Key findings and conclusions:

� Huawei CloudEngine 6850 TOR switch proved full line

rate throughput for all traffic patterns and content
type with no loss for 48*10GbE and 4*40GbE ports
� Ultra low latency for traffic forwarding in store-and-
Lab Testing forward or cut-through mode switching regardless of
Summary frame size
� High availability by design with redundant and hot
Report swappable power supplies and cooling fans
� Efficient Ethernet Certified (802.3az EEE) verified
June 2012 additional power savings up to 24% (17.3% typical)
Report SR120602

H
uawei Technologies requested Miercom evaluate the
CloudEngine 6850 Top of Rack (TOR) Data Center Switches.
The CE6850 family of fixed configuration switches is available
Product Category: in model CE6850-48S4Q-EI with 48 ports of 10GbE SFP+ interfaces or
model CE6850-48T4Q-EI with 48 ports of 10GBaseT interfaces. Both
models have four 40GbE QSFP+ uplinks.
Data Center
The CE6850 Data Center TOR switches are next-generation energy-
TOR Switch saving Layer 2/3 Ethernet switches. The CE6850 utilizes state of the art
switching technologies and superior throughput and energy efficiency
Vendor Tested: in a compact design perfect for Top of Rack applications in large
data centers.
Tests were conducted on the CE6850 TOR switches including RFC
2544 based tests for throughput, forwarding rate, and latency. Power
efficiency tests were also conducted using the Miercom Certified Green
testing methodology that indexes the industry average power
consumption for products in this class and the ATIS TEERs standards.
These tests were the first in a series of tests Miercom is conducting on
Data Center TOR Switches for all vendors in the market. This phase of

Product Tested: Figure 1: Huawei CE6850 Throughput Tests

100

CloudEngine 6850
(%) (% )

80
Series Switches
T hroughput

60
Throughput

40

20

0
CE6850-48S4Q-EI
CE6850-48S4Q-EI CE6850 CE6850-48T4Q-EI
CE6850 CE6850
48x 10GbE 4x 40GbE 48x 10GbE 4x 40GbE
SFP+ QSFP+ 10GBas e-T QSFP+
Source: Miercom, June 2012

Throughput for the CE6850-48S4Q-EI and 6850-48T4Q-EI TOR Data Center

Switch achieved full line rate all ports for all packet sizes as well as IMIX
traffic on all 10GbE and 40GbE uplink ports IAW RFC 2544.
testing of the Huawei Data Center Switches theoretical maximum for the 52 ports tested. The
focused on Layer 2 throughput, latency and CE6850 switch family is designed for Next
energy efficiency. Generation networks with an aggregate throughput
capacity and forwarding rate in excess of what
Performance the physical port speeds and capacity allow at
this time.
Throughput Tests The CE6850 was tested for
maximum throughput performance, maximum Throughput, forwarding rate and latency tests were
forwarding rate, and packet loss for Layer 2 conducted using traffic flows between the 10GbE
traffic. Fixed rate tests were conducted using interfaces, and then the tests were repeated with
packet sizes of 64, 128, 256, 512, 1024, 1280, traffic utilizing the 40GbE uplinks. The tests
1518, 2176, 9216 bytes. The Ixia XM-12 and conducted confirmed latency; throughput and
Spirent Test Center were used to conduct the forwarding rate were consistent regardless of the
RFC 2544 throughput tests. traffic switching between the 10GbE ports or
through the 40GbE uplinks.
Throughput tests were conducted both for the
10GbE ports and the 40GbE uplink ports. Line Latency Tests The CE6850 switches were tested
rate throughput was obtained for the 48-port for latency performance, at maximum forwarding
10GbE SFP+ model CE6850-48S4Q-EI and the rate for Layer 2 traffic. Fixed rate tests were
48-port 10GbE 10GBase-T model CE6850- conducted using packet sizes of 64, 128, 256, 512,
48T4Q-EI. See Figure 1 on page 1. Non blocking 1024, 1280, 1518, 2176, 9216 bytes. The Ixia
throughput to the uplink and traffic through the XM12 and Spirent TestCenter were used to
uplink were also measured at full line rate. conduct the RFC 1242 and 2544 for measuring
latency. See Figure 3 on page 3.
Full line rate performance for all ports and all
interfaces was obtained in testing. For all fixed The test and measurement equipment calculates
frame sizes and jumbo frames of 9216 bytes, the latency by transmitting frames for 30 seconds.
switch forwards all ports of traffic at full line rate Frames are tagged once a second and during half
speed without a single frame lost. See Figure 2. of the transmission duration, then tagged frames
Tests confirmed an aggregate throughput are transmitted. The receiving and transmitting
capacity of 640 Gbps and a forwarding rate of timestamp on the tagged frames are compared.
1,904.8 Mpps, both measurements are the The difference between the two timestamps is the
latency. The test uses a one-to-one traffic
mapping. For store and forward DUT switches
Figure 2: CE6850 Throughput by Frame Size latency is defined in RFC 1242 as the time interval
Compared to Theoretical Maximum starting when the last bit of the input frame reaches
100% the input port and ending when the first bit of the
100% 100% 100% 100% 100% 100% 100% 100% output frame is seen on the output port. Thus
80% latency is not dependent only on link speed, but
Throughput (%)

60%
processing time too.

40% The average latency for the CE6850-48T4Q-EI

20%
was 3.73 sec with very little variation or low jitter.
The latency varied in range from 3.32 to 3.98 sec.
0% The average latency for the CE6850-48S4Q-EI
64 128 256 512 1024 1518 2176 9216
was only 1.80 sec and also had a very tight range
Frame Size (bytes)
of variation between 1.35 and 2.05 sec.
Source: Miercom, June 2012

High Availability Testing

No packet loss was observed in test run with full line rate
mixed traffic mix patterns ingress applied to the CE6850 The CE6850 is designed with resilient hot
Switches. RFC 2544 and models CE6850-48S4Q-EI and swappable components including redundant power
CE6850-48T4Q-EI were tested.

© 2012 Miercom Huawei CE6850 TOR Data Center Switch Page 2

Figure 3: CE6850-48T4Q-EI Latency vs. Frame Size
4.0

3.9

3.8 Latency for the CE6850-

48T4Q-EI TOR Data
3.7 Center Switch for a range
of packet sizes from 64
usec

3.6
thru 9216 bytes. RFC
3.5 2544 was used for the
latency measurements
3.4 and the switch was
3.3
configured in cut through
Max Latency Avg Latency Min Latency mode.
3.2
64 128 256 512 1024 1280 1518 2176 9216
Source: Miercom, June 2012
Legend
64 128 256 512 1024 1280 1518 2176 9216
Max Latency 3.40 3.46 3.56 3.76 3.98 3.88 3.98 3.98 3.98
Min Latency 3.32 3.36 3.48 3.66 3.90 3.80 3.90 3.90 3.88
Avg Latency 3.35 3.41 3.51 3.72 3.93 3.86 3.93 3.93 3.93

Bottom Line
supplies and fans. During throughput tests, we
removed and switched power supplies and fan The Huawei CloudEngine 6850 TOR Data Center
modules. The action was detected by the Switch is designed for demanding Data Center
management console and was reported Environments where high performance and high
accurately. Performance was not impacted by the availability are essential. The superior performance
replacement of the redundant modules of the and energy efficiency of the CE6850 Switches will
switches in any way. No interruption of service or allow Huawei to meet the demand for providing TOR
other anomaly was detected during these tests. solutions for future deployments. With the increasing
demand for port density, line speed and throughput
this TOR switch will continue to provide a scalable
Energy Efficiency Testing solution for Cloud Networking applications.
The CE6850 was tested for power consumption
and energy efficiency using the Miercom Figure 4: CE6850-48T4Q-EI Energy Savings
proprietary standard for calculating Industry Employing 802.3az EEE
Average trends for data center equipment and by 320
conducting tests in accordance with ATIS- 302.6
300 Less is better
0600015.03.2009 Methodology for Energy
Efficiency for Telecommunications Equipment. 280
Normal
Watts

The CE6850-48T4Q-EI demonstrated exceptional 260 250.0 Mode

power savings by employing Energy Efficient 240 EEE

Ethernet Standard (EEE) 802.3az. We verified the Enabled
EEE power savings to be up to 24% with 17.3% 220
savings typical. See Figure 4. At lower utilization 200
levels including idle periods energy savings of up Source: Miercom, June 2012
to 24% were observed. Under heavy load
conditions of 70% utilization or more the energy
savings was reduced to 8%. A 17% savings on average power consumption with EEE
mode enabled on the CE6850-48T4Q-EI TOR. Power
usage based on all ports loaded with IMIX traffic at 30%.

© 2012 Miercom Huawei CE6850 TOR Data Center Switch Page 3

 Test Bed Diagram

Traffic Load Generators Switch Under Test

(SUT)

Ixia XM12
Multifunction
Chassis

Huawei
CE6850 TOR Data
Spirent Center Switches
TestCenter

Source: Miercom, June 2012

How We Did It
The Huawei CE6850 series of switches were evaluated for performance throughput and energy
efficiency. Testing was conducted to verify that the features in this report operated as advertised. The
Huawei CE6850 was running the latest firmware, version 8.5 OS. Two different traffic generators were
used, Ixia XM12 running IxNetwork version 5.50.121.48 and Spirent TestCenter running version
3.95.0236. Energy testing was implemented with Ixia and a Chroma ACSource power controller /
conditioner.

Miercom recognizes Ixia as an industry leader in energy efficiency testing of networking equipment.
Ixia’s unique approach utilizes coordination of energy measurements with network traffic load – allowing
energy consumption to be graphed against network traffic volume. Real-world traffic is generated by
Ixia’s test platform and test applications, principally IxNetwork for Layer 2-3 routing and switching traffic
and IxLoad for Layer 4-7 application traffic.

The tests in this report are intended to be reproducible for customers who wish to recreate them with the
appropriate test and measurement equipment. Current or prospective customers interested in repeating
these results may contact [email protected] for details on the configurations applied to the Device
Under Test and test tools used in this evaluation. Miercom recommends customers conduct their own
needs analysis study and test specifically for the expected environment for product deployment before
making a product selection.

© 2012 Miercom Huawei CE6850 TOR Data Center Switch Page 4

Miercom Performance Verified
The performance of Huawei CE6850 TOR Data Center Switch
was verified by Miercom. In hands-on testing, Huawei
demonstrated advanced performance capabilities including:
� Huawei CloudEngine 6850 TOR switch proved full line rate
throughput for all traffic patterns and content type with no loss
for 48*10GbE and 4*40GbE ports
� Ultra low latency for traffic forwarding in store-and-forward or
cut-through mode switching regardless of frame size
� High availability by design with redundant and hot swappable
power supplies and cooling fans
� Efficient Ethernet Certified (802.3az EEE) verified additional
power savings up to 24% (17.3% typical)

Huawei
CE6850 TOR Data
Center Switches

Huawei Technologies, Co., Ltd.

http://enterprise.huawei.com

About Miercom’s Product Testing Services

Miercom has hundreds of product-comparison analyses
published over the years in leading network trade
periodicals including Network World, Business
Communications Review, Tech Web - NoJitter,
Communications News, xchange, Internet Telephony and
other leading publications. Miercom’s reputation as the
leading, independent product test center is unquestioned.

Miercom’s private test services include competitive product

analyses, as well as individual product evaluations.
Miercom features comprehensive certification and test
programs including: Certified Interoperable, Certified
Reliable, Certified Secure and Certified Green. Products
may also be evaluated under the Performance Verified
program, the industry’s most thorough and trusted
assessment for product usability and performance.

Report SR120602 [email protected] www.miercom.com  Before printing, please

consider electronic distribution

Product names or services mentioned in this report are registered trademarks of their respective owners. Miercom makes every effort to ensure that
information contained within our reports is accurate and complete, but is not liable for any errors, inaccuracies or omissions. Miercom is not liable for
damages arising out of or related to the information contained within this report. Consult with professional services such as Miercom Consulting for
specific customer needs analysis.

© 2012 Miercom Huawei CE6850 TOR Data Center Switch Page 5