Power and cooling: Cisco vs. Enterasys

Brad

I noticed all your articles have turned into Cisco bashing articles. Has Cisco cracked down on your grey market gear or something?

Nice to see these comparisons. However I think that you have to include some performance comparisons as well (the first thing that people look for when buying network gear is still the specs/performance).

And I don't when the Enterasys switches where introduced, but I know that we didn't care about power and cooling in the time that the 6500 was introduced.

...there's nothing of value at all in the above data. You didn't give us the specific hardware configs in use (not even specific models) and it's hard to tell if this is an apples-to-apples comparison. I can't even tell if we are comparing two fruits. Was this a spec sheet comparison? Does it show us true power consumption under load? We've noticed that actual power draw is typically way less than maximum power draw. It's important to understand that the power supply specs don't equate to how much power is utilized. I can slap a 1200W power supply to replace the 600W in my homebuilt gaming PC, but it won't magically start drawing 1200W. It's still drawing less than the 600W it always did. But, I could go with that ultra-cool SLI config and shoot up to well over 700W and I'll be glad I upgraded the power supply. There's no appreciable difference as to how this applies to networking equipment, servers, etc. I appreciate the notion of such a comparison, but this is completely useless and pretty much wrong on all accounts. Everyone is starting to think about these aspects of their IT infrastructure, including networking equipment, and there's a lot of misinformation out there (like this) that isn't helping folks.

Brad, with titles like these, followed by a complete lack of any useful content seems to indicate you are hard up for hits. If you are getting paid to just post up an Enterasys document and then conclude facts from it, you are basically producing nothing useful for Network World readers, and frankly will be ignored. You must think NW readers are ignorant by claiming maximum capacity of a power supply is a rating of power efficiency.

Brad, Good Network engineers are probably some the most detail oriented, and analytical IT workers. We want and live by the facts, not baseless claims. This does not help us in our professions.
You have shown a lack of regard for facts, repeatedly claimed misleading and baseless accusations in which even you have to recently correct. I think you have done immense amount of damage to your own credibility in this business by letting personal biases get in the way of creating thoughtful editorials that truely are valuable. You are partially responsible for turning Network World from a valuable resource and voice for the network technology industry to a IT tabloid.

Everyone else claims to have some magic technology that can do everything at 75% of the power that a Cisco device consumes. Does anybody truely believe that every other network company has some technology that Cisco itself could not create and use for some reason.? If so, show me the patents? Because if they could, it would be darn impressive.

If the marketing is claiming something like this from ANY company, don't blindly trust it. Look at the datasheets yourself and make a real apples-to-apples comparison of REAL traffic performance, scalability, congestion handling, density, and availability.

Only by reducing ports, virtualizing, and consolidating will real measurable power gains be found. Compare an 4000W draw on a Cisco 6509 chassis to the power draw of its 384 to servers connected to it, you realize that a network consumes less than 2% of power compared to a servers.

Now, lets get some clarity on the claims that Brad had above.

N Series:
Enterasys does not list Watts or Current on their data sheets for their Diamond, Platinum, or Gold cards from their N-Series. What mix of ports? What is the performance of the system? How redundant is a configuration that draws up to 1500W? .
However, to take a page from Brad's math book: The WS-CAC-3000W from Cisco can run on a 110V-16A feed and max draw on it is 1450W. CISCO CONSUMES 50W LESS POWER THAN ENTERASYS by Brad's math!

N-series with poe:
Enterasys states that the 5 slot chassis is optimized for PoE. They claim approx 350 ports of PoE in this chassis. The 6509 can handle 384 PoE ports. Where is the information about the type of PoE devices, how they negotiated, etc. Assuming these devices are ~10W a piece, that is an additional 340W of available PoE for Cisco in a single chassis. More than likely, these would work fine with a 6000W power supply. We are not even considering the 575 ports of PoE in the 6513 which the 8700W power supplies where designed for. Enterasys would need two complete chassis to match this, which would mean 10,200W of Enterasys max draw based on power supply math. Cisco WINS BY 1500 Watts !

Nseries vs. 4506e:
Using Cisco's own power calculator and loading them up with a 4548GE Modules and a supervisor, the MAX power draw is 529.92 Watts. Guess what?? CISCO WINS BY 970 Watts!!!. Using your math, you can use a 1000W power supply from Cisco and .. CISCO STILL WINS BY 500 WATTS!

G-series vs. 3560e:
PoE or Non PoE ?? --
PoE = How does a Enterasys 1050W power supply provide 96 ports of Class 3 (15.4W) of of power?? I guess some ports cannot be used.. How about 72 ports of Class 3 power? That is still 1108W of PoE, which is 58W more than the included power supply can handle. (not including its processor) So, I guess this is now just a 48port PoE switch. Do the math.
Cisco can do 15.4 on each of the 48 ports in the 3560e with the 1150W with PLENTY of power to spare for future increased draws by devices. .. In fact, Cisco's own calculator shows that the MAX draw in the switch is only 820 Watts with the CPU with class 3 devices. Cisco WINS by 230 Watts!

Now, doing a Brad Reese comparison of non-poe versions : The 3560e 48 port switch has a 265W power supply. I guess Cisco WINS by 785 Watts again!