satellite-derived location data to improve the location calculation. The driver interface is a CRT display that can show a map, a suggested route, current location, and estimated time of arrival. The rest of the system comprises motion sensors, odometer input, receivers (GPS or other RF directional sources), a navigation computer, and a

CD-ROM. (For more information, see SAM newsletter 1991-11, "OEM Monthly: Personal

Communications Stimulates RF Market," for GPS receiver semiconductor content.) Suppliers of these systems Include Blaupunkt, Pioneer

Electronics, Teletrac (for locating stolen vehicles), Toyota (with Nippondenso), Nissan, and

Mazda.

Communications technology between the car and road system also continues to be developed. Use of such systems includes traffic congestion avoidance and automatic vehicle identification (for tolling, among others). Many international efforts are under way in this area, including a pan-European effort known as

Prometheus, a U.S. effort known as Intelligent

Vehicle Highway Systems (IVHS), and in Japan the Advanced Mobile Traffic Information and

Communication System (AMTICS). Each system is a testbed examining feasibility and proving concepts.

Semiconductor Consumption

Forecast

Figure 2 presents Dataquest's forecast of worldwide automotive and truck semiconductor consumption. Overall consumption is expected to grow at a 10.9 percent CAGR until 1995, when it reaches $4.8 billion. TTie key growth markets will be in Europe and Asia/Pacific. Growth in

Europe is driven by tougher emission laws and air bag and ABS penetration. Asia/Pacific countries such as Taiwan are imposing pollution laws as they mature and can now afford to concentrate on such issues. Because of the relative maturity of their markets. North America and Japan will grow more slowly; but applications such as advanced power train control, advanced entertainment (such as CD/DAT), ABS, and air bag controls will be the main drivers of those markets.

The challenge is investing in effective customer relations, the bestquality programs, and the right technology and products.

Key semiconduaor technologies for automotive use will be analog and mixed-signal applicationspecific ICs (ASICs) and functions (bipolar and

BiCMOS particularly), power ICs and transistors

(bipolar and DMOS-MOSFET), MCUs (8-, l6-, and 32-bit with configurable memory blocks), and to a lesser degree standalone MOS nonvolatile memories.

©1991 Dataquest Incoiporated / 1290 Eidder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012434

Semiconductor Application Markets W o r l d w i d e

Figure 2

Worldwide Automotive Semiconductor Forecast (Millions of Dollars)

Millions of Dollars

5 0 0 0 - ,

CAGR

1990-1995

10.9%

$2,794 ^ - 8 ^

$3,223

$3,850

$4,412

1990

*Estimate

Source: Dauquest (December 1991)

1991*

Figure 3

Worldwide Automotive Semiconductor Market Shares—1990

OKi 4.2%

National Semiconductor 4.3%

Philips 4.3%

Texas Instruments 4.4%

Source: Dataquest (December 1991)

Motorola 14.0%

Hitachi 9.9%

Toshiba 8.45%

Siemens 6.3%

SGS-Thomson 5.2%

NEC 4.9%

1995

©1991 Dataquest Incoipoiated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012434

10 Semiconductor Application Markets Worldwide

Industry Players

Motorola is the world's largest merchant supplier of automotive and truck semiconductors

(see Figure 3). An estimated 40 percent of its revenue is from versions of its popular 6805 and 68HC11 families. The company recentiy announced engine controller design wins with its 32-bit 683XX and 88K families for Ford, GM, and Toyota. Motorola also sells most of its broad produa line of analog ICs and discretes into this marketplace. A substantial fsercentage of Motorola's automotive revenue is for custom or semicustom versions of its standard products.

Its success in vehicle applications is dearly attributable to focused customer relations and

Baldrige-aw^ard-winning quality. An in-house customer such as the Motorola Automotive and Industrial Electronics Group is also an advantage.

Japanese suppliers such as Hitachi Ltd., NEC

Corporation, and Toshiba Corporation have benefited from global market share increases by

Japanese automakers. All three companies offer automotive product portfolios ranging from LEDs to microcontrollers. Especially interesting to note are gains made by European suppliers Philips,

SGS-Thomson Microelectronics, and Siemens— primary beneficiaries of high growth of the

European market, largest vehicle market in the world. These three grew an average 23 percent last year.

Dataquest Perspective

Car and truck semiconductor applications will continue to present an attractive opportunity through the forecast period. The reward for diligent suppliers is million-unit orders, the likes of w^hich are seen only in the consumer and PC/ peripheral markets. The challenge is investing in effective customer relations, the best-quality programs, and the right technology and products. All three are needed; if any one is missing, the business would be unattainable and unprofitable. Moreover, as in those other highvolume markets, economics is paramount. A

5 percent anomaly in a supplier's cost structure could be fatal in a market where design-ins can produce for five or more years.

By Gregory Sheppard

Inquiry Summary

Semiconductor

Application Markets

Inquiry Highlights

Dataquest's response to specific questions from clients frequently can be useful for other clients. In this article we provide responses to three recent client questions.

Q: Which are the leading semiconductor sup-

pliers to the global communications equipment market?

A: The following table lists the 1990 leaders, in terms of millions of dollars in revenue.

Company

Motorola

NEC

Toshiba

Fujitsu

AT&T

Hitachi

Texas Instruments

SGS-Tomson

Philips

National Semiconductor

Others

Source: Dataquest (December 1991)

Revenue

($M)

927

887

719

691

630

506

397

357

305

251

2,178

©1991 Dataquest Incorporated / 1290 Ridder Parle Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012434

Semiconductor Application Markets Worldwide

11

Q: What is the size of the North American

point-of-sale system market?

Q: What are the leading application candidates

for 32-bit control?

A: The following table provides size data, in terms of millions of dollars in revenue.

PC-Based Terminals

Electronic Cash

Registers

POS Terminals

Scanners

Credit Authorization

Terminals

Total

1990

610

220

175

60

125

1995

665

165

175

180

135

1,190 1.320

North American

Production (%)

Semiconductor

Content (%)

65

12

50

13

Source: EIA, Depaitment of Commerce, Dataquest

(December 1991)

CAGR (%)

1990-1995

1.7

-5.6

0

24.6

1.6

2.1

A: The following table summarizes in units some leading opportunities.

Automotive Engine Management

Graphics Boards—^High End

Imaging (Medical/Industrial)

Disk Drive—^High End/Arrays

X Window Terminal

Laser Printer

AudioA'bice PC Processing

Digital Video Boards (PC/

•Workstation)

FDD! NlCs/Bridges/Ctoncentrators

Ethemet or Token-Ring Network

Hubs/Bridges

PBX Control

High-Speed Modem (29.6K)

Digital Cellular

1990

0

450

48

400

70

3.965

800

40

4

70

1995

800'

6.100

105

900

860

9,580

18,000'

4,500*

1,900*

200*

68

590

0

105*

3,000*

4,500'

'Includes integrated transmission control; future tipp'&Miaai: active suspension and collision avoidance

'Assumes 40 percent of PC platforms have some sort of communication processor including fax and modem functions

'Worldwide preliminary estimate; assumes penetration of

10 percent of PCs

Worldwide preliminary estimates

'Assumes migration to 32-bit DSP solution for hig^-end and bytmd (AMPSAX>MA or CDMA) versions

Source: Dataquest (December 1991)

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-8000 / Fax (408) 437-0292

0012434

12

Semiconductor Application Markets Worldwide

In Future Issues

The following topics will be featured in future issues of Semiconductor Application Markets

Worldwide Dataquest Perspective:

• Opportunities in pen-based computers

• Opportunities in workstations

For More Information . . .

On the topics in this issue Semiconductor Application Markets Worldwide (408) 437-8261

About on-line access On-line Service (408) 437-8576

About other Dataquest publications Sales (408) 437-8246

About upcoming Dataquest conferences Conferences (408) 437-8245

About your subscription Customer Service (408) 437-8402

Via fax request Fax (408) 437-0292

The content of this report represents our InterpreUtion and analysb of information generally available to the public or released by responsible individuals in the subject con^anles, but is not guaranteed as to accuracy or completeness. It does not contain material provided to us in confidence by our events. Individual companies repotted on and analyzed by

Dataquest may be clients of this and/or other Dataquest services. This information is not fiimished in connection with a sale or offer to sell securities or In connection with the solicitation of an offer to buy securities. This firm and its parent and/or their offlcets, stockholders, or members of ihelr families may, from time to time, have a long or short position in the securities mentioned and may sell <»* buy such securities.

©1991 Dataquest Incorporated / 1290 Ridder Park Drive, San Jose, CA 95131-2398 / (408) 437-flOOO / Fax (408) 437-0292

0012434

AC^

.^^'

> ^

We've Changed Our Format to Meet Your Needs!

w h e n you open this cover, you'll see a bold new look for your Dataquest publications. Dataquest is committed to providing you with accurate, timely information on high-technology companies, markets, and industries. We're equally committed to presenting our analysis to you in a consistent, easy-to-use format across all of the industries that Dataquest tracks.

What's Changed?

Chances are, you're inundated with trade journals, business periodicals, and research publications—all of which you have to sift through to find the critical facts and insight you need to stay on top of the competition. Our clients told us that they wanted their Dataquest information to be more timely and easier to use. And we

listened!

We've streamlined the number of reference binders per service so that you

and your staff can quickly gain access to the specific Dataquest information that you need. Each service now provides two basic binders: a binder of twice-monthly publications called Dataquest Perspective, and an organized binder of reference and statistical documents called Source: Dataquest. You will also receive a binder for each service segment you elect to purchase.

• We've increased the frequency of our publications to provide you with timely, up-to-date information on major topics and events of interest in your market.

Dataquest Perspective is a publication that includes a synopsis for each

article, and our analysts' assessment of the overall significance of each topic covered. Now you can quickly grasp the impact of major events in your industry.

• We've indexed and cross-referenced our publications to allow you and your staff to save valuable research time when looking for information on a specific topic. A complete, updated index will be delivered quarterly, providing a year-to-date cross-reference by company name and major topic.

• We've designed the total look and format of our publications and reference documents with your needs in mind, so that your Dataquest material w i l l be easier to read, route, and file.

What Will You Find In Your Dataquest Perspective Binder?

This binder holds issues of the Dataquest Perspective, the multitopic publication that contains in-depth Dataquest analyses. A quarterly index lists the articles by title for your easy reference. Dataquest Perspective will have articles in the following areas:

Market Ana/ys/s-Provides forecasts, trends, and market share analysis by segments such as products, regions, applications, and distribution channels

Product /\na/ys/s—Presents the impact of new products on the industry

Company /\na/ys/s—Highlights new activities or organizational changes and provides analysis of strategies, finances, and major product segments

Technology Analysis—Zeroes in on key or changing technology and its impact on the markets and companies in the industry

Conferences and Exhibitions—\dent\i'\es trends and analyzes key events at recent conferences and trade shows attended by Dataquest analysts

News and V7evv5—Provides Dataquest's perspective on recent major industry events

What Will You Find in Your Source: Dataquest Binder?

Source: Dataquest is an annually updated reference binder in which you'll find the

following:

• Cu/des-Documents that clarify Dataquest definitions and methodologies

Market Sfaf/sf/cs~Detailed market shares, history, and forecasts

Company Backgrounders—A set of in-depth backgrounders on the top players in your industry. Each backgrounder contains useful information on a company's finances, product line, sales and manufacturing locations, joint ventures, and mergers and acquisitions

(

What's in the Segment Binders?

If you subscribe to a segmented service, you will receive a binder for each segment

of the service to which you subscribe in addition to the Source: Dataquest binder.

Your segment binder consists of the following:

Market Statistics-Market share and forecasts for detailed product classifications

Dataquest Perspectives—VJhere you'll file your segment-specific Dataquest

Perspective publications and the annual index

Please fold and seal your reply card before mailing

BUSINESS REPLY MAIL

FIRST CLASS PERMIT NO. 7279 SAN JOSE, CA

POSTAGE WILL BE PAID BY ADDRESSEE

Dataquest Incorporated

ATTN: Liz Levy, M/S 290-1336

1290 Ridder Park Drive

San Jose CA 95131-9980

SAM-1091

NO POSTAGE

NECESSARY

IF MAILED

IN THE

UNrrED STATES

How Are We Doing?

N o w that y o u ' v e had a c h a n c e t o read t h r o u g h y o u r first Dataquest Perspective p u b l i c a t i o n , w h a t d o y o u t h i n k o f it? W e ' v e i n c l u d e d a reply c a r d so that y o u c a n rate o u r efforts. W e need y o u r feedback so that w e can c o n t i n u e t o d e v e l o p a n d e v o l v e o u r research d o c u m e n t s t o meet y o u r needs!

Please take a f e w minutes t o f i l l in y o u r responses b e l o w . T h a n k y o u for y o u r feedback!

This Dataquest Perspective format is easier to read than the previous newsletter-style format used.

Strongly Somewhat Somewhat Strongly

Agree Agree Disagree Disagree

• • • •

I prefer the multitopic style of Dataquest

Perspective to the single-topic style of the

research newsletters.

G

1 found the articles interesting and relevant.

G

The articles contained the appropriate depth of analysis.

G

Delivery of Dataquest Perspective twice a month is satisfactory.

I would rate the overall quality of the Dataquest Perspective as:

Q

G Excellent G Good G Average G Fair

G Poor

G

G

Comments or Suggestions:

G

G

G

I'd like Dataquest to cover these topics in upcoming issues of Dataquest Perspective:

Name:

(Optional)

Company:

You m a y also call and leave a recorded message o n the Dataquest H o t Line at

(408) 4 3 7 - 7 8 7 8 .

Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement