The document summarizes a test conducted by The Tolly Group to evaluate the power consumption of Mitel's Unified Communications Platform. The test measured the watts used by Mitel IP phones and network controllers in different enterprise scenarios. A large system supporting 1,250 users took 3,885 watts, while a medium setup with 350 users across 20 sites used 1,237 watts. Even a small office of 65 users required only 181 watts according to the projections.

1. T H E
TOLLY
GROUP
No. 208345 October 2008
Mitel Test
Unified Communications Platform Summary
Enterprise Power Consumption Evaluation
Premise: Mitel has factored in power
consumption considerations and envi- Test Highlights
ronmental concerns into the functional
design requirements of its products. As Uses less than 4,000 watts of power for a large system of 1,250
a result, the power consumption of the IP phone users split across two resilient MXe controllers at HQ,
company’s most popular IP phones is 200 users at a regional site and 50 users at a branch office
amongst the lowest in the industry. The
implications of reduced power con- Uses 1,237 watts to support a medium network of 350 IP users
sumption include lower costs for cool- spread across 20 sites supported by eight CX controllers
ing, higher phone density per Power Uses just 181 watts to support 65 IP users in a small office with
over Ethernet Layer 2 switch, and a a single CX controller
reduced need for conduit, power distri-
bution upgrades and a cascading reduc- Consumes from 1 to 3.2 watts depending upon IP phone in
tion of power requirements throughout use, even with large-screen backlit display units
the data and voice infrastructure.
M
itel commissioned The Tolly Energy Consumption of the Mitel Unified Communication
Group to evaluate its Mitel® Platform in Various Enterprise Scenarios
3300 IP Communications Plat- As extrapolated from key power measurements
form (ICP) to measure the energy
required to drive critical Unified Com-
munications solutions situated in campus
4,000
headquarters, regional-office and branch- 3,885
Power consumed (Watts)
Large
office locations.
Medium
The Tolly Group measured the power 3,000
consumed (total watts) by a variety of Small
Mitel IP voice switches and IP phones.
Engineers then computed the energy 2,000
requirements to support three enterprise
scenarios — a large-scale network with 1,237
headquarters, regional office and a branch 1,000
office; a medium network with 350 users 181
distributed across 20 locations, and a self-
contained single-site, small branch office 0
Large network Medium network Small network
of 65 users. The computations factored in
scenario scenario scenario
voice switches, phones and the power
required to support them. Note: Please see Test Methodology section for a detailed description of each
network scenario.
Testing was conducted in October 2008.
Source: The Tolly Group, October 2008 Figure 1
© 2008 The Tolly Group Page 1

2. Mitel Mitel 3300 ICP System
Executive Average Energy Usage of Mitel
Summary IP Phones and Controllers
Mitel s Unified Communi- Model Watts consumed
cations Platform demon-
Mitel 5201 IP phone 1
strated very low power
draw on five of its IP
Mitel 5312 IP phone 2.4
phones and two flagship
network controllers. Mitel 5324 IP phone 2.4
Extrapolations using that
data show the Mitel Mitel 5330 IP phone 3.2
solution required under
4,000 watts to support IP Mitel 5340 IP phone 3.2
controllers and phones
across a 1,500 user Mitel ICP MXe controller 52
network, and just 181 (with single dual T1/E1 module)
watts to support a small,
Mitel ICP CX controller
65-user branch office. (with Analog Main Board and 37
Mitel said its IP phones and dual DSP module)
its MXe and CX controllers
are efficient when it comes to Figure 2
Source: The Tolly Group, October 2008
power consumption. Tests
validated this claim, showing
that Mitel’s IP phones tested projection of energy consumption The Mitel 5201 IP Phone, an entry-
draw anywhere from 1 watt across a large-scale deployment of level, desktop IP phone, consumed
to 3.2 watts of power, even 1,500 users — approximately a single watt of power.
for executive phones with 1,250 stationed at a central head-
large, backlit displays. The Mitel 5312 IP phone, a dual
quarters location, about 200 users port, multi-line display IP speaker-
Further, the company’s MXe at a remote office location and phone that supports both SIP and
and CX controllers, which approximately 50 users at a branch Mitel IP (MiNET) protocols, con-
deliver IP PBX and embed- office. Redundancy was assumed sumed 2.4 watts, as did the Mitel
ded unified communications for call control and application 5324 IP Phone.
functionality from a single delivery.
platform, draw from 37 to 52 A Mitel 5330 IP Phone, a full-
Test results show that two 3300 featured enterprise-class telephone
watts in an idle state. ICP MXe controllers with a dual with a large, backlit graphics
Users can compare these T1/E1 module consumed an aver- display and 24 programmable self-
results to other vendor net- age of 52 watts during a 40-second labeling keys, consumed 3.2 watts.
work scenario computations test in idle state.
to gauge the net power Likewise, a Mitel 5340, a full-
An MXe controller outfitted with a featured enterprise-class telephone
required by Mitel versus dual T1/E1 module and a DSP II
competitors. with a large backlit graphics
module and stationed at the display and 48 programmable self-
regional site consumed an average labeling keys, consumed 3.2 watts.
Results of 50 watts.
Using those results, Tolly Group
On the IP phone front, Tolly Group
Energy engineers tested five IP phone
engineers computed the energy
Consumption: costs for a large-scale configura-
models. (See Test Methodology tion that covered 1,250 IP users at
Large System section.) a headquarters location split across
The Tolly Group evaluated a The Mitel phones used from 1 watt two redundant 3300 ICP MXe
representative sample of to 3.2 watts during idle state. (Note controllers.
voice switches and IP that Mitel says for typical busi- The large-scale network scenario
phones, and then used those nesses, a phone is idle as much as also included 200 users at a
measurements to generate a 96% of the time.)
© 2008 The Tolly Group Page 2

3. Mitel Mitel 3300 ICP System
regional office supported by fitted with a T1 combo module and
a single 3300 ICP MXe a DSP module. Mitel
expanded controller, and 50 Engineers computed that the small 3300 IP
more users at a branch office system scenario would require just
supported by a 3300 ICP CX Communications
181 watts to power the CX Platform
controller. Phones on these controllers and associated IP
controllers can be backed up phones. (See Figure 1.)
by any other controller in this Enterprise VoIP Networks:
specific network configura- Setup & Energy Consumption
tion, including a controller at
an HQ site. Methodology
Engineers tested Mitel 3300 ICP approach was similar to power
An extrapolation of the consumption measurements in
energy consumption results MXe and CX controllers, plus five
IP phone models. Tolly Group report 208267.)
from product tests shows that
the Mitel large-scale network The controllers were configured as To measure the power consump-
would consume a total of an IP PBX and acted as a server tion of the MXe and CX control-
3,885 watts across the range that handles embedded unified lers, engineers connected the con-
of controllers and IP phones communications-class applica- troller to a California Instruments
deployed. tions. Both the MXe and CX 4801L AC power source, set the
controllers were running Release AC source to the required AC out-
Energy 9.0 software. put voltage and frequency, and
Consumption: then energized (connected) the AC
For the large network scenario, the
Medium System headquarters site was provisioned
source output. Engineers observed
the power state during controller
Tolly Group engineers also with two MXe controllers, each of boot-up and noted when the power
projected the energy costs to which carried traffic and backed draw settled to a steady state by
support a medium-sized up each other. They were config- observing readings from the front
business configuration ured with two dual T1/E1 modules, panel of the power source.
supporting 350 users spread while the regional office would
across 20 sites. The scenario have an MXe outfitted with a dual For the large network scenario
was supported by eight CX T1/E1 module plus a DSP II mod- extrapolation of 1,500 users, the
controllers, providing 1:1 ule (digital signal processing). percentage of IP phones factored in
redundancy where every Additionally, there was a single was as follows: 20% Mitel 5201
phone is backed up by an- CX controller to support employ- 25% Mitel 5312, 25% Mitel 5324,
other controller and a T1 ees in a branch office. 20% Mitel 5330, and 10% Mitel
combo trunk module for 5340. Together, the IP phones
For the medium network scenario, accounted for 3,690 watts of
WAN links. It was assumed eight CX controllers were used,
that there were two dedicated energy.
two of which were configured with
CX controllers at two a T1/E1 module. All of the CXs Two MXe controllers were fac-
regional sites, and there was had analog main board (AMB) tored in for the headquarters, and
one controller per three of modules. one for a regional site, together
the smaller branch sites. accounting for 154.3 watts. The
To test the power consumption of addition of a CX controller in a
Using energy consumption each phone, power was provided
data gathered from testing simulated branch office accounted
by a Xantrex laboratory-grade for 40.7 watts, bringing the total
each component, Tolly power supply set to 48 volts DC.
Group engineers calculated controller watts to 195 watts.
This was connected to the phone When added with the 3690 watts
that the Mitel solution would via a power test fixture that also
consume 1,237 watts. (See from the IP phones, the Mitel solu-
passed through the data connection tion required 3,885 watts in total.
Figure 1.) between the phone under test and
the controller. The test fixture, For the medium network scenario
Energy extrapolation of 350 users, the
combined with a Fluke Electronics
Consumption: 8010A digital multimeter, facili- percentage of IP phones factored in
Small System tated the precise measurement of was as follows: 20% Mitel 5201
the IP phone DC input voltage and 25% Mitel 5312, 25% Mitel 5324,
This scenario involved a 20% Mitel 5330, and 10% Mitel
small self-contained office current, the product of which
yielded the power consumed by 5340. Together, the IP phones
with 65 users. The site was accounted for 861 watts of energy.
configured with a single the IP phone. A breakdown of the
3300 ICP CX controller out- watts required for each phone can
be found in Figure 2. (Note: This
© 2008 The Tolly Group Page 3

4. Mitel Mitel 3300 ICP System
Test Bed Diagram for MXe and CX Controllers The Tolly Group is a lead-
ing global provider of third-
party validation services for
Power supply
Test point B vendors of IT products,
components
Layer 2 switch
Layer 2 Switch
and services.
CAT 5e
Test fixture CAT 5e
The company
is based in
Boca Raton,
FL and can be
reached by phone at (561)
Test point A
391-5610, or via the
Internet at:
AC power
Web: http://www.tolly.com,
E-mail: sales@tolly.com
Source: The Tolly Group, September 2008 Figure 4
A total of eight CX control- For the small network scenario One CX controller was factored in
lers were factored in for the extrapolation of 65 users, the per- to serve the branch office site, and
medium network scenario, centage of IP phones factored in it was projected to use 37.6 watts.
representing a net energy cost was as follows: 20% Mitel 5201 When added with the 143 watts
of 376 watts, or 1,237 watts 40% Mitel 5312, 30% Mitel 5324, needed for the IP phones, the Mitel
when added with the energy 10% Mitel 5330, and 0% Mitel small network solution required
required to power the IP 5340. Together, the IP phones 180.6 watts, or 181 watts of energy
phones. accounted for 143 watts of energy. total.
Test Equipment Summary
Vendor Product Web URL
California Instruments Model 4801L http://www.calinst.com
Fluke Electronics 8010A Digital Multimeter http://us.fluke.com
208345-njmupo1-cdb-09Dec08
© 2008 The Tolly Group Page 4