HP 9000 midrange servers from HP—
The HP 9000 rp7440 and rp8440 Servers
Executive summary............................................................................................................................... 3
Adaptive, flexible, and scalable: The HP 9000 rp7440 and rp8440 Servers ............................................. 4
Introduction ..................................................................................................................................... 4
New, innovative HP chipset delivers more power, availability, and manageability ............................... 4
HP 9000 rp7440 and rp8440 Server specifications ................................................................................ 6
System architecture .............................................................................................................................. 9
HP 9000 rp7440 Server architecture ................................................................................................. 9
HP 9000 rp8440 Server architecture ............................................................................................... 10
HP Super-Scalable Processor Chipset sx2000.................................................................................... 10
Cell boards ................................................................................................................................... 11
Cell design details ...................................................................................................................... 12
Cell configurations...................................................................................................................... 13
Memory latency ......................................................................................................................... 13
Cell hot-plug .............................................................................................................................. 14
Crossbar backplane ....................................................................................................................... 14
Crossbar chips and links ............................................................................................................. 14
I/O subsystem ............................................................................................................................... 15
I/O controller chips .................................................................................................................... 16
PCI-X backplane ......................................................................................................................... 16
Core I/O ...................................................................................................................................... 18
HP 9000 rp7440 Server core I/O ............................................................................................... 18
HP 9000 rp7440 Server internal peripheral bay............................................................................ 20
HP 9000 rp8440 Server core I/O ............................................................................................... 21
HP 9000 rp8440 Server internal peripheral bay............................................................................ 22
HP Server Expansion Unit 2............................................................................................................. 22
AC power subsystem ...................................................................................................................... 24
Fault-tolerant power compliance in the HP 9000 rp7440 Server ...................................................... 24
Fault-tolerant power compliance in the HP 9000 rp8440 Server ...................................................... 25
AC power consumption .................................................................................................................. 26
Performance and scalability ................................................................................................................ 27
Important speeds and feeds............................................................................................................. 27
Scalability ..................................................................................................................................... 28
Racking ............................................................................................................................................ 28
Third-party racks ............................................................................................................................ 29
Server virtualization ........................................................................................................................... 29
Hard partitions .............................................................................................................................. 30
vPars ............................................................................................................................................ 30
HP Essentials Global Workload Manager ......................................................................................... 30
HP Instant Capacity ........................................................................................................................ 31
High availability ................................................................................................................................ 31
Partition reliability .......................................................................................................................... 32
Processor protection.................................................................................................................... 32
ECC on caches .......................................................................................................................... 32
Automatic processor deconfiguration ............................................................................................ 32
Processor cooling ....................................................................................................................... 32
HP Instant Capacity .................................................................................................................... 33
Memory protection ..................................................................................................................... 33
Double Chip Spare..................................................................................................................... 33
Dynamic memory resiliency ......................................................................................................... 33
Hardware memory scrubbing....................................................................................................... 33
Protection for I/O ....................................................................................................................... 34
HP 9000 rp7440 and rp8440 Server crossbar and I/O backplane protection .................................. 34
Reliability in the cabinet infrastructure............................................................................................... 34
Dual AC line cord support ........................................................................................................... 35
Resilience to service processor failures .......................................................................................... 35
Enhanced serviceability .................................................................................................................. 35
Investment protection.......................................................................................................................... 35
Upgrading the HP 9000 rp7420 and rp8420 Servers ....................................................................... 36
Ease of management.......................................................................................................................... 36
Management processor .................................................................................................................. 36
HP System Insight Manager............................................................................................................. 36
HP-UX ........................................................................................................................................... 37
HP Services for the HP 9000 rp7440 and rp8440 Servers ..................................................................... 38
Evolve your infrastructure confidently with a partner that stands accountable ......................................... 38
Full lifecycle services....................................................................................................................... 39
The HP difference .............................................................................................................................. 41
Proven experience and expertise ..................................................................................................... 41
For more information.......................................................................................................................... 42
Executive summary
The HP 9000 rp7440 and rp8440 Servers and the PA-8900 processor bring you all the computing
power you need to operate in the most demanding IT environments. Leading performance and
compute density, coupled with the flexibility and scalability of the HP 9000 rp7440 and rp8440
Servers, allow you the agility and investment protection you need to meet today’s commercial and
technical computing demands while improving your return on IT. HP 9000 servers easily outpace the
competition by providing a broader range of solutions with more applications at a higher
performance In addition, with flexible configurations, competitive pricing, and financial incentives, the
HP 9000 rp7440 and rp8440 Servers make high-availability computing an affordable reality for
your enterprise.
Powered by PA-RISC and HP-UX, the HP 9000 rp7440 and rp8440 Servers give you the
performance and financial flexibility to power every application from Web serving to mission-critical
enterprise resource management. Additionally with in-box upgrades, HP offers a solid roadmap to
protect your investment, to current and future Intel® Itanium® processors, offering you tremendous
capacity and scalability over time.
3
Adaptive, flexible, and scalable: The HP 9000 rp7440 and
rp8440 Servers
Introduction
Today’s HP 9000 servers outpace competitive servers by providing a broader range of solutions with
more applications at a higher performance, across both commercial and technical computing. The HP
9000 server family offers hardware and software solutions for every tier of the enterprise, from entrylevel to high-end enterprise computing, featuring the HP 9000 rp3410, rp3440, rp4410, rp4440,
rp7440, and rp8440 Servers and the HP 9000 Superdome. This paper covers the midrange
HP 9000 rp7440 and rp8440 Servers—two highly flexible and scalable members of the HP 9000
server family, powered by PA-8800 and PA-8900 processors.
New, innovative HP chipset delivers more power, availability, and manageability
The new HP enterprise systems chipset, the HP Super-Scalable Processor Chipset sx2000, is designed
to provide the scalability, reliability, manageability, and performance to meet your most demanding
server needs. The chipset, used in the latest HP 9000 high-end and midrange servers, supports the PARISC PA-8900 processor along with Intel Itanium 2 processors and creates a foundation for future
upgrades to Itanium 2 processors. This capability offers the flexibility and longevity to maximize your
return on IT investment.
This white paper introduces you to the technical details of the HP 9000 rp7440 and rp8440 Servers
beginning by describing their modular designs, system architectures (including the HP Super-Scalable
Processor Chipset sx2000), cell board design and configurations, and I/O subsystems. Every aspect
of the operation and design of the HP 9000 rp7440 and rp8440 Servers is covered, including
performance and scalability, unit racking, partitioning, manageability, serviceability, and high
availability features.
4
Figure 1. HP 9000 rp7440 and rp8440 Servers
5
HP 9000 rp7440 and rp8440 Server specifications
HP 9000 rp7440 Server
with 8 PA-8900 processors
HP 9000 rp8440 Server
with 16 PA-8900
processors
HP 9000 rp8440 Server
Expansion Unit (SEU-2)
Cell boards
1–2
1–4
—
Dual-Core PA-8900
processors
1–8 processors
1–16 processors
—
(2–16 cores)
(2–32 cores)
1.068 GHz processors
2–8 processors
2–16 processors
—
Memory (1-GB, 2-GB, or
4-GB DIMMs)
2–128 GB
2–256 GB
—
Hot-plug PCI-X I/O slots
15 slots
16 slots
16 slots
Cell controller to I/O
controller bandwidth
11.5 GB/s
11.5 GB/s
11.5 GB/s
PCI-X slot single bus
bandwidth (qty.)
533 MB/s (2)
533 MB/s (2)
533 MB/s (2)
PCI-X slot Fat Link
bandwidth (qty.)
1,066 MB/s (6)
1,066 MB/s (6)
1,066 MB/s (6)
PCI-X slot dual Fat Link
2,128 MB/s (8)
2,128 MB/s (8)
2,128 MB/s (8)
Internal disk storage slots/
max. capacity
4/1,200 GB
4/1,200 GB
4/1,200 GB
Internal removable media
slots (DVD, DAT)
2 (with slim-line DVD+RW)
2
2
Hard partitions
2
2 (4 with SEU-2)
—
Hot-swap redundant power
supplies (N+1 included)
2
6
2
Hot-swap redundant fans
(N+1 included)
Yes
Yes
Yes
High availability
features
Hot-swap redundant fans and bulk power supplies
Redundant power line inputs for dual grid protection
Error checking and correcting (ECC) on all processor and memory paths
Double Chip Spare
Cell Link level retry and redundant links
Parity-protected I/O data paths
Operating systems
HP-UX 11i v1, v2, v3 (v2 support is planned in mid 2007; v3 support in planned in 2H
2007)
(533 MHz FSB )
Bandwidth (qty)
1 (with half-height devices)
Dimensions:
Height—Rack
—Pedestal
Width
445 mm (17.5 in.)
755 mm (29.7 in.)
400 mm (15.75 in.)
527 mm (20.75 in.)
833 mm (32.8 in.)
—
482 mm (19.0 in.)
482 mm (19.0 in.)
482 mm (19.0 in.)
6
HP 9000 rp7440 Server
with 8 PA-8900 processors
HP 9000 rp8440 Server
with 16 PA-8900
processors
HP 9000 rp8440 Server
Expansion Unit (SEU-2)
Depth
762 mm (30.0 in.)
762 mm (30.0 in.)
762 mm (30.0 in.)
Typical maximum power
consumption (for maximum
configuration)
2120VA (2078 W)
3866VA (3789 W)
613VA (601 W)
Weight (max.)
101.6 kg (224 lb)
171.4 kg (378 lb)
81.6 kg (180 lb)
Figure 2. HP 9000 rp7440 Server
Front and side view
Rear view
Figure 2 shows major components of the HP 9000 rp7440 Server, as well as the system’s mechanical
and architectural features. The HP 9000 rp7440 Server is shown with its front plastic bezel and top
and left side panels removed.
A peripheral bay located at the top front of the HP 9000 rp7440 Server provides space for four hotplug disk drives and one removable-media device (DVD or DAT) (two slimline DVDs +RWs are
available for the HP 9000 rp7440 Server). Directly below the peripheral bay are two redundant PCIX power bricks, which supply DC power for the PCI-X backplane. Below the power supplies are two
redundant hot-swappable cooling fans. These fans pull cool air in from the front and force air to the
rear, cooling the system’s internal components. At the bottom is the bulk power supply (BPS) bay,
which houses two redundant (2N) hot-swap power supplies with dual grid support.
Figure 2 also shows the right side of the HP 9000 rp7440 Server, with a view of the cell board bay,
which supports up to two cell boards. The cell boards contain processors, memory, and cell controller
chips.
The rear view of the HP 9000 rp7440 Server shows the location of the two hot-swappable 150-mm
exhaust fans and the I/O bay bulkhead directly above them. The core I/O cards are located at the
right edge of the unit.
The system backplane board houses the linkages used for communications between cell boards, I/O,
and internal peripherals. The HP 9000 rp7440 Server I/O card bay is located at the top rear of the
system. It contains 15 PCI-X card slots, all supporting hot-plug functionality.
The bottom rear of the HP 9000 rp7440 Server has inputs for the 2+2 redundant line cords. Because
of the high degree of connectivity offered in the HP 9000 rp7440 Server, a cable management arm
(not shown) is provided for dressing cables and simplifying cable routing.
7
The HP 9000 rp7440 and rp8440 Servers share many system components and design features as
seen below. Figure 3 shows a front/side and rear view of the HP 9000 rp8440 Server with its front
plastic bezel and top and left side panels removed.
A peripheral bay located at the top front of the HP 9000 rp8440 Server provides space for four hotplug disk drives and two removable media devices (DVD or DAT). Directly below the peripheral bay
are two redundant PCI-X power bricks, which supply DC power for the PCI-X backplane. Below the
power supplies are nine redundant hot-swappable cooling fans. These fans pull cool air in from the
front and force air to the rear, cooling the system’s internal components. At the bottom is the bulk
power supply bay, which houses up to six redundant (2N+1) hot-swap power supplies.
Figure 3 also shows the side of the HP 9000 rp8440 Server, with a view of the cell card cage. This
cage supports up to four cell boards on which processors, memory, and cell controller chips reside.
Figure 3. HP 9000 rp8440 Server
Front and side view
Rear view
The rear view of the HP 9000 rp8440 Server shows the location of the 12 hot-swappable 120-mm
exhaust fans and the I/O bay bulkhead directly above them. The core I/O cards are located at the
right edge of the unit.
This view shows the system backplane board, which houses the high-bandwidth crossbar used for
communications between cell cards, I/O, and internal peripherals. The HP 9000 rp8440 Server I/O
card bay is located at the top rear of the system. It contains 16 PCI-X card slots, all supporting hotplug functionality.
The bottom rear of the HP 9000 rp8440 Server has inputs for the 2+2 redundant line cords. Because
of the high degree of connectivity offered in the HP 9000 rp8440 Server, a cable management arm
specific for the HP 9000 rp8440 Server (not shown) is provided for dressing cables and simplifying
cable routing.
8
System architecture
HP 9000 midrange servers are built around HP sx2000 chipset which is based on a modular cellbased architecture. Cell-based architecture allows for components (cell boards, IO and more) to be
configured to effectively cover a wide range of computing needs. Both the HP 9000 rp7440 and
rp8440 Servers support a variety of system configurations, ranging from one to sixteen PA-8900
processors. Both servers can be configured as large symmetric multiprocessing (SMP) systems or as
multiple independent hard partitions called nPartitions (nPars). The basic components in the HP 9000
midrange servers are the cell board with processors and memory, the PCI-X based I/O subsystem,
and, in the HP 9000 rp8440 Server, the crossbar backplane. These components fit together to
provide a high-performance computing platform that is scalable, highly available, and flexible.
HP 9000 rp7440 Server architecture
The HP 9000 rp7440 Server architecture is designed around the ability to operate the system as a
single one- to eight-processor SMP server or to divide it into two independent hard partitions (nPars).
Figure 4 shows the primary components of the HP 9000 rp7440 Server architecture. When the
system is configured as a non-partitioned server, all resources shown in Figure 4 are available to
perform together as one logical server. When it is configured as two nPars, system resources are
divided into two logical servers, or independent partitions, each containing a cell board with a
dedicated set of I/O resources. For example, in Figure 4, imagine that the solid line connecting the
upper and lower cell boards is no longer there. The drawing would then reflect a system divided into
two independent partitions. The cell board, I/O bay, core I/O, and peripheral bay in the upper half
of the drawing would be an independent hard partition, which is isolated from the second partition
shown in the lower half of the drawing.
Figure 4. Architecture of the HP 9000 rp7440 Server, showing the basic modular building blocks of the system
Cell boards
I/O bay
7 P CI- X 266
IO s lots
Processor
3x cros s bar links
P eripheral bay
L AN/
SCSI
Processor
Cell
controller
Processor
Core I/O
I/O controller
MP/
SCSI
Processor
Memory
Cell board
Processor
Processor
Cell
controller
Processor
I/O controller
Processor
Memory
Cell board
7/8 P CI- X 266
IO s lots
MP/
SCSI
L AN/
SCSI
All OL AR
9
HP 9000 rp8440 Server architecture
The HP 9000 rp8440 Server architecture builds on that of the HP 9000 rp7440 Server with the
addition of a crossbar backplane and two more cell boards. The crossbar backplane provides a nonblocking connection between up to four cells, plus a connection to the external I/O resources in the
HP SEU-2. Similarly to the HP 9000 rp7440 Server, the HP 9000 rp8440 Server can be configured
as one 1- to 16-processors system, or it can be divided into smaller independent nPars. The HP 9000
rp8440 Server can be divided into four hardware-isolated partitions when connected to the SEU-2.
Figure 5. Architecture block diagram of the HP 9000 rp8440 Server, showing the modularity of the system
P CI B ackplane
8 P CI- X I/O s lots
I/O controller
All hot- plug
8 P CI- X I/O s lots
I/O controller
B ulk
power
s upply
Core I/O
P eripheral bay
Core I/O
Core I/O
dis ks
S ys tem cros s bar
backplane
Cross bar chips
Out to S E U
Cell
controller
Cell
controller
Cell
controller
Cell
controller
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
P roces s or
Memory
Cell board
Memory
Cell board
Memory
Cell board
Memory
Cell board
HP Super-Scalable Processor Chipset sx2000
The HP sx2000 is the HP enterprise class chipset that provides more than twice the memory
bandwidth and over four times the I/O and fabric bandwidths of the sx1000 chipset, along with
significant reductions in systems latencies. This increased capabilities of the HP sx2000 in
combination with the Dual-Core PA-8900 processors delivers significantly greater performance for
your applications. All sx2000 systems use error correcting and self-healing technologies throughout to
reduce hardware outages. Unique technologies include Double Chip Spare (DCS), which immediately
restores chip-spare protection after a DRAM has failed and protects against the majority of memory
buffer and data bus failures, and link self-healing, which recovers from any connector, backplane, or
cable failure without loss of performance.
10
Figure 6. Primary components of the HP sx2000 chipset
Cell board
Memory
I/O
controller
16 DIMMs x 36 DDR
II
8.5 GB/sMem buff
4.3 GB/s each bus
17 GB/s total
CPU socket
CPU socket
8.5 GB/s each bus
17 GB/s total
Cell
controller
5.75 GB/s each bus
11.5 GB/s total per
cell
I/O
controller
CPU socket
CPU socket
11.5 GB/s each bus x 3 / cell
34.6 GB/s per cell
I/O backplane
crossbar
System backplane
Cell boards
The cell, or cell board, is one of the basic building blocks of the HP 9000 midrange servers. A cell
board is a module that primarily houses processors, memory, and the cell controller applicationspecific integrated circuits (ASICs).
11
Figure 7. Layout of the HP 9000 rp7440 and rp8440 Server cell board
Cell design details
Each cell board is a self-contained unit, with an SMP, main memory, and all necessary hardware.
• Processors—Up to four Dual-Core PA-8900 processors (or 8 processor cores/cell board)
• Cell Controller ASIC (CC)
• Memory Mux ASICs
• Main memory DDR-2 DIMMs (up to 16 DIMMs per cell board)
• Voltage regulator modules (VRMs)
• Cell management and processor-dependent hardware (PDH) circuitry
The CC is at the heart of each cell board. It provides the communications link between processors,
memory, I/O, PDH, and adjacent cells. The CC contains interface logic and maintains cache
coherency throughout the system. Adjacent to the cell controller ASIC are up to four PA-8900
processors and up to 256 GB of main memory (currently supports up to 64 GB of main memory using
4-GB DIMMs). Each cell interfaces with adjacent cells and I/O resources directly or, in the case of the
HP 9000 rp8440 Server, to other cells through the crossbar backplane.
The primary function of the Memory Mux ASIC is to multiplex and de-multiplex data between the CC
and the DIMMs in the memory subsystem. When the CC issues a read transaction to the memory
interface command bus, the Memory Mux ASIC buffers the memory read data and returns it as soon
as possible. When the CC issues a write transaction, the memory controller ASIC receives the write
data from the CC and forwards it to the memory.
The sx2000 chipset has twice the memory bandwidth of the sx1000 and over 25% lower memory
access latency when compared to the sx1000.
12
The CC also supports DCS, an increased level of memory availability. The memory error correcting
logic reestablishes chip-spare correction after another DRAM in the same error checking and
correcting (ECC) codeword has failed. This strategy is more cost-effective than memory mirroring
methods for protecting memory. See the HP sx2000 chipset technical white paper for more
information on this feature.
Cell configurations
The HP 9000 rp7440 Server supports up to two cells. The HP 9000 rp8440 Server supports up to
four cells. With Dual-Core PA-8900 processors, each cell can be configured with up to four active
processors. When configured with PA-8900 processors, each cell can be purchased with one to four
active PA-8900 processors. A fully loaded HP 9000 rp7440 Server therefore contains eight PA-8900
processors. A fully loaded HP 9000 rp8440 Server therefore contains sixteen PA-8900 processors.
They an also be purchased in combination with inactive HP Instant Capacity (iCAP) processors.
Within the cell, CPU-to-CC peak bandwidth has been increased to 17 GB/s (13.6 GB/s sustained).
The minimum supported cell configuration is one active Dual-Core processor and 2 GB of memory per
cell board. The maximum configuration includes four active Dual-Core processors and 64 GB of
memory per cell board in the HP 9000 rp7440 Server. The HP 9000 rp8440 Server supports a
maximum of four active Dual-Core processors and 64 GB of memory per cell board. DIMM modules
are sold in sets of two, with available DIMM sizes of 1 GB, 2 GB, and 4 GB. Memory pairs of
different sizes can be mixed within a chassis and within a cell. However, for optimum memory
interleaving and performance, HP recommends that one memory size be selected, distributed evenly
across available cells, and loaded in increments of eight DIMMs.
Within a cell, the CC-to-memory peak bandwidth is 17.2 GB/s, a 2.1X improvement compared to
sx1000. Memory is accessed directly through the CC, so all memory slots are accessed, regardless of
the number of processors loaded on the cell.
Memory latency
There are two types of memory latency within the HP 9000 rp7440 Server:
• Memory latency within the cell refers to the case in which an application runs on a partition that
consists of a single cell or uses cell local memory.
• Memory latency between cells refers to the case in which the partition consists of two cells and cell
interleaved memory is used. In this case, 50% of the addresses are to memory on the same cell as
the requesting processor, and the other 50% of the addresses are to memory on the other cell.
The HP 9000 rp7440 Server average memory latency depends on the number of processors in the
partition. Assuming that memory accesses are equally distributed across all cell boards and memory
controllers within the partition, the average idle memory latency (load-to-use) is shown in the following
table.
13
Number of processors per partition
Average memory latency
4 processors (one cell)
~185 ns
8 processors (two cells)
~249 ns
There are two types of memory latency within the HP 9000 rp8440 Server:
• Memory latency within the cell refers to the case in which an application runs on a partition that
consists of a single cell or uses cell local memory.
• Memory latency between cells refers to the case in which the partition consists of two or more cells
and cell interleaved memory is used. For example, for an HP 9000 rp8440 Server with four cells in
the partition, 25% of the addresses are to memory on the same cell as the requesting processor,
and the other 75% of the addresses are to memory on the other three cells.
The HP 9000 rp8440 Server average memory latency depends on the number of processors in the
partition. Assuming that memory accesses are equally distributed across all cell boards and memory
controllers within the partition, the average idle memory latency (load-to-use) is shown in the following
table.
Number of processors per partition
Average memory latency
4 processors (one cell)
~185 ns
8 processors (two cells)
~249 ns
16 processors (four cells)
~334 ns
Cell hot-plug
The HP 9000 rp7440 and rp8440 Servers support cell hot-plug. Coupled with the partitioning
capability 1 of the servers and appropriate support from the operating system, cell hot-plug allows for
the servicing of cell boards within a single partition while the other partition continues normal
operation. Any number of configuration changes can be made to the partition being serviced,
including replacing the complete cell board, adding or deleting processors and memory, or even
increasing or decreasing the number of cells in that partition. (Cell hot-plug is supported only in
systems with two or more partitions.)
Crossbar backplane
The next basic building block of the HP 9000 rp8440 Server is the crossbar backplane. The crossbar
backplane contains two crossbar chips that provide non-blocking connections between four cells and
their associated memory and I/O. Each cell has three connections to the crossbar fabric. The HP
9000 rp7440 Server does not have a crossbar backplane, so communication between its cells is over
three direct-connect links.
Crossbar chips and links
The Crossbar ASIC is yet another part of the HP sx2000 chipset. The Crossbar ASIC and CC
communicate over a high-speed SERDES-based link using 8 bit/10 bit encoding with clock/data
recovery. As a result, there is no critical clock signal in any link that could be a single point of failure
for that link.
The peak bandwidth for each port has been increased as well to 11.5 GB/s (9.2 GB/s typical),
giving aggregate peak bandwidth of 34.5 GB/s (27.6 GB/s typical) per cell. In addition to
The HP 9000 rp8440 Server can be configured as a single large SMP server or hardware-partitioned into up to four smaller logical servers. See
“nPartitions” for more details about partitioning.
1
14
providing higher bandwidth, the crossbar fabric consisting of the Crossbar ASIC and high-speed links
have been redesigned for the sx2000 to provide greater availability and reliability.
See the HP sx2000 chipset technical white paper for more information on the benefits provided by
new Crossbar ASIC and links.
I/O subsystem
Each HP 9000 rp7440 and rp8440 Server contains an embedded high-performance I/O subsystem.
In addition, the HP 9000 rp8440 Server can optionally connect to external I/O resources located in
the HP SEU-2 through a high-performance I/O cable link. The components within the I/O subsystem
are the I/O controllers, internal peripheral bay, and multifunction core I/O. Basic block diagrams for
the HP 9000 rp7440 and rp8440 Server I/O subsystems are shown in Figures 8 and 9.
Figure 8. Basic block diagram of the HP 9000 rp7440 Server I/O subsystem
I/O bay
Core I/O
7 available PCI-X I/O slots
LAN/
SCSI
I/O controller 1
MP/
SCSI
Peripheral bay
Optional:
slim line DVD
I/O controller 0
7/8 PCI-X I/O slots
MP/
SCSI
LAN/
SCSI
all hot-plug
15
Figure 9. Basic block diagram of the HP 9000 rp8440 Server I/O subsystem
I/O controller chips
The HP 9000 rp7440 and rp8440 Servers contain two master I/O controller chips located on the
PCI-X backplane. Each I/O controller contains 16 high-performance, 12-bit-wide links. These links
connect to 16 I/O controller chips supporting the PCI-X card slots and core I/O.
In both systems, two links—one from each master controller—are routed through the system backplane
and are dedicated to core I/O. The remaining 30 links are divided among the sixteen 64-bit PCI-X
card slots, with each slot on a dedicated PCI-X bus. This one-card-per-bus architecture leads to greater
I/O performance, better error containment, and higher availability.
Each controller chip is also directly linked to a host cell board, which means that two cell boards,
located in cell slots 0 and 1, must be purchased to access all available I/O card slots. (With one cell
board, access to half of the available slots is enabled.)
PCI-X backplane
Figures 10 and 11 show detailed views of the HP 9000 rp7440 and rp8440 Server PCI-X
backplanes. The I/O slot implementations between the two servers are almost identical—the
difference is the use of one or two slots by the HP 9000 rp7440 Server core I/O. In both figures,
note that eight of the 16 I/O card slots are supported by dual high-performance fat links. These duallink I/O slots provide a maximum of 2.1 GB/s of peak bandwidth for the slot. Six of the 16 I/O card
links are supported by high-performance fat links. These fat link slots provide a maximum of 1.06
GB/s of peak bandwidth for the slot. The remaining two I/O slots are single links and provide a
maximum of 530 MB/s of peak bandwidth. Aggregate I/O slot bandwidth is approximately
25 GB/s or a 4.4x increase in bandwidth compared to the sx1000 platforms.
These platform supports eight 266-MHz slots, which means that these I/O slots allows the industry’s
highest-performing PCI-X cards to run at their maximum design speed.
16
Figure 10. The PCI-X backplane of the HP 9000 rp7440 Server has fat link high-performance links for 14 of the 16 I/O card
slots.
T win fat L ink: 2120MB /s
F at L ink: 1060MB /s
E nhanced L ink: 530MB /s
L ink to cell boards
1
0
All hot-plug
P CI- X
P CI- X
P CI- X
I/O
chass is 1
P CI- X
P CI- X
P CI- X
S lot 8
Core L AN/
S CS I
fat link
fat link
fat link
P CI- X
twin fat link
twin fat link
twin fat link
S lot 1
P CI- X
T win fat link
twin fat link
I/O controller 0
P CI- X
fat link
I/O controller 1
S lot 1
P CI- X
T win fat link
P CI- X
T win fat link
I/O
chass is 0
P CI- X
T win fat link
P CI- X
fat link
fat link
E nhanced link
E nhanced link
P CI- X
P CI- X
or
S lot 8
Core L AN/
SCS I
Core I/O
MP /S CS I
Core I/O
MP /S CS I
Figure 11. The PCI-X backplane of the HP 9000 rp8440 Server has fat link high-performance links for 14 of the 16 I/O card
slots.
Twin fat Link: 2120MB/s
Fat Link: 1060MB/s
Enhanced Link: 530MB/s
Link to cell boards
1
0
All hot-plug
PCI-X
PCI-X
PCI-X
I/O
chassis 1
PCI-X
PCI-X
PCI-X
Slot 8
PCI-X
fat link
fat link
fat link
twin fat link
twin fat link
twin fat link
twin fat link
Twin fat link
I/O controller 0
PCI-X
fat link
I/O controller 1
Slot 1
Twin fat link
Twin fat link
Twin fat link
fat link
fat link
Enhanced link
Enhanced link
Core I/O
MP/SCSI
PCI-X
Slot 1
PCI-X
PCI-X
PCI-X
I/O
chassis 0
PCI-X
PCI-X
PCI-X
PCI-X
Slot 8
Core I/O
MP/SCSI
17
In practice, PCI-X I/O cards requiring the largest amount of bandwidth should be configured into the
dual-link slots. Because each I/O slot has a dedicated bus, any slot can be hot-plugged or serviced
without affecting other slots.
Core I/O
The HP 9000 rp7440 and rp8440 Servers are purchased with one or two core I/O card products. In
both systems, core I/O provides console, Ultra320 SCSI, Gigabit LAN, and Management Processor
(MP) functionality, along with SCSI controllers for the peripheral bay. The second core I/O product
can be used to enable dual partitioning, provide access to a second set of disk drives, and provide
redundant MP functionality. In the HP 9000 rp8440 Server, the second core I/O product also
enables the use of an additional removable media device. Although core I/O provides the same
functionality in both the HP 9000 rp7440 and rp8440 Servers, the physical implementation is
different. To accommodate size limitations, the core I/O in the HP 9000 rp7440 Server is divided
across two separate boards, while the core I/O in the HP 9000 rp8440 Server is implemented on
one physical board. The following sections detail the core I/O implementation in each system.
HP 9000 rp7440 Server core I/O
The HP 9000 rp7440 Server chassis supports up to two core I/O card sets. Each set contains two
cards (MP/SCSI and LAN/SCSI), which are installed in different locations: MP/SCSI cards are
installed along the right rear vertical edge of the chassis; LAN/SCSI cards are installed in the PCI-X
card bay. A minimum of one core I/O card set must be ordered with each system; the optional
second core I/O card set can be used to enable hardware partitioning or to utilize the full capacity of
the built-in mass storage bays.
Both core I/O card sets are identical. In the primary and secondary core I/O set, the LAN/SCSI
board is supported by a single 530 MB/s link. In the primary and secondary core I/O set, the two
SCSI controllers—one in the LAN/SCSI card and one in the MP/SCSI card—each support a single
internal disk drive.
Figure 12. Block diagram showing basic core I/O card set (primary) in the HP 9000 rp7440 Server
Link to PCI
converter
530 MB/s bus
530 MB/s bus
Link to PCI
converter
MP/SCSI card
LAN/SCSI card
Ultra320 SCSI
controller
Ultra320 SCSI
controller
Management
processor
10/100Base-TX
LAN
rs -232
peripheral bay
management
LAN
10/100/
1000Mb LAN
Local
Console
Ultra320 LVD
Ultra320 LVD
10/100/
1000Mb LAN
18
Figure 13. Basic core I/O card set (secondary)
Link to PCI
converter
530 MB/s bus
530 MB/s bus
Link to PCI
converter
MP/SCSI card
LAN/SCSI card
Ultra320 SCSI
controller
Ultra320 SCSI
controller
Management
processor
10/100Base-TX
LAN
rs -232
peripheral bay
management
LAN
10/100/
1000Mb LAN
Local
Console
Ultra320 LVD
Ultra320 LVD
10/100/
1000Mb LAN
Management Processor—The MP is a dedicated processor located on each MP/SCSI card that
simplifies and extends system management and enhances serviceability. The MP reduces the need for
the system administrator to be physically at the system to perform tasks such as diagnostics, system
management, or even hard resets. The HP 9000 rp7440 and rp8440 Server MP offers the following
features, among others:
• System management over the Internet or intranet
• System console redirection
• Console mirroring
• System configuration for automatic restart
• Viewing history log of system events and console activity
• Setting MP inactivity timeout thresholds
• Remote system control
• Remote power cycle (except for MP housekeeping power)
• Viewing system status
• Event notification to system console, e-mail, pager, or HP Response Centers (e-mail and pager
notification work in conjunction with HP Event Monitoring Service [EMS])
• Automatic hardware protection of critical environmental problems
• Automatic system restart
• Remote resetting of hardware partitions
• Forward progress indicator (through a virtual front panel)
• Out-of-band manageability and processor-dependent code (PDC) firmware update
• Configuration of manageability and console security
• MP failover (systems with both core I/O boards)
• Secure Sockets Layer (SSL)
19
External LAN—The external LAN port is a 10/100/1000BaseT external LAN port that uses an RJ-45
connector.
External SCSI—The external SCSI port is an Ultra320 LVD external SCSI port for connections to mass
storage or media.
HP 9000 rp7440 Server internal peripheral bay
The HP 9000 rp7440 Server internal peripheral bay is located at the top front of the system chassis.
The peripheral bay holds up to four low-profile hot-plug disks and one removable media device (either
half-height DVD or DAT) or supports up to two low-profile (slim-line) DVDs.
Each HP 9000 rp7440 Server core I/O card set contains dual-channel Ultra320 SCSI controller chips
that support the SCSI devices in the internal peripheral bay. Each core I/O card set supports two
internal disks. Separate controllers and SCSI buses manage the two disks supported by the primary
core I/O card set.
Internal disks—The HP 9000 rp7440 Server holds up to four SCSI disks, which are accessible from
the front of the server. These hot-plug disks can be removed and inserted while the HP 9000 rp7440
Server continues to operate. The system supports the following disks: 36 GB, 15,000 rpm; 73 GB,
15,000 rpm; 146 GB, 10,000 rpm; and 300 GB, 10,000 rpm.
Removable media—The HP 9000 rp7440 Server contains one removable-media bay, which supports
a half-height DVD drive, a half-height DAT drive, or two low-profile (slim-line) DVDs. Access to these
devices is also from the front. The DVD drives provides enhanced features while preserving backward
read compatibility with CD-ROM drives. Data transfer rates of up to 6.75 MB/s are achieved with the
DVD format; 4.8 MB/s can be achieved using the CD-R format. The DAT drive has a maximum
storage capacity of 72 GB, with a peak transfer rate of 21.6 GB/hour (compressed).
20
HP 9000 rp8440 Server core I/O
The HP 9000 rp8440 Server chassis supports up to two core I/O cards, installed in core I/O slots
located along the right-rear vertical edge of the chassis. A minimum of one core I/O card must be
ordered with each system; the optional second core I/O card can be used to enable hardware
partitioning or to utilize the full capacity of the built-in mass storage bays.
Figure 14. Diagram showing the basic core I/O in the HP 9000 rp8440 Server
Link to PCI
converter
Management Processor
10/100Base
-TX LAN
10/100/1000Mb
LAN
SCSI Controller
530 MB/s Bus
SCSI Controller
RS-232
Management
Processor LAN
10/100/1000 Ultra320
Mb LAN
Peripheral Bay
Local
Console
The core I/O management processor, external LAN port, and external SCSI port functionality in the
HP 9000 rp8440 Server is the same as described in the HP 9000 rp7440 Server core I/O section.
However, there are slot count and bus routing differences in the peripheral bay implementation. The
following section pertains specifically to the HP 9000 rp8440 Server peripheral bay.
Access to internal peripheral bay—The first core I/O card enables half of the HP rp8440 Server
peripheral bay, which includes one removable-media device and two low-profile disks. The second
core I/O card enables the remaining internal peripherals, two disks, and one removable media bay.
If you require access to more than two internal disks or more than one removable media slot, you
need the second core I/O card and a minimum of two cell boards.
21
HP 9000 rp8440 Server internal peripheral bay
The HP 9000 rp8440 Server internal peripheral bay is located at the top front of the system chassis.
The peripheral bay holds up to four low-profile hot-plug disks and two removable media devices.
Each HP 9000 rp8440 Server core I/O card contains two dual-channel SCSI controller chips that
support the SCSI devices in the internal peripheral bay. Each core I/O card supports two internal
disks and one removable media device, each on a dedicated 40 MB/s SCSI channel. If use of more
than two internal disks or one removable-media device is needed, the HP 9000 rp8440 Server
requires both core I/O cards. This architecture also provides an added degree of availability by
supporting full disk mirroring across independent buses, controllers, core I/O cards, and master I/O
controller chips.
Hot-plug disk drives—The HP 9000 rp8440 Server holds up to four SCSI disks, which are accessible
from the front of the server. These hot-plug disks can be removed and inserted while the HP 9000
rp8440 Server continues to operate. Three disk sizes are currently supported: 36 GB, 15,000 rpm;
73 GB, 15,000 rpm; 146 GB, 10,000 rpm; and 300 GB, 10,000 rpm.
Removable-media bays—The HP 9000 rp8440 Server contains two removable-media bays, which
support a half height DVD drive or a half-height DAT drive. Access to these devices is also from the
front of the server. The DVD drive provides enhanced features while preserving backward read
compatibility with CD-ROM drives. Data transfer rates of up to 6.75 MB/s are achieved with the DVD
format; 4.8 MB/s can be achieved using the CD-R format. The DAT drive has a maximum storage
capacity of 72 GB, with a peak transfer rate of 21.6 GB/hour (compressed).
HP Server Expansion Unit 2
The HP Server Expansion Unit 2 (SEU-2) is an add-on chassis containing I/O resources that
complement the I/O and partitioning capabilities within the HP 9000 rp8440 Server. The HP 9000
rp8440 Server contains built-in ability to connect to the HP SEU-2. Connection to the SEU-2 doubles
the amount of the HP 9000 rp8440 Server I/O resources, and the SEU-2 enables the creation of two
additional nPars within the chassis. The following features are provided by the SEU-2:
• Sixteen high-performance PCI-X I/O slots
• Eight slots that are 266-MHz x 64-bit capable
• Six slots that are 133-MHz × 64-bit capable
• All slots support hot-plug
• All slots supported by independent dual or single links
• Four disk drive bays
• Two removable half-height media slots (either DVD or DAT)
• Two additional nPars in the HP 9000 rp8440 Server host are enabled
• Two core I/O slots
• Redundant and hot-swap fans and bulk power supplies
• Redundant line cords for dual grid support
• 9U rack-mount chassis
• Certified under Uptime Institute’s Fault Tolerant Power Compliance Specification
22
Figure 15. The HP SEU-2 (left) and the SEU-2 + HP 9000 rp8440 Server mounted in an HP 2-m cabinet (right)
The HP SEU-2 mirrors the I/O resources embedded within the HP 9000 rp8440 Server chassis, both
physically and electrically. The high-speed connection between the HP 9000 rp8440 Server and the
SEU-2 is provided through a remote I/O (E-Link) cable. The E-Link cable mounts directly to the system
backplane board in the HP 9000 rp8440 Server and to the I/O backplane board in the SEU-2.
23
Figure 16. The HP SEU-2 block diagram
P CI- X Backplane
8 P CI I/O s lots
I/O controller
All hot-plug
8 P CI I/O s lots
I/O controller
Core I/O
Core I/O
P eripheral bay
Core I/O
Dis ks
Bulk
power
s upplies
E - L ink cable as s embly
Connection
to host server
AC power subsystem
The HP 9000 rp7440 and rp8440 Servers were designed to take full advantage of the multi-grid
power inputs found in today’s high-end data centers, which means that both servers can connect to
two independent power grids at the same time and tolerate a grid failure without causing computing
interruptions. The power subsystems are so robust that they both are certified (without deviations)
under the Uptime Institute’s Fault Tolerant Power Compliance Specification. The specific AC power
subsystem details are covered separately in the following sections.
Fault-tolerant power compliance in the HP 9000 rp7440 Server
The AC input to the HP 9000 rp7440 Server is divided into four separate circuits. Each circuit is fed
by any 50- to 60-Hz high line source through four line cords (“high line” refers to 200 to 240 V
operation). A minimum of two power cords is used to maintain normal operation of the HP 9000
rp7440 Server. A second set of two cords is added to improve system availability by protecting, for
example, against power grid failures, failed power supplies, or accidentally tripped circuit breakers.
Four power cords are used to enable redundancy and hot-swap functionality of the bulk power
supplies. This power is routed from four individual 20-A circuit breakers through input line filters to
two internal bulk power supplies (BPSs). These four lines are labeled A0, B0, A1, and B1 at the line
filter inputs on the back panel of the HP 9000 rp7440 Server.
Figure 17 shows the HP 9000 rp7440 Server power configuration. Each AC inlet feeds one HP 9000
rp7440 Server BPS, and those two separate AC inlets feed each bulk supply. The design of the
individual BPSs and the configuration of the interconnect meet all the conditions for fault-tolerant
power compliance.
The benefit of this design is that when utilizing all four power cords, the HP 9000 rp7440 Server has
2N redundant power protection.
The dual AC modular bulk power supplies provide:
24
• Redundancy for both hardware failures and power input failures
• Hot-plug capability for any BPS in a redundant configuration
• Better data security, maintenance scheduling, and maintenance operations without system
interruption
Figure 17. Power inputs and interconnects in the HP 9000 rp7440 Server are designed for fault-tolerant power compliance.
Fault-tolerant power compliance in the HP 9000 rp8440 Server
The AC input to the HP 9000 rp8440 Server is divided into four separate circuits. Each circuit is fed
by any 50- to 60-Hz high line source through four line cords (“high line” refers to 200 to 240 V
operation). A minimum of two power cords is used to maintain normal operation of the HP 9000
rp8440 Server. A second set of two cords is added to improve system availability by protecting, for
example, against power grid failures or accidentally tripped circuit breakers. Four power cords are
used to enable redundancy and hot-swap functionality of the bulk power supplies. This power is
routed from four individual 20-A circuit breakers through input line filters to six internal bulk power
supplies. These four lines are labeled A0, B0, A1, and B1 at the line filter inputs on the back panel of
the HP 9000 rp8440 Server.
Figure 18 shows the HP 9000 rp8440 Server power configuration. Each AC inlet feeds three HP
9000 rp8440 Server BPSs, and two separate AC inlets feed each bulk supply. The design of the
individual BPSs and the configuration of the interconnect meet all the conditions for fault-tolerant
power compliance.
Each BPS shown in Figure 18 can be thought of as two sub-power supplies housed in a single BPS
module. If only Grid A power cords are used, only one of the two sub-power supplies is in use. The
other supply is off until the Grid B cords are used. The benefit of this design is that when utilizing all
four power cords, the HP 9000 rp8440 Server has 2N+1 redundant power protection.
The dual AC modular power supplies provide:
• The ability to purchase only the amount of power conversion necessary for the intended HP 9000
rp8440 Server configuration
• Redundancy for both hardware failures and power input failures
• Hot-plug capability for any BPS in a redundant configuration
• Better data security, maintenance scheduling, and maintenance operations without system
interruption
25
Figure 18. Power inputs and interconnects in the HP 9000 rp8440 Server are designed for fault-tolerant power compliance.
AC power consumption
The power consumption of the HP 9000 rp7440 and rp8440 Servers varies greatly, depending on
the hardware configuration and the input line voltages supplied at your site. HP represents power
consumption in terms of volt-amperes (VA) to take into account power factor correction.
• Maximum Theoretical Power or Maximum Configuration—The calculated sum of the maximum
worst case power consumption for every subsystem in the server. This number is not exceeded by a
functioning server for any combination of hardware and software under any conditions.
• Marked Electrical Power—The server Marked Electrical Power is the rating given on the chassis
label and represents the input power required for facility AC power planning and wiring
requirements. This number represents the expected maximum power consumption for the server
based on the power rating of the bulk power supplies. Also, this number can safely be used to size
AC circuits and breakers for the system under all conditions.
• Typical Maximum Power, User Expected Maximum Power, or Typical Configuration—The measured
maximum worst case power consumption. This number represents the larges power consumption
that HP engineers were able to produce for the server with any combination of hardware under
laboratory conditions using aggressive software applications designed specifically to work the
system at maximum load. This number can safely be used to compute thermal loads and power
consumption for the system under all conditions.
Power numbers per configuration are shown below. For more detail information on power
consumption details for these servers, please refer to the HP 9000 rp7440 or rp8440 Server
Installation guide at http://docs.hp.com.
Typical value represents the expected power consumption of a given configuration. The typical value
is the approximate power consumption that you most likely experience. It can be used for power
budgeting purposes.
Following are the typical power consumption values for two different configurations of each server.
HP 9000 rp7440 Server fully loaded configuration with eight PA-8900 processors—Consists of eight
1.068-GHz PA-8900 processors, 64 GB of memory, 14 PCI-X cards, two cell boards, four internal
hard drives, one DVD drive, two core I/O card sets, and two bulk power supplies.
26
– Typical maximum power: 2120 VA (2078 W) (10.6 A @ 200 VAC across two cords)
– Marked Electrical for the server: 2640 VA (12A @ 220 VAC across two cords)
– Marked Electrical per line cord: 1320VA (6A @ 220 VAC across each cord)
– Maximum Theoretical Power: 3130 VA (3092 W)
HP 9000 rp7440 Server average configuration with four PA-8900 processors—Consists of four 1.068
GHz PA-8900 processors, 10 GB of memory, five PCI-X cards, two cell boards, two internal hard
drives, one DVD drive, one core I/O card set, and two bulk power supplies.
– Typical power consumption: 1080 VA (5.4 A at 200 VAC across 2 cords)
HP 9000 rp8440 Server fully loaded with 16 PA-8900 processors—Consists of 16 1.068 GHz PA8900 processors, 128 GB of memory, 16 PCI-X cards, four cell boards, four internal hard drives, two
DVD drives, two core I/O cards, and six bulk power supplies.
– Typical maximum power: 3,866 VA (3,789 W) (19.33 A @ 200 VAC across 2 cords)
– Marked Electrical for the server: 5400 VA (30A @ 180 VAC across 2 cords)
– Marked Electrical per line cord: 2700 VA (15A @ 180 VAC across each cord)
– Maximum theoretical power: 5,837 VA (5,720 W)
HP 9000 rp8440 Server average configuration with eight PA-8900 processors—Consists of eight
1.068 GHz PA-8900 processors, 16 GB of memory, eight PCI-X cards, two cell boards, two internal
hard drives, one DVD drive, two core I/O cards, and four bulk power supplies.
– Typical power consumption: 1830VA (9.15 A at 200 VAC across 2 cords)
Performance and scalability
The HP 9000 rp7440 and rp8440 Servers continue to provide leading-edge performance and
scalability to the midrange server market. After factoring in racking density, HP midrange servers
maintain performance density at least double that of the competition.
The following section summarizes the main performance and scalability characteristics of the HP
9000 rp7440 and rp8440 Servers.
Important speeds and feeds
HP 9000 rp7440 Server
HP 9000 rp8440 Server
Crossbar bandwidth per cell (peak)
NA
34.5 GB/s
Cell controller to I/O subsystem bandwidth (peak)
11.5 GB/s for each cell
11.5 GB/s for each cell
I/O slot bandwidth (peak)
23 GB/s
46 GB/s*
Memory bus bandwidth (peak)
34 GB/s
68 GB/s
27
Scalability
HP 9000 rp7440 Server
HP 9000 rp8440 Server
Cell boards
1–2
1–4
PA-8900Processors
1–8
1–16
Memory
2–128 GB
2–256 GB
Hot-plug PCI-X I/O slots
15 slots
16 slots
Partitions
1–2
1–2 (4*)
Hot-plug internal disks
0–4
0–4 (8*)
Removable media
0–1 (2 with low profile DVD)
0–2 (4*)
*Including SEU-2
Racking
Both the HP 9000 rp7440 and rp8440 Servers provide industry-leading performance density and
availability in a racked configuration. At 10 EIA units (44.45 cm [17.5 inches]), up to four HP 9000
rp7440 Servers can be mounted into a single HP 2-m cabinet. At 17 EIA units (75.57cm [29.75
inches]) each, two HP 9000 rp8440 Servers can be mounted into a single HP 2-m cabinet, with 7 EIA
units of extra space for mounting external peripherals.
The industrial design and packaging of the HP 9000 rp7440 and rp8440 Servers allow easy and
quick access to all of the system components. The most frequently handled removable media devices
and disks are directly accessible at the front of the system. Removing the front bezel allows complete
servicing of hot-swap fans, hot-swap BPSs, and redundant PCI-X power supplies. At the rear, core I/O
and more hot-swap fans are directly accessible.
The rack-mount kit for the HP 9000 rp7440 and rp8440 Servers consists of the following:
• Slide kit—Slides that safely support a fully configured product that must be serviced from the top
and sides, so the slides must extend away from the rack front columns
• Anti-lock mechanism—A method of preventing sliding of more than one server from the rack at a
time to prevent tipping over
• Ballast weight (only required for HP RBII rack, ordered separately using order number J1479D)—A
method of safely counterbalancing the HP Rosebowl II (RBII) rack enclosure when the product is
deployed for servicing
• Anti-tip foot (only required for HP 10000 G2 Series Rack, ordered separately using order number
AF064A)—A method of safely counterbalancing the HP 10000 G2 Series Rack when the product is
deployed for servicing
• Cable management arm—A design to manage cable bundles at the rear of the product as the
product slides
For access to all other components, the rack-mounted versions come with rack slides. These slides
enable the servers to slide forward out of the cabinet for servicing of internal components such as
fans, cell boards, and I/O cards—even while the system is still operating.
The slides also allow for servicing or replacement of any field replaceable unit (FRU) without removing
the chassis from the cabinet. In fact, you can access and remove any FRU within 15 minutes or less.
This design reduces the downtime associated with system upgrades in the rare event of a component
failure.
28
The interlock is a safety feature that is used when more than one product is to be installed in a single
rack. This feature works only when two products are mounted on slides adjacent to each other (above
and below) in a rack. The function of the anti-lock mechanism is to prevent sliding of more than one
server from the rack at any given time. In the absence of this mechanism, if two servers are pulled out
at the same time, the rack can tip over.
For stability during servicing, ballast kits are used on HP cabinets in which the HP 9000 rp7440 or
rp8440 Servers are installed. The ballast weight base mount is designed to counteract the movement
of one product in an HP RBII rack when extended out on its slides for servicing. The ballast weight
also works in conjunction with the interlock mechanism to counteract the movement of one or two
products in an HP RBII rack being extended for servicing. Every system that HP ships, excluding
pedestal models, includes a ballast kit. The ballast weight easily attaches to the rear anti-tip foot that
comes standard with every HP Rack System cabinet. Use of the ballast kit is mandatory for safety and
for warranty validation, and the kit should be installed immediately.
Also included with every HP 9000 rp7440 and rp8440 Server is a cable management arm (CMA).
The cable management arm is a two-member trough system that resides at the back of the chassis and
guides cables during extension and retraction of the product. The CMA neatly secures data cables
and prevents them from becoming entangled while the system is being serviced.
Third-party racks
HP servers are designed to maximize performance density when installed into HP system cabinets.
Moreover, HP system cabinets maintain the high level of safety and reliability that you have come to
expect. Although HP strongly recommends racking in HP cabinets, some circumstances might prohibit
this. Therefore, HP has developed guidelines that enable safe, reliable HP server installations in thirdparty cabinets. Because of the wide variety of cabinets in the marketplace, it is extremely important
that the guidelines be followed explicitly. See Chapter 4 in the HP Server Configuration Guide for the
racking guidelines.
Server virtualization
Virtualization means that the physical resources are separated from the logical view of the server
infrastructure. Server virtualization helps administrators to improve the usage and simplify the
management of single or multiple server environments by configuring them as reusable pools of
resources. The HP Virtual Server Environment (VSE) for the HP 9000 rp7440 and rp8440 Servers
does just that—sharing, pooling, and allocating of server resources to optimizing utilization, adapting
to changing IT needs while driving down cost.
With the HP VSE, you can increase your utilization of the HP 9000 rp7440 and rp8440 Servers,
maintain nearly continuous service levels, and pay for only what you use. The HP VSE consists of tools
for intelligent control, features, and tools for partitioning the server and availability and provides the
flexibility of utility pricing.
The HP 9000 rp7440 and rp8440 Servers and the HP VSE offers the broadest range of hard and
virtual partitioning solutions in the industry to provide data center security and uptime, while
simultaneously maintaining the highest degree of flexibility. Resource virtualization can be done at the
server, hard partition, or soft partition level. The two strategies for partitioning available to you are
hard partitioning (nPars) and virtual partitioning (vPars). Each partitioning strategy splits the resources
in the server (processors, memory, and I/O) into instances that can each run an operating system
instance.
29
Hard partitions
Hard partitions (nPars) are electrically isolated hard partitions with security provided in the hardware.
Facilitated by the MP and secure firmware, you can configure the sx2000 fabric to isolate resources
in an nPar from the rest of the system, which creates a hardware firewall to prevent other operating
system instances from disrupting that partition. The firewall also reduces the chance that a single
failure (in hardware or software) can take down multiple partitions. This feature is required to perform
hardware maintenance on one partition while the other partitions continue to operate.
The size of nPar partitions can range from a single cell board to the entire system. nPars are available
in every HP 9000 rp7440 and rp8440 Server with no additional purchase or licensing requirements.
vPars
The other partitioning strategy, vPars, runs within a hard partition as shown in Figure 19. Available
on HP-UX 11i, vPars offer partitioning granularity down to the processor core level and the
performance necessary for key applications because there is minimal overhead to coordinate among
the guest operating environments. vPars also provide software fault isolation and security with the
ability to dynamically move processor cores from one vPar to another within the same physical server
or nPar, enabling you to dynamically allocate processing cores to and from the partition, scaling it
with your usage.
Figure 19. The HP Partitioning Continuum for the HP 9000 rp7440 and rp8440 Servers
nPartitions
Hard partitions
within a node
Single Physical
Node Single OS image per
node within a cluster
•
•
•
•
Dedicated CPU,
RAM & I/O
nPar 2
OS image with HW fault
isolation
•
•
vPar 1
•
•
•
•
Dedicated CPU,
RAM & I/O
nPar n
OS image with
HW fault isolation
Application stacking allowing for
higher utilization of resources
Hard Partition
nPar 1
OS image with
HW fault isolation
nPar 3
Node
Virtual Partitions running on Hard
partitions within a node
OS + SW fault isolation
Dedicated CPU, RAM
Application 1
•
vPar 2
OS + SW fault isolation
Dedicated CPU, RAM
Application 2
•
Hard Partition
vPar 1
•
•
Guaranteed compute
resources (shares or
percentages)
Application 3
OS + SW fault isolation
Dedicated CPU, I/O &Memory
vPar 2
•
•
compute resources
(shares or percentages)
Guaranteed
OS + SW fault isolation
Dedicated CPU, I/O & Memory
Dedicated CPU
RAM & I/O
Isolation
Application n
•
Guaranteed compute
resources (shares or
percentages)
Flexibility
HP Essentials Global Workload Manager
HP Essentials Global Workload Manager (gWLM) is the intelligent policy engine for the multisystem
HP VSE. HP gWLM improves the utilization of resources in a multiserver, heterogeneous rp7440,
rp8440, and other HP Integrity and HP 9000 servers using HP-UX 11i, OpenVMS, and Linux
platforms. HP gWLM flexes hard partitions by activating and deactivating temporary instant capacity
30
processors and the virtual server resources that can run within the server hard partitions (resource
management groups, soft partitions) across an entire multitier application based on the demand
generated at the front end. By ensuring that each service can access the resources it requires to meet
its service levels, HP gWLM reduces the complexity and management challenges associated with
multitier applications and web services.
For more information on virtualization and the HP VSE, see the white paper HP Virtual Server
Environment for HP Integrity and HP 9000 servers—Optimize server utilization in real time.
HP Instant Capacity
With HP iCAP solutions running on HP-UX, the HP 9000 rp7440 and rp8440 Servers can be fully
populated with high-performance Dual-Core PA-8900 processors at a significantly lower cost. It is no
longer necessary to fully pay for inactive processors—with iCAP, you only pay when you start using
the extra capacity. These additional processors can be activated with a simple command, providing
immediate increases in processing power to accommodate application traffic demands.
iCAP is also an availability feature. In the unlikely event that a processor fails, the HP system
automatically replaces the failed processor on the cell board at no additional charge—without
rebooting. In online mode, the iCAP processor brings the system back to full performance and
capacity levels, reducing downtime and ensuring no degradation in performance.
HP Temporary Instant Capacity (TiCAP) is the ability to turn iCAP processors that are already installed
in the system on and off for short periods to provide added capacity. TiCAP gives you the ability to
adjust to unplanned or planned spikes in computing.
Cell board iCAP extends the iCAP value proposition to include a complete standby cell board
(processors and memory) in the system for a fraction of the cost. When processing capacity is
needed, simply activate the cell board (memory and at least one processor) to immediately increase
the compute power of the server.
HP iCAP tools use the following algorithm to activate new processors:
1. Verify that there is at least one active processor per cell board.
2. Activate processors round robin style across cell boards within a partition—the number of active
processors per cell board differs by, at most, one across the partition.
3. Enable processors on a cell board in the order 0, 1, 2, 3 to spread the processors across the two
internal cell controller buses and allocate processors in the best thermal fashion.
4. When a failed processor is replaced, choose one from the same cell board when possible; if that
is not possible, choose the next available processor, following rules 2 and 3.
For increased performance, all cell boards in the same partition should contain the same number of
active processors. For high availability reasons, each cell board should contain at least two active
processors.
High availability
High availability (HA) continues to be the hallmark of HP computer systems. But HP knows that
delivering solutions that fully enable the highly available 24 × 7 operations demanded of today’s
businesses requires more than just delivering laundry lists of unusable HA features—or HA features
with limited utility. The high availability features of the HP 9000 rp7440 and rp8440 Servers address
the real causes of downtime, as determined by actual field data from midrange computer users.
The HA features of the HP 9000 rp7440 and rp8440 Servers can be classified as those that address
per partition reliability and those that address intra-partition reliability—that is, single points of failure
between hard partitions.
31
Partition reliability
The HP 9000 rp7440 and rp8440 Servers have a design that is significantly “hardened” over other
systems in their class. In fact, many of the features in these midrange systems can only be found in
mainframes (or in HP Superdome). The reliability features within each HP 9000 rp7440 and rp8440
Server partition have been field proven to provide high system reliability. And many users who have
used these features report significantly lower hardware failure rates than with competitive systems.
Processor protection
The processor is often a major cause of system downtime. For instance, processor cache errors are
demonstrated to be a large contributor (in many cases, the greatest contributor) to unplanned system
downtime. Furthermore, addition or modification of processor resources is among the highest ranking
causes of planned hardware downtime. But in the HP 9000 rp7440 and rp8440 Servers, HP has
designed specific features to combat processor caused downtime, including the following:
• Full error checking and correcting (ECC) on all caches
• Automatic deconfiguration of “faulty” processors (known as Dynamic Processor Resilience [DPR]) on
servers running HP-UX
• A highly effective and reliable processor cooling scheme
• Processor hot-spares using HP iCAP on servers running HP-UX
• Redundant processor power converters
ECC on caches
The processor caches in the HP 9000 rp7440 and rp8440 Servers are fully protected from single-bit
hard errors and random soft errors generated from cosmic rays or other intermittent error-generation
sources. Some competitive systems in the same class are not similarly protected, resulting in errors that
are hard to debug and that are, in many cases, blamed on the customer environment. Such cache
errors in these unprotected systems can result in failures that bring down multiple partitions.
Another advantage of the HP 9000 rp7440 and rp8440 Server processor cache is its layout, which
significantly reduces the chance of a multibit error due to a random cosmic ray strike. Such attention
to detail is not found in many designs available from other vendors.
Automatic processor deconfiguration
DPR refers to the ability of the system to detect, deallocate, and swap in spare processors online for
processors that are generating an excessive quantity of recoverable cache errors. This feature protects
you against the extremely unlikely event of a double-bit cache error. This is one example of the selfhealing features of the HP hardware. Implementation of this feature results in no downtime or
performance loss. This feature is not currently supported with Windows.
Processor cooling
Heat is the big enemy of electronic components. The two-level cooling scheme of the HP 9000
rp7440 and rp8440 Servers offers outstanding cooling capacity at a nominal cost. The turbocooler
fans of the servers draw air directly into the heatsinks of the processor and cell VLSI. At the extreme
operating ranges of the HP 9000 rp7440 and rp8440 Servers, the turbocooler fans keep
temperatures well below the maximum values allowed. Even though the turbocoolers might not be
required under normal operating conditions, running them ensures that the silicon chips operate at the
lowest temperature, helping to ensure maximum lifetime.
To further improve reliability of the HP 9000 rp7440 and rp8440 Servers, manageability software
monitors the speeds of all fans, including turbocooler fans. The smart fan controller of the HP 9000
rp7440 and rp8440 Servers can detect the first hint of slowdown associated with bearing wear,
ensuring you get plenty of warning before a fan fails.
32
HP Instant Capacity
HP iCAP is a means of adding and removing processors in a partition. With iCAP, you do not need
to worry about:
• Interleaved memory
• Application locked memory
• Server switchovers due to false failures
• Physically handling processors or memory boards
• Rebooting
iCAP is the most reliable means of reducing planned downtime for hardware upgrades.
Memory protection
Main memory failures are the single largest cause of customer downtime. The HP 9000 rp7440 and
rp8440 Servers have several features designed to reduce failures of memory:
• Double Chip Spare (DCS)
• Dynamic memory resiliency (DMR)
• Automatic deconfigure on reboot
• Hardware memory scrubbing
• Industry leadership address/control parity protection
Double Chip Spare
The HP 9000 rp7440 and rp8440 Servers support extensive state-of-the-art error detection and error
correction features. This protection is used on paths within the VLSI components and between the
chipset components and processors, memory, and I/O, including processor buses, memory buses,
I/O and fabric links, and I/O slots.
An enhanced feature in the sx2000 chipset over the sx1000 chipset is the DCS. This capability of the
memory error correcting logic reestablishes chip spare correction after another DRAM in the same
ECC codeword has failed. With DCS, firmware recognizes when the first DRAM has failed and
deletes its bits from the ECC calculations, which enables the ECC logic to correct for a second DRAM
failure in the same ECC codeword. This erasure can be applied to a single DRAM, all DRAMs sharing
a bit of a bus, or all buses of a memory subsystem. This maximizes the coverage of this unique
protection mechanism. This enhanced feature is a result of the HP advance research in memory
technology and does not require more DRAMs per MB than was required for the sx1000 single chip
sparing support. This strategy is more cost-effective than memory mirroring methods for protecting
memory. The primary benefit of this HA feature is that it reduces system downtime to replace failed
DRAMs.
Dynamic memory resiliency
DMR is the ability of the system to de-allocate failed memory pages online. This feature is similar to
DPR; if a location in memory proves to be questionable (that is, exhibits persistent errors), the memory
is de-allocated online with no visible impact. Assuming the HP 9000 rp7440 and rp8440 Servers are
equipped with adequate memory to begin with, it is unlikely that the failed memory might have to be
replaced over the life of the product, resulting in a significant reduction in both planned and
unplanned downtime. DMR is superior to industry available hardware-only techniques because
hardware-only techniques can quickly run out of spares. The HP page deallocation technique solves
this problem, resulting in more spares than can possibly be used over the life of the machine.
Hardware memory scrubbing
Software-based memory scrubbers are limited in function because many operating systems and
applications “lock down” memory, resulting in no possible access. The HP hardware scrubber
33
“cleans” memory without operating system or application knowledge, resulting in much better
coverage.
Protection for I/O
I/O errors are another significant cause of hardware errors and downtime because the number of
I/O cards in a typical system is significant and the I/O cards themselves are a part of the system most
exposed to frequent human interaction in the data center.
To prevent downtime resulting from I/O errors, HP has designed the following features into the HP
9000 rp7440 and rp8440 Servers:
• Online replacement of PCI-X cards
• Hardware firewall of I/O errors to cell
• High mean time between failures (MTBF) for I/O cards
• Separate PCI-X buses for each I/O card
Taken together, these features reduce hardware downtime by at least 20% over similar servers.
HP 9000 rp7440 and rp8440 Server crossbar and I/O backplane protection
The backplane of the HP 9000 8440 Server connects everything together. Because all partitions
share the backplane, high reliability and true domain isolation are very important. The specific
features that address these areas are as follows:
• Highly reliable ASICs—The backplane ASIC is manufactured and tested with a process that results
in 10X demonstrated reliability over comparable chips. This reliability results in virtually zero
backplane ASIC failures in the field.
• Redundant DC-DC converters—The DC-DC converters that power the system backplane chips are
fully redundant, reducing downtime associated with power conversion. (Power conversion is
normally a significant contributor to failure rate.) Redundant DC-DC converters have also been
added for the I/O backplane extending the same concept to the I/O subsystem.
• Full end-to-end error correction and independent partition design—The HP 9000 rp8440 Server
backplane is built from two crossbar ASICs with point-to-point connections. Traffic within a partition
is contained in that partition, so there is no sharing of links in a properly configured system. Each
port of the crossbar chip is fully independent, allowing cells of different partitions to coexist without
affecting each other in any way. In other bus-based systems, all domains participate in the
coherency scheme and share address buses. Therefore, in these systems all domains are linked in
some fashion, resulting in shared failure modes that might crash multiple partitions.
Also, unlike other snoopy coherency systems that must accept and respond to all coherency requests
from all domains, HP 9000 rp8440 Server partitions have hardware firewalls dedicated to
guarding partitions from errant transactions generated on failing partitions. A failure in one HP
9000 rp8440 Server partition does not affect any other partitions.
Finally, all data paths in the fabric are resistant to both random single-bit errors and persistent
single-wire “stuck at” faults. Therefore, the fabric is resilient to any single-bit failure, including pin,
connector, or solder problems.
Reliability in the cabinet infrastructure
In keeping with its focus on maintaining high availability, the HP 9000 rp7440 and rp8440 Servers
include protection against failure within the cabinet infrastructure. The HA features in this area include
true dual AC line cord support and complete resilience to service processor failures.
34
Dual AC line cord support
As described earlier in this paper, the HP 9000 rp7440 and rp8440 Servers can run on one or two
totally independent power sources. Moreover, these two power sources do not need to be in phase or
the same voltage.
Resilience to service processor failures
The HP 9000 rp7440 and rp8440 Server hardware has been designed to enable service processor
failover when redundant core I/O cards are in place. Future firmware and manageability code
releases are expected to allow a secondary service processor to take over for a failed service
processor and also enable a resilient console (on reboot). Future operating system releases may allow
the console to fail over as well.
HP continues to make great strides in implementing features that reduce the time to upgrade
components or diagnose and repair component failures. The HP 9000 rp7440 and rp8440 Servers
were designed with the objective of coupling state-of-the-art diagnostic tools with hardware features to
reduce unplanned downtime. Hot-swap and hot-plug technology is implemented throughout the server,
allowing addition or replacement of components while the system continues to run.
For components that cannot be serviced while the server is running, both servers were designed to
provide access and removal of any FRU within 15 minutes or less. In addition, the HP 9000 rp7440
and rp8440 Servers are loaded with design innovations that greatly simplify servicing.
Enhanced serviceability
The following is a list of enhanced serviceability features:
• Hot-plug functionality for internal disks and PCI-X cards
• Hot-swap functionality for bulk power supplies and cooling fans
• 67.9-cm (26.75-inch) product length from mounting column to the connector mating surface at rear
(approximately 76.2 cm [30 inches] overall with the front bezel sitting forward of the rack column),
leaving 10.2 cm (4 inches) for cable bending in third-party racks
• PCI-X card access from the top of the chassis so that ceiling light enhances visibility
• Access panels as large as possible to enhance the service access area
• Five-sided access to increase accessibility
• Independent access to major FRUs (Any FRU can be accessed and removed within 15 minutes or
less.)
• Access to most commonly serviced components from the front or rear without moving the product in
the rack
Investment protection
The HP midrange server family provides outstanding customer investment protection and lasting value,
with a system infrastructure designed to accommodate several generations of processor upgrades. For
the HP 9000 rp7440 and rp8440 Servers, this means supporting PA-8900 processors today, with the
ability to perform in-the-box upgrades to multiple generations of future Intel Itanium processors. No
other competitor in this arena can offer the investment protection provided by HP midrange servers.
This makes upgrading easy and economical: you simply remove all the cells and the memory
contained within the cells, transfer the memory to the processor cell boards, and install the processor
cell boards into the cabinet. In addition, HP is investing in the Intel Itanium processor, enabling you to
move to the new architecture when you are ready, not when your vendor forces you to. The HP
microprocessor roadmap helps make the HP 9000 rp7440 and rp8440 Servers the safest and fastest
midrange servers in the market.
35
Upgrading the HP 9000 rp7420 and rp8420 Servers
The HP 9000 rp7420 and rp8420 Servers are the current generation of the HP midrange server
family that share several different components with their equivalent HP Integrity rx7640 and rx8640
Servers. These PA-8900 RISC-based servers can be easily upgraded to an HP 9000 rp8440 Server
containing PA-8900 processors. Upgrading the HP 9000 rp7420 and rp8420 Servers is easy:
1. Remove the cell boards.
2. Upgrade system backplane.
3. Plug the cell board into the cabinet.
4. Remove or replace existing core I/Os.
5. Remove or replace I/O backplane board.
6. Upgrade system firmware (as needed).
7. Install desired operating system.
8. Change system bezel to reflect the correct system name and color scheme.
Today’s HP midrange servers are already prepared for the next generation of processors so that they
can stay ahead of tomorrow’s performance demands. HP servers deliver investment protection
through multiple significant in-chassis upgrades.
Ease of management
As the number of servers grows and server farms proliferate, IT professionals have come to realize
that the cost of managing these servers can add up to many times the actual cost of hardware. In the
HP 9000 rp7440 and rp8440 Servers, HP has provided features to make management easier and
less taxing. Add HP management software that interfaces seamlessly with the HP 9000 rp7440 and
rp8440 Servers, and provides lower total cost of ownership (TCO) and higher overall efficiency.
Management processor
Both servers have a dedicated MP that simplifies and extends system management and enhances
serviceability. The MP feature set was designed to reduce the need for the system administrator to be
physically at the system to perform tasks such as diagnostics, system management, or even hard
resets. The MP enables you to remote managed the server, regardless of the system state. It can be
use to power on or off the server, set up the hardware partitioning, view system status and event logs.
The MP also enables remote console and console redirection for the multiple operating systems
running on the server. The MP is embedded into the HP 9000 rp7440 and rp8440 Server core I/O
and does not take up a PCI-X slot.
HP System Insight Manager
HP Systems Insight Manager (SIM) is the central point of administration for management applications
that address the HP 9000 rp7440 and rp8440 Server management requirements. HP SIM delivers
powerful monitoring and control, notifying the administrator of potential hardware or software
problems before they occur. It also provides inventory-reporting capabilities that dramatically reduce
the time and effort required to track server assets. HP SIM provides secure communications as well as
role-based security to ensure that its powerful capabilities are kept secure from unauthorized users.
36
HP-UX
HP Essentials for HP-UX 11i are advanced plug-ins to HP SIM that provides modular, integrated
system management software for complete HP server management. It integrates with many other HPUX–specific system management tools, including the following tools available on HP 9000 servers:
• Ignite-UX addresses the need for HP-UX system administrators to perform fast deployment for one or
many servers. It provides the means for creating and reusing standard system configurations,
enables replication of systems, permits post-installation customizations, and is capable of both
interactive and unattended operating modes.
• Software Distributor-UX (SD-UX) is the HP-UX administration toolset used to deliver and maintain
HP-UX operating systems and layered software applications. Delivered as part of HP-UX, SD-UX can
help you manage your HP-UX operating system, patches, and application software on HP servers.
• System Management Homepage (SMH) is used to manage accounts for users and groups, perform
auditing and security operations, and handle disk and file system management and peripheral
device management. HP Systems Insight Manager allows these tasks to be distributed to multiple
systems and delegated using role-based security.
• HP-UX Kernel Configuration is used for self-optimizing kernel changes. The new HP-UX Kernel
Configuration tool enables users to tune both dynamic and static kernel parameters quickly and
easily from a web-based GUI to optimize system performance. This tool also sets kernel parameter
alarms that notify you when system usage levels exceed thresholds.
• Partition Manager creates and manages nPars for high-end servers. After the partitions are created,
the systems running on those partitions can be managed consistently with all the other tools
integrated into SIM.
• HP-UX 11i Webmin-based Admin is a web-based system management framework that allows a
wide variety of open source Webmin system management modules to be plugged in. HP supports
this tool for the configuration of the HP-UX 11i Apache-based Web Server and the HP-UX 11i
Tomcat-based Servlet Engine.
• HP-UX Bastille is a security hardening/lockdown tool that enhances the security of an HP-UX 11i
UNIX® host. It accommodates the various degrees of hardening required of servers used for webs,
applications, and databases.
• Security Patch Check efficiently improves systems security by performing analysis of file sets and
patches installed on an HP-UX 11i system and generating a report of recommended security
patches.
• System Inventory Manager is for change and asset management. It enables you to easily collect,
store, and manage inventory and configuration information for HP-UX–based servers. It provides an
easy-to-use, web-based interface, superior performance, and comprehensive reporting capabilities.
• Event Monitoring Service (EMS) keeps the administrator of multiple systems aware of system
operation throughout the cluster, and it notifies the administrator of potential hardware or software
problems before they occur. HP Systems Insight Manager can launch the EMS interface and
configure EMS monitors for any node or node group that belongs to the cluster, resulting in
increased reliability and reduced downtime.
• HP Process Resource Manager (PRM) controls the resources that processes use during peak system
load. PRM can manage the allocation of processor, memory resources, and disk bandwidth. It
allows administrators to run multiple mission-critical applications on a single system, improve
response time for critical users and applications, allocate resources on shared servers based on
departmental budget contributions, provide applications with total resource isolation, and
dynamically change configuration at any time—even under load.
• HP-UX Workload Manager (WLM) provides automatic processor resource allocation and
application performance management based on prioritized service level objectives (SLOs). In
addition, WLM allows administrators to set real memory and disk bandwidth entitlements
37
(guaranteed minimums) to fixed levels in the configuration. The use of workload groups and SLOs
improves response time for critical users, allows system consolidation, and helps manage user
expectations for performance.
• HP OpenView Operations Agent provides a fully integrated, single-pane-of-glass management
solution for systems, networks, applications, and databases. A powerful ability to monitor, filter,
correlate, and respond to events enables IT organizations to establish central management control
over their managed environments and improve overall availability and reliability.
• HP OpenView Performance Agent monitors and analyzes the performance of systems and
applications to compare SLOs with actual application performance, and it enables real-time
performance monitoring as well as action on alarm.
• HP OpenView Glanceplus is a powerful system monitoring and diagnostic tool that provides online
performance information, examination of system activities, identification and resolution of
performance bottlenecks, and system fine-tuning.
• HP OpenView Data Protector (Omniback II) provides reliable, high-performance data protection for
enterprise-wide heterogeneous environments without impacting system or application performance.
It centralizes and automates backup and recovery operations and tracks file versions and media to
enable swift recovery of information.
• HP OpenView Network Node Manager (NNM) management station runs on Itanium 2-based HP-UX
servers. NNM provides a powerful network management solution that includes concise, in-depth
views of network devices and their status in an intuitive graphical format. NNM helps network
managers evaluate network performance, pinpoint problem sources, and proactively manage their
networks and network availability.
All other HP OpenView management tools, such as HP OpenView Operations, Service Desk, and
Service Reporter, can collect and process information from the agents running on HP 9000 servers
with HP-UX.
HP Services for the HP 9000 rp7440 and rp8440 Servers
Services and support for each operating system are available at the time the specific operating system
is available or supported by the server.
Evolve your infrastructure confidently with a partner that stands
accountable
When you are ready to take advantage of the performance improvements Itanium-based computing
offers, HP has a full range of multi-OS services to help make the transition as seamless and painless as
possible. HP can help you quickly and confidently introduce HP 9000 systems into your existing IT
environment and capitalize on their potential for your business. HP offers assessment services to
precisely define porting requirements and chart a course to deployment, implementation services to
install and configure equipment rapidly, and education services to provide your staff with the
expertise required to achieve better system performance. Throughout the evolution process, HP
accepts full accountability for delivering on the service commitments that HP and its partners have
made. The HP commitment to your satisfaction does not stop with the transition process itself. The HP
multi-OS support offerings—from simple reactive to comprehensive mission-critical—reduce the risks
associated with downtime after your HP systems are installed. HP is looking ahead to help with your
long-term success by working with leading independent software vendors (ISVs) in both the technical
and commercial markets to tailor their applications to the Intel Itanium 2 architecture thereby
exploiting the full potential of your HP servers.
HP Services delivers end-to-end solutions that offer consistent quality and service levels across multiple
platforms such as UNIX, Windows, OpenVMS, Linux, and NonStop Kernel as well as systems from
38
other well known vendors. With the introduction of the midrange servers based on the Intel Itanium 2
micro-architecture, HP is the only vendor that offers the services to support the implementation of
multiple operating systems (UNIX, Windows, OpenVMS, and Linux) on a single Itanium-based server.
HP Services can utilize its wide range of offerings and its experienced services personnel to help
companies fully exploit the Intel Itanium micro-architecture capabilities while protecting their existing
infrastructures.
Full lifecycle services
HP Services offers a portfolio of services that meet your business requirements across your IT lifecycle.
Whether it’s a discrete short-term engagement or a full-scale project deployment, HP helps you fully
exploit the Intel Itanium architecture capabilities while protecting your existing infrastructures and IT
investments by positioning you to meet your evolving business needs.
• Evaluation and planning—HP Services experts help your company determine the best strategy for
seamlessly integrating Itanium-based systems into your existing infrastructure to improve
performance, reduce costs, and gain control of your IT environment. Working closely with your
team, HP helps you develop an architecture that matches your IT environment to your business
requirements and then creates a detailed migration plan that provides specific recommendations
based upon your IT needs.
• Porting and migration—HP offers flexible porting and migration services that help you retain or
enhance your application functionality, reduce disruptions during migration, and take full
advantage of the distinctive capabilities of the Intel Itanium architecture. HP can work with you to
determine your porting and migration needs, devise a strategy for moving applications to Itaniumbased platforms, and create a detailed migration plan. And HP can manage all the details,
including porting and migrating your applications, migrating your data, integrating your
applications with existing ones, and performing application tuning, if that is what you need.
• IT consolidation—HP offers IT consolidation solutions that help you enhance the use of your IT
resources, achieve new performance and productivity levels, and evolve your IT environment to
meet changes in demand. HP Services personnel help you identify your company’s business and IT
objectives, create an investment justification and architectural blueprint, develop a detailed design
plan that incorporates specifics on the configuration and technologies of your company’s solution,
and deliver a complete, tested infrastructure implementation.
• Deployment—HP Services personnel have the experience and expertise to deliver services that can
help reduce the risks, time, and costs associated with deploying new technology. Acting as your
single point of contact, HP develops and implements a deployment plan—including configuration,
testing, and installation of all equipment—that helps provide the seamless deployment of Itaniumbased solutions into full production.
39
• Ongoing support—HP offers a variety of support levels that cover your entire IT infrastructure and
meet your specific needs. Companies can count on HP Services to help them acquire the assistance
they need to maintain control of and deliver business value through their IT investments. You benefit
from proactive onsite services, defined escalation processes, and rapid fixes. Corresponding HP
services for your storage and network systems provide an integrated support solution for your
environment.
– Hardware and software support: HP has a full spectrum of proven preventive, diagnostic,
remedial, and repair services that include high-quality remote and onsite support, upgraded
hardware service response times and coverage periods, fast and reliable access to HP response
centers for software phone-in assistance, and software updates for selected HP and third-party
products.
– Integrated support: Support Plus and Support Plus 24 are integrated hardware and software
services that are available when you require assistance that complements your internal IT
resources.
– Mission-critical support: To reduce exposure to downtime, HP can deliver mission-critical support
for businesses running critical applications. HP uses proven processes, best practices, and
leading remote support technologies to help provide maximum IT availability and performance.
HP Proactive 24 Service includes proactive advice and assistance that helps you improve the
effectiveness of your IT environment. The Critical Service offering combines proactive and reactive
services, including availability assessments and modeling, benchmarking, performance
optimization, remote monitoring, security, capacity planning, configuration, availability, and
performance management services. No other vendor matches the HP worldwide standard sixhour call-to-repair commitment (subject to specific terms and conditions). In addition, through the
HP Mission Critical Partnership, you have the opportunity to create a custom agreement with HP
to achieve your business objectives through customized business-level commitments that reduce
exposure to abrupt negative acts.
• Education—HP offers a full curriculum of education courses that can help improve the productivity
and performance of your workforce. Through these courses, programmers and IT administrators can
quickly become in-house experts on the Intel Itanium architecture and the operation of Itanium-based
solutions for HP-UX, Windows, OpenVMS, Linux, and NonStop Kernel. Courses are delivered using
a variety of methods—from customized onsite classes to self-paced, web-based sessions—that meet
your company’s learning requirements.
• Technical services—Even the most technically proficient IT staff might need assistance when
implementing and managing its Itanium-based platforms, which is why HP offers technical
services—discrete, focused engagements that provide on-demand expertise to supplement your own
IT resources. HP delivers proactive technical consulting services across the entire IT lifecycle, from
assessment and planning to high availability and performance to security and system
administration.
• Business continuity services—HP business continuity offerings—including consulting, disastertolerant, business recovery, and backup and recovery services—can protect your critical high-value
business processes against potentially serious outages. HP uses a proven best practices-based
methodology to design and implement a solution and continuity plan that fits your company’s
specific business needs.
• Managed services—HP offers comprehensive managed services that deliver innovative on-demand
and traditional outsourcing services to simplify the management of your IT infrastructure and reduce
costs while helping you quickly address changes in the marketplace. Our proven, flexible approach
enables you to outsource your entire IT infrastructure—hardware, processes, and people—with the
knowledge that you can receive predictable support levels based upon approved service level
agreements.
40
The HP difference
HP Services provides a full range of services to help companies like yours quickly and confidently
introduce Itanium-based systems into their IT infrastructures. Through our world-class methodologies,
proven processes, IT expertise, advanced support technologies, and partnerships with industry
leaders, HP Services can help you gain the full benefits from this technology—better resource
utilization, reduced costs, and improved return on IT investment.
HP provides consistent delivery of its services through a global network of operations, education, and
competency centers. With more than 69,000 services professionals operating in 170 countries,
acknowledged technology leadership, and a heritage of innovation in services, HP Services can point
to an extensive track record of helping customers improve their ability to support their changing
business needs.
Proven experience and expertise
Our global network of services personnel have unmatched experience and expertise deploying go-tomarket solutions using best-in-class processes across the lifecycle. And no matter what services you
choose—from assessment or porting and migration to complete deployment and education offerings
to ongoing support or full outsourcing solutions—HP can help you take advantage of this nextgeneration architecture quickly and cost-effectively.
To learn more about HP Services, see http://www.hp.com/hps/support.
41
For more information
For more information, see:
• The HP 9000 server family overview website at http://www.hp.com/go/hp9000
• The HP 9000 rp7440 Server website at http://www.hp.com/go/rp7440
• The HP 9000 rp8440 Server website at http://www.hp.com/go/rp8440
© 2004–2007 Hewlett-Packard Development Company, L.P. The information
contained herein is subject to change without notice. The only warranties for HP
products and services are set forth in the express warranty statements
accompanying such products and services. Nothing herein should be construed as
constituting an additional warranty. HP shall not be liable for technical or editorial
errors or omissions contained herein.
Intel and Itanium are trademarks or registered trademarks of Intel Corporation or its
subsidiaries in the United States and other countries. Microsoft, Windows, and
Windows NT are U.S. registered trademarks of Microsoft Corporation. UNIX is a
registered trademark of The Open Group.
4AA0-0137ENW, January 2007