Tolly Group - Evaluation of Power Consumption for ShoreTel vs. Cisco

Tolly Group - Evaluation of Power Consumption for ShoreTel vs. Cisco
T
H
E
TOLLY
G R O U P
No. 208267
October 2008
ShoreTel, Inc.
Test
Summary
Unified Communications Systems
Evaluation of Power Consumption vs.
Cisco Unified Communications Systems
S
horeTel, Inc. commissioned The Tolly
Group to evaluate its Unified Communications (UC) System for its power
consumption when running critical IP
Telephony components in campus headquarters, regional office and branch-office
locations.
The Tolly Group measured the power consumed (total watts) by a variety of ShoreGear® IP voice switches and ShorePhone®
IP phones. Engineers then computed the
energy requirements to support three enterprise scenarios — a large-scale network with
headquarters (HQ), a medium network with a
central HQ and 19 remote offices, and a selfcontained single-site small office. The computations considered all ShoreGear voice
switches, phones and associated servers and
the power required to support them.
The Tolly Group compared the ShoreTel
Unified Communications System’s energy
consumption to a Cisco Systems, Inc. solution
used for a comparable deployment. A representative sample of Cisco equipment was
tested for power consumption and those
results were used to extrapolate the data for
the three enterprise scenarios.
Test Highlights
Uses 27% less power — 9,235 watts versus 12,594 watts
versus a Cisco solution in a large enterprise scenario
Uses 62% less power — 1,953 watts versus 5,191 watts versus
a Cisco solution in a medium-sized multi-site deployment
Uses 56% less power — 489 watts versus 1,104 watts versus a
Cisco solution in a small-office deployment
ShoreTel’s monochromatic GbE IP phone consumes 37% less
power than a Cisco equivalent in idle state
Energy Consumption of ShoreTel and Cisco Unified
Communications Systems for Different System Scenarios
As extrapolated from key power measurements
14,000
Power consumed (Watts)
Premise: Enterprises of all sizes are
becoming increasingly conscious of the
energy consumption of their business tools,
including communications systems. Not only
does lower energy consumption decrease
the total cost of ownership, it has a net positive impact on CO2 emissions. Buyers are
advised to factor in the “green” footprint,
as well as the features and functionality of a
unified communications system.
12,000
10,000
8,000
12,594
Cisco
27% less energy
9,235
62% less energy
6,000
5,191
4,000
56% less energy
2,000
1,104
1,953
489
0
Large system
Medium system
Small system
System scenario
Note: Large system was comprised of ~1,500 users across a main headquarters site, a
remote office (200 users) and a satellite office (50 users). Medium scenario included 350
users across one headquarters site and 19 branch offices. Small scenario included 65
users at a single site.
Source: The Tolly Group, September 2008
© 2008 The Tolly Group
ShoreTel
Figure 1
Page 1
ShoreTel
This evaluation showed that
ShoreTel’s Unified Communications System, which
is comprised of the
company’s ShoreGear IP
voice switches and IP
phones, uses up to 62% less
energy than Cisco’s comparable Cisco Unified Communications Manager.
The Tolly Group measured
key VoIP components and
calculated the energy consumption of the solutions
under test for three
enterprise-class scenarios
— a large multi-site business with ~1,500 users, a
medium-sized multi-site
distributed business with
350 users, and a small office
with 65 users. (Note: the
wired LAN/WAN infrastructure was not included
in the configuration except
where that infrastructure
contained VoIP components
such as T1 cards.)
The large system configurations was based on RFP
responses each vendor submitted to the VoiceCon
2008 Telephony RFP.
Minor modifications, noted
below, were made to
Cisco’s large configuration
to provide a consistent
number of VoIP devices for
© 2008 The Tolly Group
Results
Energy
Consumption: Large
System
The Tolly Group evaluated a
representative sample of voice
switches and IP phones, and then
Energy Usage of ShoreTel and Cisco Unified
Communications Systems in a Large System
As extrapolated from key power measurements
14,000
Power consumed (Watts)
ShoreTels Unified
Communications (UC)
System and its ShoreGear
voice switches and
ShorePhone telephones
used up to 62% less energy across three
enterprise-class
scenarios than a solution
based upon Cisco Unified
Communications
Manager.
comparison. (Please see sidebar
on page 8 for more information
on the RFP.) In every scenario,
the ShoreTel Unified Communications System required less energy to power VoIP communications than the Cisco solution.
12,000
Lower bars are better
12,594
10,000
Cisco
ShoreTel
9,235
8,000
7,766
7,743
6,000
4,000
4,829
2,000
1,492
0
Total
IP Phones
VoIP Servers/Switches
Note: Large system was comprised of approx. 1,500 users across a main
headquarters site, a remote office (200 users) and a satellite office (50 users).
Figure 2
Source: The Tolly Group, September 2008
Energy Usage of ShoreTel and Cisco Unified
Communications Systems in a Medium System
As extrapolated from key power measurements
Power consumed (Watts)
Executive
Summary
Unified Communications System
6,000
5,000
Lower bars are better
5,191
4,000
Cisco
3,000
ShoreTel
3,580
2,000
1,953
1,611
1,000
1,600
353
0
Total
IP Phones
VoIP Servers/Switches
Note: Medium system supported 350 users across one headquarters and 19
branch offices.
Source: The Tolly Group, September 2008
Figure 3
Page 2
ShoreTel
Unified Communications System
Energy Usage of ShoreTel and Cisco Unified
Communications Systems in a Small System
As extrapolated from key power measurements
Lower bars are better
ShoreTel Inc.
Unified
Communications
System
Power consumed (Watts)
1,200
1,000
Enterprise VoIP Networks:
Energy Consumption
Cisco
1,104
ShoreTel
800
600
400
618
489
Calibration Caveat
486
378
200
111
0
Total
IP Phones
VoIP Servers/Switches
Note: Small system supported a single site with 65 users.
Source: The Tolly Group, September 2008
used those measurements to
generate a projection of
energy consumption across
a large-scale deployment of
1,500 users — approximately 1,250 stationed at a
central headquarters location with two equipment
rooms, about 200 users at a
remote office location and
approximately 50 users at a
branch office. The system
was fully UC-enabled with
voice, voice messaging, IM
and converged conferencing. It also had a contact
center.
Redundancy was assumed
for call control, voicemail
and for the contact center.
The large system configuration was based on ShoreTel
and Cisco’s response to a
VoiceCon RFP, but the
small and medium system
configurations were based
on typical real-world
deployment scenarios for
the small and medium Enterprise environments.
The Cisco solution specified
725 GbE “worker” phones,
whereas ShoreTel specified
850. For the purposes of the
calculations, 125 phones
© 2008 The Tolly Group
Figure 4
and one server were added to
Cisco’s configuration.
Test results show that a single
ShoreGear-90 voice switch (SG90) used 10.9 watts during the
40-second test in both idle state,
and active state with traffic flowing across ports (voicemails
recorded, voice connections
made, etc).
Testers note that an inconsistent
relationship exists between the
maximum power consumption
figures published on vendor data
sheets and the actual power consumption measured under test.
To date, experience has shown
that actual power consumption is
typically less than or equal to the
stated power draw. The range, as
seen in the current test, can be
anywhere from 25% to 100% of
the published figures.
Engineers based the calibration
factors used in this evaluation
both on measurements made specifically for this study, as well as
on previous experience with
measuring network infrastructure
components.
An IBM System x3250 server,
used by ShoreTel for system
management and voicemail
applications, required 80.7 watts
of energy.
As the calibration factor will
impact the overall results, the
reader should make their own
determination of power consumption by benchmarking their
specific requirements.
The Cisco MCS 7825-H2
appliance, on the other hand, consumed 115 watts during the same
test, which translates into about
42% more power consumed.
Source: The Tolly Group, Sept. 2008
On the IP phone front, Tolly
Group engineers tested four models each from ShoreTel and from
Cisco. (See Setup & Methodology
section.)
The ShoreTel phones used 6.05
watts each during idle state (just
the phone powered, no traffic)
while the Cisco IP phones
required 6.3 watts. During active
state (with an open voice connection), the ShoreTel phones used
6.8 watts each, on average, while
the Cisco phones required 7.05
watts each, on average.
Users who deploy Gigabit
Ethernet to the desktop should
note that ShoreTel’s monochrome
Gigabit Ethernet ShorePhone IP
560G IP phone consumed 4.9
watts in the idle state with backlight off versus 7.8 watts for a
comparable Cisco Unified IP
Phone 7961G-GE. This shows
that this ShoreTel IP phone uses
37% less power than the comparable Cisco IP phone in the idle
state. With the same IP phones
serving an active voice call, the
Page 3
ShoreTel
Unified Communications System
Unified Operations Manager and
Provisioning Manager, eight IP
Contact Center and IP IVR, two
Presence servers, one Unity, one
MeetingPlace, one Mobility
server, six Cisco 2821 Integrated
Services Routers (three at each
HQ site), plus a Cisco 2821 ISR
at the remote office and a Cisco
3825 ISR at the regional office.
ShoreTel phone consumed
25% less power than the
Cisco phone by measuring 6
watts for ShoreTel and 8
watts for Cisco.
Using those results, Tolly
Group engineers then computed the energy costs for a
large-scale configuration
that covered ~1,300 IP
phones, 60 ShoreGear voice
switches, seven ShoreTel
servers, a conference bridge
and a third-party speech
engine. Both Cisco and
ShoreTel assumed that
approximately 200 of the
phones in the large configuration were traditional analog phones. Those were not
counted for the power consumption.
For the Cisco configuration,
engineers factored in four
Cisco Unified Communications Manager appliances, a
ration supporting 350 users
across 20 sites (One headquarters
and 19 branch offices). The scenario for the ShoreTel solution
consisted of a single headquarters
site with 65 users supported by
three SG-50 voice switches providing N+1 redundancy, a VoIP
application server, conference
bridge and an SG-T1 to support
T1/PRI trunks.
Each of 19 branch offices was
supported by an SG-30 voice
switch and 15 users with IP
phones. (See Figure 10.)
Test results show that the ShoreTel UC solution would use 9,235
watts versus 12,594 watts for the
Cisco solution. (See Figure 1 &
2.) This proves that the ShoreTel
UC system consumes 27% less
energy than the comparable Cisco
UC solution in the large system
configuration.
ShoreTel’s ShoreGear voice
switches and IP phones would
require 1,953 watts to support the
HQ and 19 branch office sites. A
comparable Cisco solution would
require 5,191 watts, meaning
users would need 2.7 times more
energy to power the comparable
Cisco VoIP solution. (See Figures
1 and 3.)
Energy
Consumption:
Medium System
Tolly Group engineers also projected the energy costs to support
a medium-sized business configu-
ShoreTel - System Component Power Consumption Summary
Published power
consumption
(Watts)
Power consumed Calibration
as measured
factor
(Watts)
Power consumption value used for
calculations
(Watts)
Description
Model
Basic phone
IP 115
3.7
N/A
0.75
2.8
Worker phone (Mono, FE)
IP 230
4.4
3.5
N/A
3.5
Worker phone (Color, FE)
IP 265
5.9
6
N/A
6
Worker phone (Mono, FE)
IP 560
6.4
N/A
0.75
4.8
Worker phone (Mono, GbE)
IP 560G
7.1
6
N/A
6
Executive color phone (Color, GbE)
IP 565G
6.9
6
N/A
6
Conference phone
IP 8000
8.2
7.3
N/A
7.3
VoIP switch
SG T1
18
N/A
0.35
6.3
VoIP switch
SG 30
23
N/A
0.35
8.05
VoIP switch
SG 50
23
N/A
0.35
8.05
VoIP switch
SG 90
31
10.9
N/A
10.9
VoIP switch
SG 120
63
16.2
N/A
16.2
VoIP switch
SG 220T1A
29
11.3
N/A
11.3
Administration/Messaging server
IBM System x3250
475
80.75
N/A
80.75
Speech recognition server
IBM System x3250
475
N/A
0.17
80.75
Converged Conference bridge
IBM System x3250
475
N/A
0.17
80.75
Source: The Tolly Group, September 2008
Figure 5
© 2008 The Tolly Group
Page 4
ShoreTel
Unified Communications System
Energy
Consumption:
Small System
watts to power the equipment.
The comparable Cisco solution
would require 1,104 watts. The
results show that the ShoreTel
small office configuration provides 56% greater energy efficiency than the Cisco small office
solution. (See Figure 1 and 4.)
This test scenario involved a
small self-contained office
with 65 users.
The single site with the
ShoreTel solution would use
three SG-50 voice switches
for N+1 redundancy for call
control, 66 IP phones including a conference IP phone
and an SG-T1 for PRI trunking.
The small office Cisco configuration would consist of
two Cisco Call Managers for
redundancy, a Unity voicemail server, a Cisco 2821
WAN gateway with T1
access, and 66 IP phones.
ShoreTel’s small office UC
solution would need 489
and Unified Communications
Manager Business Edition 6.1
software suite.
Engineers also tested eight IP
phones with a mix of display and
connection types connected to a
GbE switch via a PowerDsine
3100G single-port PoE injector:
ShorePhone IP 230 and
Cisco IP 7941G (no PoE
capability, tested with AC
adaptor), monochrome display and Fast Ethernet
connection;
ShorePhone IP 560G and
Cisco IP 7961G-GE, monochrome display and GbE
connection;
Setup &
Methodology
Engineers tested the ShoreTel
ShoreGear-90 (SG-90) voice
switch connected to an IBM System x3250 rack-mount server
loaded with Windows Server
2003 and ShoreTel 8 Software.
For Cisco, engineers measured
the power consumption of the
Cisco MCS 7825-H2 Unified
Communication Manager Appliance equipped with a preinstalled Linux operating system
ShorePhone IP 265 and
Cisco IP 7970G, color
display and Fast Ethernet
connection;
Cisco - System Component Power Consumption Summary
Published power
consumption
(Watts)
Power consumed Calibration
as measured
factor
(Watts)
Power consumption value used for
calculations (Watts)
Description
Model
Basic phone
7906G
5
3.2
N/A
3.2
Worker phone (Mono, FE)
7941G
6.3
3.2
N/A
3.2
Worker phone (Mono, FE)
7962G
6.3
N/A
0.5
3.2
Worker phone (Mono, GbE)
7961G-GE
12
8
N/A
8
Worker phone (Color, GbE)
7965G
12
N/A
0.5
6
Executive phone (Color, GbE)
7975G
12
6.5
N/A
6.5
Conference phone
7937G
18.2
N/A
0.5
9.1
Call processing/application server
MCS 7835-H2
450
N/A
0.25
115
Call processing/application server
MCS 7845-H2
800
N/A
0.25
200
Call processing/application server
MCS 7816-H3
420
N/A
0.25
105
Call processing/application server
MCS 7825-H2
450
115
N/A
115
Call processing/application server
MCS 7845-I2
835
N/A
0.25
208.8
Integrated Service Router (Call
server)
3825 Router w/
SRST
360
N/A
1
360
Integrated Service Router (PSTN
gateway)
2821 Router w/ T1
280
N/A
1
280
Integrated Service Router (PSTN
gateway & Call server)
2801 Router w/
SRST
150
N/A
1
150
Analog voice gateway
VG248
72
N/A
1
72
Source: The Tolly Group, September 2008
Figure 6
© 2008 The Tolly Group
Page 5
ShoreTel
Unified Communications System
derive the power consumed
purely by the IP phone alone.
Tolly Group engineers measured
the power consumption of the
various VoIP components and
extrapolated the measured data
into the small, medium and large
system configurations. The key
VoIP components included voice
switches, VoIP servers and IP
phones. Engineers used an Extech
380801 Power Analyzer to measure the watts consumed by the
devices.
ShoreTel’s and Cisco’s RFP
responses for the large system
scenario matched with respect to
number of VoIP phones.
ShorePhone IP 565G and
Cisco IP 7975G, color
display and GbE port.
The ShoreTel Call Manager
was connected to an SG-90
voice switch, which
provided ShoreTel IP
phones with connection
across the GbE network,
and the Cisco MCS 7825H2 Unified Communication
Manager connected to Cisco
Unified IP phones in the
same manner.
Before measuring the
amount of power drawn by
the IP phones, engineers
measured the power drawn
by the PowerDsine 3100G
PoE injector (1.6 watts) as a
baseline. Engineers then
subtracted the baseline from
the overall power consumption of each IP phone to
In the case of the large system
scenario, the ShoreTel response
called for 1290 total VoIP phones
and the Cisco response called for
1165 phones. They differed in the
number of so-called “Worker
phones with Gigabit Ethernet” by
125. In order to make a more
apples-to-apples comparison,
Tolly Group engineers increased
the Cisco GbE phone count to
match ShoreTel. As Cisco’s
original response called for one
server for approximately 100
phones, one server was added to
the Cisco configuration.
The extrapolated data was calculated using a calibration factor
derived from the difference
between the measured power
consumption of each device and
the official published power consumption found in each vendor’s
Web site official data sheet for
each component. The calibration
factor enabled engineers to derive
an effective value, representing
the likely power consumed by the
device. The component summary
tables provide details of the val-
System Component Summary in a Large System
VoiceCon 2008 RFP for 1500-user VoIP System
Vendor
Category
VoIP
gateway
ShoreTel
IP phone
VoIP
server
VoIP
gateway
IP phone
Cisco
VoIP
server
Model
HQ 1 & 2
SG-90
SG-120
SG-220T1A
IP 115
IP 560
IP 560g
IP 565g
IP 8000
ShoreTel Servers
IncendoNet Speech Bridge
Converged Conf. Bridge
3825 Router w/ SRST
2821 Routers
VG248
7906G
7962G
7965G
7975G
7937G
MCS 7835-H2
MCS 7845-H2
MCS 7816-H3
MCS 7825-H2
MCS 7845-I2
6
13
25
50
75
850
110
34
5
1
1
6
3
50
75
850
110
34
10
2
2
1
2
Remote
office
3
4
3
15
15
120
15
8
1
1
15
15
120
15
8
-
Satellite
office
4
2
5
5
25
5
5
1
1
5
5
25
5
5
-
Sub total
13
19
28
70
95
995
130
47
7
1
1
1
7
3
70
95
995
130
47
10
2
2
1
2
Total
60
1,337
9
11
1,337
17
Source: The Tolly Group, September 2008
Figure 7
© 2008 The Tolly Group
Page 6
ShoreTel
Unified Communications System
System Component Summary in a Medium System
350 Users across a Headquarters Site and 19 Branch Offices
Vendor
VoIP
gateway
ShoreTel
IP phone
VoIP
server
VoIP
gateway
Cisco
Model
HQ (1)
Branch
offices (19)
Sub total
SG-T1
SG-50
SG-30
IP 115
IP 230
IP 265
IP 560g
IP 565g
IP 8000
ShoreTel Servers
Converged Conf. Bridge
1
3
6
52
7
1
1
1
19
38
171
76
-
1
3
19
44
171
76
52
7
1
1
1
2821 Router w/ T1
2801 Router w/ Analog
Trunk
7906G
7941G
7961G-GE
7965G
7937G
MCS 7825-H2
MCS 7816-H3
MCS 7835-H2
1
-
1
-
19
19
6
52
7
1
2
1
1
38
171
76
-
44
171
52
83
1
2
1
1
Category
IP phone
VoIP
server
Total
23
351
2
20
351
4
Note: See Figure 10 for graphical representation.
Figure 8
Source: The Tolly Group, September 2008
System Component Summary in a Small System
65 Users at a Single Site
Vendor
Category
Model
VoIP
gateway
SG-T1
SG-50
IP 115
IP 560g
IP 565g
IP 8000
ShoreTel IP phone
Cisco
VoIP
ShoreTel Servers
server
VoIP
2821 Router w/ T1
gateway
7906G
7961G-GE
IP phone
7965G
7937G
MCS 7825-H2
VoIP
VoIP
s
server
MCS 7835-H2
Source: The Tolly Group, September 2008
© 2008 The Tolly Group
Small
office (1)
1
3
6
52
7
1
Total
4
66
1
1
1
1
6
52
7
1
2
1
ues used. (See Figures 5 & 6.)
Given this approach, Tolly Group
engineers could extrapolate a
close estimate to the total power
consumed from all unmeasured
devices for the final results. After
applying the calibration factor an
effective value of maximum
power consumed by the devices
was derived. Using the effective
values for power consumption of
the various components tested,
Tolly Group engineers estimated
the total power consumption from
all devices in the three network
scenarios.
66
3
Figure 9
Page 7
ShoreTel
Unified Communications System
ShoreTel Medium System Configuration
Note: See Figure 8 for detailed
system component summary
SG-50 (Qty 3)
Phones
SG-T1
T1 trunk
Router
PSTN
WAN
Branch offices (X 19)
HQ Server
Conf. Bridge
Router
Headquarters
SG-30
PSTN
The Tolly Group is a leading
global provider of third-party
validation services for vendors of IT products, components and services.
Phones
Analog trunk
The company is
based in Boca
Raton, FL and
can be reached
by phone at (561) 391-5610,
or via the Internet at:
Web: http://www.tolly.com,
E-mail: sales@tolly.com
Figure 10
Source: ShoreTel, Inc., 2007
VoiceCon® RFP
RFP and responses can be found
on the VoiceCon Web site at:
were traditional analog phones
rather than VoIP phones.
Allan Sulkin of TEQConsult
Group developed a comprehensive VoIP system RFP in conjunction with the VoiceCon Orlando
2008 conference held in March
2008. The results were unveiled at
a conference workshop and subsequently published.
The time frame for the product
information contained in the RFP
responses was Fall 2007. The
http://www.voicecon.com/orlando
/about/real_world_rfp.php.
The Tolly Group referred to
VoiceCon RFP for the large
system configuration only. The
small and medium configurations
were derived by ShoreTel, Inc.
(Note: The Tolly Group is not
affiliated with the VoiceCon conference, the TEQConsult Group
and had no part in the creation of
the RFP and makes no claims of
any kind with respect to the RFP.)
As both Cisco Systems and ShoreTel responded to this RFP
request, the gear outlined in those
responses was used as a basis for
this comparison.
Both Cisco and ShoreTel assumed
that approximately 200 of the
phones in the large configuration
Test Equipment Summary
Vendor
Extech Instruments
Product
Extech 380801 Power Analyzer
Web URL
http://www.extech.com
208267-tifsxppe2-cdb-08Oct08
© 2008 The Tolly Group
Page 8
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising