Cisco 300 Series LAN Switch Comparison

#211103
February 2011
Commissioned by Cisco Systems, Inc.
Cisco Small Business 300 Series Managed Switches:
Performance, Power Consumption and Features
vs. D-Link, HP Networking and NETGEAR
Executive Summary
The Bottom Line
Small businesses today require high-performance, high-functionality
switches for both 10/100 and Gigabit Ethernet deployments. Businesses
deploying VoIP solutions will also require switches that deliver Power over
Ethernet (PoE), quality of service (QoS) and Voice VLANs. With respect to
economics, switches should be easy to use, to help lower administrative
costs, offer attractive cost-per-Gigabit and, where appropriate, cost-perWatt of PoE delivered.
Cisco Systems commissioned Tolly to evaluate several models of the new
Cisco Small Business 300 Series of managed switches along with
comparable models from D-Link, HP Networking and NETGEAR. In all,
eleven switches were tested.
Tolly evaluated a range of features and capabilities including: Power over
Ethernet, Layer 2 throughput and latency, power consumption, feature/
functionality and advanced recovery and performance options, such as
link aggregation and MSTP. Engineers also calculated cost-per-Gigabit
and cost-per-Watt-delivered as appropriate.
Small business users have a wide range of LAN switch implementations
to choose from: fixed/stackable/chassis, managed/smart/unmanaged,
Fast Ethernet (10/100)/Gigabit Ethernet, PoE/non-PoE, etc.
The Cisco 300 Series 10/100 and GbE
Managed Switches delivered:
1
2
3
4
5
6
7
Wire-speed, non-blocking, Layer 2 throughput at all
frame sizes tested from 64 to 1518 bytes
Consistently low latency at all frame sizes
Best Price/Performance among switches tested
Most extensive feature set: IPv6, traffic shaping and
rate limiting, scope of GUI-based configuration
Lowest power consumption in 2 of the 3 classes
tested, and best-in-class power efficiency overall
Most extensive set of IPv6 protocol and application
support
Best usability with a simplified user interface
delivering both basic and advanced capabilities in an
intuitive fashion
This evaluation was restricted to fully managed, non-stackable (fixed) switches offering a basic complement of at least 24 ports. Within that class,
switches were further categorized as follows: 1) 10/100 (Fast Ethernet) with PoE, 2) 10/100 without PoE, and 3) Gigabit Ethernet without PoE.
The Cisco Small Business 300 Series features a switch in each of these categories.
All three Cisco 300 Series switches delivered wire-speed throughput and consistently low latency at all frame sizes tested (64-byte through 1518bytes). The power consumption of each switch was calculated using the industry-standard ATIS methodology. Cisco switches ranked first or
second in energy efficiency in every category - in some cases consuming only 25% as much power as competing switches. The Cisco switches
also offer additional energy efficiency capabilities not present in the competing switches such as power scaling on cable length and reduction of
power on ports with endpoints disconnected or powered down.
All three of the Cisco switches tested run on the identical code base, thus, provide support for the same advanced feature set which simplifies
switch maintenance. In addition to supporting features like QoS, MSTP, Layer 3, Voice VLAN, LLDP-MED, DHCP Options 66/67/82, IGMP snooping,
Querier, and security features, such as ACLS and time-based 802.1X, the Cisco boxes were unique in this test for their extensive support for IPv6
as well as sophisticated, configurable rate limiting and traffic shaping features. (Not all features listed were tested as part of this evaluation.)
While all of the systems tested provide a graphical user interface (GUI), the Cisco GUI had the most modern look-and-feel with Ajax-based
features. Testers also verified that, even with the Cisco switch running 100% load on 24 of its ports, that the GUI (accessed in-band via port 25)
remained responsive and that the switch could maintain wire-speed throughput.
© 2011 Tolly Enterprises, LLC
Tolly.com
Page 1of 12
Cisco Systems 300 Series Managed Switches
#211103
As network architects focus more on total
cost of ownership (TCO), the cost of energy
and the recurring costs associated with
running network infrastructure devices
7x24x365 have become increasingly
important.
Introduction
The evaluation included eleven LAN
switches from four prominent vendors and
explored a wide range of areas that included
price/performance, Power over Ethernet,
feature sets and usability.
Cisco Systems, Inc.
300 Series
Switches
LAN Switch:
Performance,
Power
Consumption,
PoE, Advanced
Features
Cisco Systems, apparently, has made energy
efficiency a focus for the 300 Series Switches.
Certain areas, such as feature sets and
usability, typically remain consistent across
the various models of a single vendor. Other
information, such as price/performance, will
be tied to a specific switch.
In this series of tests, the Cisco switches
delivered “best-in-class” power efficiency for
the Gigabit Ethernet and 10/100 PoE
switches and were in a virtual tie for best
power efficiency in the 10/100 Non-PoE
switch category. See Figure 1.
Results will be detailed first for those areas
that generally apply to all products tested
from a given vendor. Then, product-specific
results will be detailed for each of the three
product categories tested.
Additionally, Cisco has implemented
features that can respond to actual, run-time
conditions to reduce power consumption.
Power
Specifically, the energy detect feature
enables the switch to sense whether the
attached device is powered on and to put
In recent years, the importance of how a
device consumes and manages power has
surged.
Tested
December
2010
the switch port into a power-conserving
sleep mode when the device is powered off.
The power scaling feature provides for
dynamic detection of cable length thus
allowing the switch to reduce port power for
shorter cable lengths.
ATIS Power Consumption (Watts)
LAN Switch: Layer 2 Power Consumption (Watts)
24 Ports Benchmarked - ATIS Power Rating
Usage as reported by Watts Up? PRO ES
(Lower numbers are better)
49.3
42.2
23.1
10/100 PoE
Cisco SF300-24P
D-Link DES 3528P
HP Networking E2610-24/12PWR
13.0 14.5 15.2
18.5
12.5
Cisco SF300-24
D-Link DES 3528
HP Networking E2510-24
HP Networking E2610-24PT
NETGEAR FSM726
38.4
26.0
10/100 Non-PoE
47.1
Gigabit Ethernet
Cisco SG300-28
HP Networking E2810-24G
NETGEAR GSM7224
Note: No power over Ethernet was being delivered during this test. The ATIS power is 0.1*(at 0)+.8*(at 10) +.1*(at 100) where (at n) are utilization rates.
HP Networking and NETGEAR GbE switches have 24 ports where Cisco has 28. ATIS values for Cisco used 24 ports for comparison purposes. Even running
28 ports Tolly engineers confirmed that the Cisco ATIS was the lowest of the three Gigabit Ethernet switches.
Source: Tolly, December 2010
Figure 1
© 2011 Tolly Enterprises, LLC
Tolly.com
Page 2 of 12
Cisco Systems 300 Series Managed Switches
#211103
Cisco was the only vendor in this evaluation
that provided these enhanced energy
management features. See Table 1.
D-Link provided more granular control with
per-port rate limiting for ingress/egress and
both per-port and per-queue rate limiting.
Features
The Cisco offering, however, provided the
most extensive set of rate limiting and traffic
shaping options. Unique among the
products in the test, Cisco offers two modes
for traffic management: Basic - providing
per-port functionality, and Advanced adding more granular per-flow control. The
support was very granular and allows
network administrators to manage traffic at
per-port levels and/or at ingress or egress to
the switch. Furthermore, the traffic shaping
policy could be applied to a single port or to
a set of ports belonging to a logical link
aggregation group.
Tolly engineers validated a range of features
on each of the devices and reviewed each
system’s configuration screens to identify
the presence or absence of other functions.
Table 2 summarizes those results and
highlights are discussed below.
Traffic Prioritization
Engineers confirmed that all of the systems
under test supported 802.1p/Q prioritization
and VLAN tagging. Engineers also validated
that all the systems under test provide
support for 802.3ad link aggregation which
allows multiple physical links to function as
a single logical link between switches.
Spanning Tree and Multicast
Support
Engineers investigated, but did not test,
additional proprietary features each vendor
makes available to provide more granular
control over bandwidth.
Tolly engineers validated that all systems
under test supported both multiple
spanning tree (MSTP) and rapid spanning
tree (RSTP) protocols.
Products from HP and NETGEAR offered
some additional control over the traffic
stream with HP providing multiple traffic
queues but no explicit rate limiting and
NETGEAR offered a per-port limit for
transmit rate.
Tolly engineers validated the support for
various multicast functions and verified that
the Cisco solution supported IGMP
Snooping, Querier and MLD Snooping
functions. The HP solutions offered IGMP
Snooping and Querier functions - no MLD
Snooping options were found in either the
GUI or CLI configuration environments.
IPv6 Support
With the IPv4 address set already fully
allocated, IPv6 is a must-have feature for
network equipment in businesses of all
sizes.
While not part of the testing phase, Tolly
engineers used product documentation to
identify the current IPv6 support offered by
each product.
HP’s products did not appear to have any
IPv6 functionality and the D-Link and
NETGEAR products offered minimal support
for IPv6 addressing and management.
In contrast, the Cisco solution offered an
extensive set of IPv6 protocol and
application support. This is the most
extensive IPv6 support that Tolly has seen, to
date, in an SMB-class product.
As noted earlier, a summary of feature
support can be found in Table 2. For Cisco,
the protocol support includes dual-stack
IPv6/IPv4, ACLs, QoS and MLD Snooping
with a full range of IPv6 network
management functions and applications
including Telnet/SSH, RADIUS and DNS.
LAN Switch Energy Conservation Features
Vendor Products Tested
Feature
Cisco Systems
DLink Systems
HP Networking
NETGEAR
Energy Detect
2
2
2
Power Scaling
1
2
2
2
Notes: 1) Available on Cisco Gigabit Ethernet switch only. 2) Switch documentation did not reference these functions or similar functions.
As feature availability is the same across all products tested from each vendor, only the vendor name is referenced.
Source: Tolly, December 2010
© 2011 Tolly Enterprises, LLC
Table 1
Tolly.com
Page 3 of 12
Cisco Systems 300 Series Managed Switches
#211103
Gigabit Ethernet LAN Switch: Layer 2 System Latency (μsec) Under 100% Load
24 Ports in Full Mesh Configuration
(as reported by Ixia IxAutomate 6.90)
Lower numbers are better
10/100 Non-PoE
!
!
!
!
" $
!"
!
$
!$
!$
! " #"
10/100 PoE
Gigabit Ethernet
#
%
!
$
"
#
!
#
!
" $
!"
!
$
!$
!$
!
#
!
#
#
" "
#
#
Note: Cisco’s GbE switch has 28 ports where HP Networking and NETGEAR’S have 24 ports. For purposes of comparison, 24 ports were used on the
Cisco switch. For the HP E2510-24 up to .5% of traffic was dropped at all frame sizes during the 100% load, latency test. For the D-Link DES-3528P
approximately .5% of traffic was dropped at 64-byte frames only.
Source: Tolly, December 2010
© 2011 Tolly Enterprises, LLC
Figure 2
Tolly.com
Page 4 of 12
Cisco Systems 300 Series Managed Switches
#211103
LAN Switch Feature Summary
Based on Vendor Products Tested
Feature
Cisco Systems
DLink Systems
HP Networking
NETGEAR
802.1p/Q
MSTP 802.1s
RSTP 802.1w
Link Aggregation 802.3ad
IGMP Snooping, Querier and
MLD Snooping
Partial, no MLD
Snooping3
(Minimal)
(Minimal)
(Extensive)
Dual Stack IPv6/IPv4
IPv6 ACLs
IPv6 QoS (DSCP)
MLD Snooping
Transition mechanism – ISATAP
IPv61
IPv6 Applications – SNMP, Telnet/SSH, RADIUS, Syslog,
Web/SSL, DNS, etc
VLAN Mirroring
Layer 32
Partial, only on
E2610 model
(Extensive)
Rate Limiting/Traffic
Shaping1
(Partial)
(Minimal)
QoS but no explicit
(Minimal)
Per-port rate limiting for
Basic and advanced modes. Per-port and
rate limiting. Also
ingress or egress, can
per-flow policies and mapping can control
some settings
QoS supports a per-port cap
specify maximum per
ingress/egress/both. Applicable to single
require
for transmit rate.
port and per queue (7)
port or LAG.
configuration via
rate limiting.
CLI.
Notes: 1. IPv6 and rate limiting configuration options reviewed but not tested. 2. Layer 3 does not imply full layer 3 routing functionality. 3. The HP
Networking E2510 does not support data-driven IGMP.
As feature availability is the same across all products tested from each vendor, only the vendor name is referenced.
Source: Tolly, December 2010
© 2011 Tolly Enterprises, LLC
Table 2
Tolly.com
Page 5 of 12
Cisco Systems 300 Series Managed Switches
#211103
Ethernet LAN Switch: Price/Performance
Cost per Gigabit Throughput & Cost per PoE Watt Delivered
Lower numbers are better
Calculations based on US MSRP January 2011
10/100 PoE
$354.16
$185.42
$6.75
$225.83
$2.56
Cost per PoE Watt Delivered
Cost per Gigabit Throughput
Cisco SF300-24P
D-Link DES-3528P
HP Networking E2610-24/12PWR
10/100 Non-PoE
Gigabit Ethernet
$225.83 $229.16
$152.56
$3.23
$129.16
$89.58
Cost per Gigabit Throughput
Cisco SF300-24
HP Networking E2610-24
NETGEAR FSM726
$60.12
D-Link DES-3528
HP Networking E2510-24
$29.87
$17.86
Cost per Gigabit Throughput
Cisco SG300-28
HP Networking E2810-24G
NETGEAR GSM7224
Note: The cost per Gigabit was calculated using the aggregate bidirectional throughput of each switch when transmitting 1518-byte frames at the switch’s maximum rate.
The MSRP is divided by the throughput. The cost per PoE Watt was calculated by dividing the MSRP by the measured power budget.
For PoE, HP recommended using the version of the E2610 (J9087A) that provided PoE on all 24 ports but that switch was not in stock for testing.The cost per PoE watt
delivered would have been $3.20 based on a CDW price of $1,299 and an advertised 406W PoE budget.
GbE uplinks not included in 10/100 cost per Gbps calculation. All 10/100 switches were equipped with 4 GbE uplinks except the NETGEAR FSM726 and the HP E2510 which
had 2 uplinks.For GbE, HP and NETGEAR switches offer 24 ports where Cisco provides 28 ports. All available ports used for cost calculations on the Gigabit Ethernet switches.
Source: Tolly, December 2010
Figure 3
Usability
Graphical User Interface
While usability cannot be measured in the
same way as system performance, it is an
important area when considering total cost
of ownership (TCO) especially in an SMB
environment where dedicated IT resources
might be constrained or even non-existent.
While all managed switches will be able to
claim to offer a “GUI”, this cannot be treated
as a “check mark” item that a product either
does or does not offer.
© 2011 Tolly Enterprises, LLC
The GUI environments offered by the other
vendors were on par with what has been
available in the industry for years. HP
Tolly.com
required a JAVA download to function. The
GUI’s offered minimal “help” functions and
Tolly engineers found that they had to leave
the “easy” GUI environment and use the
command-line interface (CLI). In summary,
the D-Link, HP and NETGEAR represented
the status quo GUI environments that one
has come to expect.
Page 6 of 12
Cisco Systems 300 Series Managed Switches
Cisco has raised the bar on switch
management GUIs. Cisco’s GUI was
designed with new generation technology
and could be seen immediately to provide a
state-of-the-art user experience offering
Ajax-style interaction with the user. More
importantly, the GUI does not just provide
access to basic functionality. Even the
granular traffic shaping functions discussed
earlier can be configured via the GUI. And,
the GUI offers both a “basic” and an
“advanced” screen set for this function so
that network managers concerned only
with basic traffic shaping functions do not
need to deal with screens and functions
related to advanced traffic management
functions. Also, the Cisco GUI offers a very
complete, context-sensitive help not
present in the other products evaluated.
Common Code Base
Additionally, usability needs to go beyond
the traditional definition of the look-and-feel
of the management interface.
While the Cisco switches tested ranged from
10/100 to Gigabit Ethernet with and without
Power over Ethernet, they all share the same
code base. Thus, users of the Cisco 300
family only need to maintain and deploy a
single code base and, consequently, have
identical software functionality in all models.
Of the vendors tested, only Cisco used the
same code base across all of the products in
this test.
#211103
Product-Specific
Comparisons
10/100 Power over Ethernet
Switches
The Cisco SF300-24P was compared with a
D-Link DES-3528P xStack and an HP
Networking E2610-24/PWR. (See Table 4 for
a complete listing of switches evaluated in
this report.)
Both the Cisco and D-Link models tested
were able to deliver PoE on any of the 24
ports. While HP has a model of the switch
tested that has this capability, the model in
stock and used for this test supported PoE
on only 12 ports.
Power Efficiency
The Cisco SF300-24P is the most power
efficient of the three tested with power
usage calculated using the ATIS method that
are roughly half of the HP and D-Link
offerings. See Figure 1.
Tests were run to determine how many
ports could deliver 15.4W simultaneously.
The maximum for HP was 8. Cisco delivered
maximum power on 11 ports and D-Link on
23 ports. See Table 3.
Cost Per PoE Watt Delivered
Using price information available on CDW
and CompSource as of January, 2011, Tolly
engineers calculated the “cost per Watt” of
PoE power delivered.
For Cisco, that value was $2.56 compared
with $3.23 for D-Link and $6.75 for HP. Using
HP’s advertised power budget of 406W for
t h e E 2 6 1 0 - 2 4 / 2 4 P W R s w i tc h H P
recommended for test that provides PoE on
24 ports, engineers calculated the cost per
PoE Watt delivered of that unit as $3.20
based on a CDW price of $1,299.99. See
Figure 3 for details and calculation.
© 2011 Tolly Enterprises, LLC
Tolly.com
Performance
Tolly engineers performed a series of
standard throughput tests using the 24
10/100 ports on each device. See Table 1.
All devices performed at wire-speed for all
frame sizes tested except the D-Link which
dropped some frames when tested at 64bytes.
Engineers also benchmarked and verified
that the Cisco switch exhibited much lower
latency than the other two switches.
With 64-byte frames, the Cisco switch
latency was 10.1μs compared with 1
224.4μs for HP. With the largest frames,
1518-bytes, the Cisco latency was only
127μs compared to 814.1 for D-Link and
345.1μs for HP. See Figure 2 and Table 5.
Cost Per Gbps of Throughput
Engineers related throughput to device cost
by calculating the cost of each Gbps of
throughput. The Cisco switch had the lowest
cost per Gbps at $185.42 followed by D-Link
at $225.83 and HP at $354.16. See Figure 3
for details and calculation.
The cost per Gbps includes only the 24
10/100 ports as those represent the core
capacity of the switch. All switches in this
category provided 4 GbE uplink ports.
10/100 Ethernet Switches
(Non-PoE)
The Cisco SF300-24 was compared with a DLink DES-3528, an HP Networking E2510-24,
and HP Networking E2610-24PT and a
NETGEAR FSM726.
Power Efficiency
The Cisco SF300-24 is the second most
power efficient of the five tested with power
usage calculated using the ATIS method that
is lower than both HP switches and the DLink offering. See Figure 1.
Page 7 of 12
Cisco Systems 300 Series Managed Switches
#211103
Performance
Tolly engineers performed a series of
standard throughput tests using the 24
10/100 ports on each device.
Competitive Interaction
All devices performed at wire-speed for all
frame sizes tested except the HP
Networking E2510 that passed between
99.481 and 99.979% of the theoretical
maximum traffic load at all frame sizes.
Tolly acquired the DLink, HP Networking and
NETGEAR switches via normal product distribution
channels. The Tolly Group invited representatives
from those companies to participate in the testing
as per The Tolly Group’s Fair Testing Charter.
(See http://www.tolly.com/FTC.aspx).
As with the PoE switch category, Cisco
switch exhibited much lower latency when
tested at 100% load than the other switches.
In fact, the Cisco latency was identical to its
PoE results.
While NETGEAR latency was competitive,
the D-Link switch and the HP E2610 had
latency that was some 3x longer than Cisco
at larger packet sizes. (The HP E2510
dropped frames in every latency test and
cannot be compared accurately.) See Figure
2 and Table 5.
Cost Per Gbps of Throughput
Engineers related throughput to device cost
by calculating the cost of each Gbps of
throughput. The Cisco switch had the lowest
cost per Gbps at $89.58 followed by
NETGEAR at $129.16 and the HP Networking
E2510-24 at $152.56. See Figure 3 for details
and the calculation used.
The cost per Gbps includes only the 24
10/100 ports as those represent the core
capacity of the switch. The NETGEAR
FSM726 and the HP Networking E2510-24
provide two Gigabit Ethernet uplinks where
the other switches in the category provide 4
GbE uplinks each.
Gigabit Ethernet Switches
The Cisco SG300-28, a 28-port switch, was
compared with two 24-port switches: an HP
Networking E2810-24, and a NETGEAR
GSM7224.
© 2011 Tolly Enterprises, LLC
All three vendors participated.
HP recommended using the E261024/24PWR, with PoE on 24 ports, but the
recommended switch was not available in the timeframe for this test.
The various vendors reviewed their results and did not dispute the accuracy of the
results.
For more information on the Tolly Fair Testing Charter, visit:
http://www.tolly.com/FTC.aspx
Power Efficiency
The Cisco SG300-28 is the most power
efficient of the three tested with power
usage calculated using the ATIS method that
is lower than both of the other GbE switches.
While the comparisons shown in Figure 1
are based on 24 ports for each of the
products, Tolly engineers confirmed that
even with all 28 ports active, the Cisco GbE
switch still has the lowest ATIS value.
Performance
Tolly engineers performed a series of
standard throughput tests using 24 GbE
ports on each device. See Table 1.
All devices performed at wire-speed for all
frame sizes tested. Additionally, Tolly
engineers re-ran the performance test on
Cisco using all 28 GbE ports and, again, the
Cisco switch delivered wire-speed
throughput.
Tolly.com
While all three GbE switches exhibited low
latency, the Cisco latency was the lowest of
all switches tested for all frame sizes. See
Figure 2 and Table 5.
Cost Per Gbps of Throughput
Engineers related throughput to device cost
by calculating the cost of each Gbps of
throughput. The Cisco switch had the lowest
cost per Gbps at $17.86 followed by
NETGEAR at $29.87and HP at $60.12. See
Figure 3 for details and calculation.
The Cost Per Gbps includes all of the Gigabit
Ethernet ports available on each switch as
those represent the core capacity of the
switch. The fact that Cisco provides 28 ports
where HP and NETGEAR provide 24 has a
beneficial impact on Cisco’s price/
performance results.
Page 8 of 12
Cisco Systems 300 Series Managed Switches
#211103
LAN Switch: Power over Ethernet (PoE) Support
as reported by Sifos PSA-3000 PowerSync Analyzer
Vendor
Product
Cisco Systems
D-Link Systems
SF 300-24P
DES-3528P xStack
HP Networking
E2610-24/12PWR PoE Switch
HP Networking
Power Budget
(W)
173.6
363.0
Power over Ethernet
IEEE 802.3af2003 (up to 15.4W)
Yes. Available on all ports. 11 ports can be powered at a full 15.4W
Yes. Available on 24 ports. 23 ports powered simultaneously at full power
126.0
Yes. Available on 12 ports. 8 ports powered simultaneously at full power
Note: The Power Budget is the total amount of power that the switch can make available to PoE ports, thus, a higher number is better. Power Budget is
typically shared among all PoE ports. For reference, Tolly notes that virtually every VoIP phone tested by Tolly has required less than 7.6W. Devices tested
without any additional external PoE power source. HP Networking also offers the E2610-24/24PWR switch with PoE on 24 ports and an advertised power
budget of 406W.
Table 3
Source: Tolly, December 2010
Systems Under Test: 10/100 & Gigabit Ethernet
Non-Stackable, Fully Managed, Non-PoE and PoE LAN Switches
Vendor
Cisco
Systems
D-Link
Systems
HP
Networking
NETGEAR
Product
Cisco SF300-24
(SRW224-G4-K9-NA)
Cisco SF300-24P
(SRW224-G4P-K9-NA)
Cisco SG300-28
(SRW2024-K9-NA)
D-Link DES-3528 xStack
(P1UQ3A8003662)
D-Link DES-3528P xStack
(P4LX199000012)
HP Networking E2610-24PT
(J9085A)
HP Networking E2610-24/12PWR
(J9086A)
HP Networking E2510-24
(J9019B)
HP Networking E2810-24G
(J9021A)
NETGEAR ProSafe FSM726
NETGEAR ProSafe GSM7224
Product Class
Software/Hardware
Version
10/100 Non-PoE
10/100 PoE
$215.00
11.0.0.27
0 0 27 21SEP2010
(Same software for
all Cisco SUTs)
GbE Non-PoE
10/100 Non-PoE
10/100 PoE
CDW Price (USD)
$445.00
$499.99
2.60.017
(Same software for
both D-Link SUTs)
$508.43
$1,174.99
(CompSource)
10/100 PoE
11.54
(Same software for
both HP 2610 SUTs)
$794.42
10/100 Non-PoE
Q11.26
$342.57
GbE Non-PoE
N11.25
$1,432.92
10/100 Non-PoE
8.0.1.9
$309.00
GbE Non-PoE
8.0.1.4
$716.99
10/100 Non-PoE
$528.99
Note:. Systems have at least 24 copper ports with some systems having 2+ uplink/stacking ports. Prices as listed as selling price on CDW website on Jan.
17, 2011 except for the the D-Link DES-3528P which was not listed on CDW, so the price on the CompSource website was used. HP recommended using
the E2610-24/24PWR (J9087A), with PoE on 24 ports, but it was not available in the test window.
Table 4
Source: Tolly, January 2011
© 2011 Tolly Enterprises, LLC
Tolly.com
Page 9 of 12
Cisco Systems 300 Series Managed Switches
Test Methodology
Whenever possible, tests were run using
default SUT configurations and using
generally accepted test methodologies
defined in relevant RFCs, Ixia test
configurations and/or Tolly Common Test
Plan methodologies. Thus, this section will
present notes of interest and/or variations
from standard test procedures.
Devices Under Test
The test compared LAN switches that were
fully managed and without proprietary
stacking ports. While actual port count
varied slightly among switches, all had at
least 24 ports of either Fast Ethernet
(10/100) or Gigabit Ethernet copper ports.
Most switches had 2 or more additional
Gigabit Ethernet links and/or dualpersonality links. See Table 4 for system
details.
Test Tools
The test traffic was generated using an Ixia
Optixia XM2 chassis outfitted with 2, 16-port
Gigabit Ethernet line cards. The Power over
Ethernet tests were run using the Sifos
PSA-3000 PowerSync analyzer.
Energy consumption was measured using
the Watts up? PRO ES power meter. This
device recorded power consumption in onesecond intervals to provide granular power
consumption data.
Throughput and Latency
For this test, engineers connected all ports
on the DUT to Ixia XM2 test ports, keeping
the switch in its default configuration.
Engineers then ran the RFC 2889
throughput test included as a standard test
in the Ixia platform, varying the frame size
and using a binary search to determine the
maximum throughput. Latency metrics
were taken using the RFC 2889 latency test,
built in to IxAutomate, using the Cut-
© 2011 Tolly Enterprises, LLC
#211103
through calibrated metric at 100% line rate
on all ports. In cases where the standard
deviation across the port pairs was greater
than 5%, additional tests were run and
averaged.
Power Consumption
To measure the power consumption of the
switches under load, engineers modified the
RFC 2544 template in IxAutomate to yield a
test which varied the load at 50% line rate
increments over both 64 and 1518 byte
frames, which was run with 24 links active
on the DUT. Measurements were recorded
using a Watts Up? PRO ES power meter. Each
metric was recorded, with additional data
collected including the Gigabit uplink ports.
In no scenario was the percent error greater
than 0.5% of the mean.
Industry Solutions website at http://
www.atis.org.
Power over Ethernet
To measure PoE capabilities, engineers
enabled PoE on each of the supported
switches. Then, they connected all 24 ports
to the Sifos PSA-3000 PowerSync Analyzer.
Engineers used the integrated
mp_power_cap test, which requested a
minimal amount of power draw from the
switch on all ports, then gradually increased
the request on each port, terminating when
the switch drops power from one port to
keep a higher priority port active. The
maximum supplied power was recorded,
and tests were run a total of three times to
accurately measure the power budget of the
DUT.
D e v i c e s we re t e s t e d i n d e f a u l t
configurations. Testers noted that HP
Networking offers an option that turns off
the LEDs on the switch. Informal tests show
that enabling this feature reduces switch
power consumption by 1W.
ATIS Power Rating
The ATIS power rating refers to the
recommended methodology promulgated
by the Alliance for Telecommunications
Industry Solutions (ATIS).
The ATIS method is based on measuring the
switch in three different states: idle, 10%
load and 100% load. The ATIS calculation
consists largely of the power consumption
at 10% load. That is 80% of the ATIS value.
The remaining part of the ATIS value
consists of 10% of the idle power
consumption and 10% of the power
consumption at 100% load.
The test methodology used for this
report relies upon test procedures,
metrics and documentation
practices as defined in various Tolly
Common Test Plan documents.
To learn more about Tolly Common
Test Plans, go to:
http://www.CommonTestPlan.org
The formula for the ATIS power rating is 0.1*
(at 0)+.8*(at 10) +.1*(at 100) where (at n) are
utilization rates. For more information, see
the Alliance for Telecommunications
Tolly.com
Page 10 of 12
Cisco Systems 300 Series Managed Switches
#211103
Gigabit Ethernet LAN Switch: Layer 2 System Throughput & Latency Under 100% Load
24 Ports in Full Mesh Configuration
(as reported by Ixia IxAutomate 6.90)
10/100 Non-PoE
System Under
Test
Cisco
SF300024
DLink
DES33528
Frame size Throughput Latency Throughput Latency
(bytes)
(%)
(µs)
(%)
(µs)
HP Netwoorking
E251024
Throughput
(%)
HP Netw
working
E261024PT
Latency Throughput Latency
(µs)
(%)
(µs)
NETGEAR
FSM7726
Throughput
(%)
Latency
(µs)
64
100
10.10
100
208.3
99.481
10.1 1
100
191.1
100
15.5
128
100
15.2
100
216.7
99.701
15.2 1
100
190.9
100
20.9
256
100
25.7
100
228.3
99.841
25.5 1
100
212.1
100
31.3
512
100
46.5
100
249.5
99.915
46 1
100
213.9
100
52.1
1024
100
87.5
100
290.8
99.957
87 1
100
265.7
100
92.5
1280
100
107.9
100
311.5
99.963
107.4 1
100
278.4
100
113
1518
100
127
100
330.7
99.979
225.7 1
100
307.7
100
132.5
10/100 PoE
System
Under Test
Cissco
SF300024P
Gigabit Ethernet
DLink
DES3528P
HP Nettworking
System
E2610224/12PWR Under Test
Cissco
SG300028
HP Nettworking
E281024G
NETG
GEAR
GSM
M7224
Frame size Through Latency Through Latency Through Latency Frame size Through Latency Through Latency Through Latency
(bytes) put (%) (µs) put (%) (µs)
put (%) (µs)
(bytes) put (%) (µs) put (%) (µs)
put (%) (µs)
64
100
10.10
99.481
14 1
100
224.4
64
100
3.20
100
4.1
100
4.1
128
100
15.2
100
703.4
100
223
128
100
3.7
100
4.6
100
4.6
256
100
25.7
100
713.6
100
244.2
256
100
5
100
5.6
100
5.6
512
100
46.5
100
733.9
100
250.1
512
100
7.4
100
7.6
100
7.7
1024
100
87.4
100
774.6
100
300.4
1024
100
11.5
100
11.7
100
11.7
1280
100
107.9
100
795
100
314.6
1280
100
13.5
100
13.7
100
13.8
1518
100
127
100
814.1
100
345.2
1518
100
15.4
100
15.6
100
15.7
Note: Throughput results are listed as the percentage of maximum theoretical throughput of 24 10/100 or Gigabit Ethernet ports as appropriate. Uplinks
not used for the test. Note: Cisco’s GbE switch has 28 ports where HP and NETGEAR’s have 24 ports. For purposes of comparison, 24 ports were used on the
Cisco switch.
1. Frames were dropped during the 100% load, latency measurement test.
Source: Tolly, December 2010
© 2011 Tolly Enterprises, LLC
Table 5
Tolly.com
Page 11 of 12
Cisco Systems 300 Series Managed Switches
About Tolly
The Tolly Group companies have been
delivering world-class IT services for
more than 20 years. Tolly is a leading
global provider of third-party
validation services for vendors of IT
products, components and services.
You can reach the company by e-mail
at sales@tolly.com, or by telephone at
+1 561.391.5610.
Visit Tolly on the Internet at:
http://www.tolly.com
#211103
Test Equipment Summary
The Tolly Group gratefully acknowledges the providers
of test equipment/software used in this project.
Vendor
Product
Web
Chassis Type: XM2
Interfaces: 32x 1Gbps
Card Type: 2 x LSM1000 XMV16-01
Software: IxAutomate
6.90.98.5GAPatch4, IxExplorer IxOS
5.70.600.9 EA-SP1
http://www.ixiacom.com
PSA-3000 PowerSync Analyzer
http://www.sifos.com
Terms of Usage
This document is provided, free-of-charge, to help you understand whether a given product, technology or service merits additional
investigation for your particular needs. Any decision to purchase a product must be based on your own assessment of suitability
based on your needs. The document should never be used as a substitute for advice from a qualified IT or business professional. This
evaluation was focused on illustrating specific features and/or performance of the product(s) and was conducted under controlled,
laboratory conditions. Certain tests may have been tailored to reflect performance under ideal conditions; performance may vary
under real-world conditions. Users should run tests based on their own real-world scenarios to validate performance for their own
networks.
Reasonable efforts were made to ensure the accuracy of the data contained herein but errors and/or oversights can occur. The test/
audit documented herein may also rely on various test tools the accuracy of which is beyond our control. Furthermore, the
document relies on certain representations by the sponsor that are beyond our control to verify. Among these is that the software/
hardware tested is production or production track and is, or will be, available in equivalent or better form to commercial customers.
Accordingly, this document is provided "as is", and Tolly Enterprises, LLC (Tolly) gives no warranty, representation or undertaking,
whether express or implied, and accepts no legal responsibility, whether direct or indirect, for the accuracy, completeness, usefulness
or suitability of any information contained herein. By reviewing this document, you agree that your use of any information
contained herein is at your own risk, and you accept all risks and responsibility for losses, damages, costs and other consequences
resulting directly or indirectly from any information or material available on it. Tolly is not responsible for, and you agree to hold Tolly
and its related affiliates harmless from any loss, harm, injury or damage resulting from or arising out of your use of or reliance on any
of the information provided herein.
Tolly makes no claim as to whether any product or company described herein is suitable for investment. You should obtain your
own independent professional advice, whether legal, accounting or otherwise, before proceeding with any investment or project
related to any information, products or companies described herein. When foreign translations exist, the English document is
considered authoritative. To assure accuracy, only use documents downloaded directly from Tolly.com. No part of any document
may be reproduced, in whole or in part, without the specific written permission of Tolly. All trademarks used in the document are
owned by their respective owners. You agree not to use any trademark in or as the whole or part of your own trademarks in
connection with any activities, products or services which are not ours, or in a manner which may be confusing, misleading or
deceptive or in a manner that disparages us or our information, projects or developments.
211103-nbs-1-kt-22Feb11VerK
© 2011 Tolly Enterprises, LLC
Tolly.com
Page 12 of 12
Download PDF
Similar pages