Study: Power drawn by Ethernet switches varies widely
In-Stat sees wide variation in energy efficiency
By
Tim Greene
,
Network World
, 05/15/2008
- Share/Email
- Tweet This
- Print
A recent study by In-Stat says that fixed-configuration 24- and 48-port switches can vary by more than 600% in terms of how much switching
capacity they support in relation to how much electric power they draw.
The range measured in gigabits per second per Watt ranges from a low of less than .5Gbps/Watt to a high of more than 3Gbps/Watt,
the study says. (Compare Ethernet access switches.)
For 24-port switches, In-Stat found the best performing in order were made by 3Com, SMC, Netgear, H3C and D-Link. For 48-port switches, the top five in order were made by 3Com, Force10, Netgear, Extreme Networks and SMC.
Foundry Networks and Cisco placed as the bottom two among 24-port switches and Cisco and HP ProCurve placed as the bottom two among 48-port switches. The most efficient was the 24-port 3Com switch with more than 3Gbps
per Watt, and the least efficient was the 24-port Foundry switch at less than .5Gbps per Watt.
The 13 vendors rated were: 3Com, Allied Telesis, Cisco, D-Link, Enterasys, Extreme, Foundry, Force10, H3C, HP ProCurve, NetGear,
Nortel and SMC.
In-Stat finds that in general, energy efficiency drops as the number of switch ports increases. Energy efficiency does not
vary much between low- and high-end models of switches made by the same manufacturer, the study says.
More important is who makes the switch. "In-Stat determined that even among similarly equipped switches capable of performing
identical tasks, there are significant vendor-specific differences in energy efficiency," the study says.
Efforts to improve power efficiencies are underway in the IEEE, but may not bear fruit for some time yet, says David Law, the chairman of the IEEE's 802.3 committee
that oversees Ethernet issues.
A working group on energy-efficient Ethernet is considering standards for determining when a link is idle and automatically dropping it down to lower power — just enough
to keep the link alive, then throttling power back up when traffic demands it, says Law, who also works for 3Com.
The key is getting standards so devices made by different vendors can participate in varying power in response to demand,
he says. The power reductions would apply to end devices such as workstations and servers.
"This is an active idle, a low-power idle — in fact off, if possible — with occasional keep-alive bursts," Law says.
Partner Content
Simplify Your Branch Infrastructure
Learn how to simplify your branch infrastructure while dramatically increasing app performance with Citrix Branch Repeater.
Download the Free Info Kit
Next-Gen Load Balancing
Free Guide: "Next Gen Load Balancing: 8 Things You Need to Handle Today's Network Traffic" shows you the functionality needed in your next load balancer.
Download the Free Guide
Accelerate Your Web Apps by up to 5x
Free Guide: "The Secret to Getting Maximum Speed from your Web Applications."' Learn how you can deliver Web apps up to 5x faster.
Download the Free Guide
Comments (4)
In-Stat Study is Flawed!By xcountryflyer on May 15, 2008, 7:03 pmI have personally read the report and the author uses poor and inconsistent methodology. The author compares SoHo products with enterprise class products. Of course...
Reply | Read entire comment
3 major problems with studyBy Anonymous on May 20, 2008, 8:53 amLooking at the In-Stat report, there were several flaws or misleading assumptions made. Another thing to note is that the methodology is missing specific products...
Reply | Read entire comment
Ok, but...By Jon on May 30, 2008, 3:17 amAccording to a few other similar articles, 3Com and Force10 were in the top slots and Foundry was near the bottom. Are Force10 and 3Com not considered Foundry's...
Reply | Read entire comment
Flaw: you can't rely on published specificationsBy brycenesbitt on September 12, 2009, 3:23 amI tried making purchase decisions based on marketing materials, but found it impossible. The stated power consumption of the devices, including the ones branded...
Reply | Read entire comment
View all comments