Cisco's UCS Next Year's Servers This Fall

When the rest of the blogosphere went crazy over Cisco's UCS announcement pronouncing it the biggest change in the data center since the microprocessor I was skeptical to say the least. While others touted they had management software that did the admin's job for him and revolutionary networking that made IBM and HP blade servers look like Model T Fords. I failed to understand the excitement.

Howard Marks

July 2, 2009

4 Min Read
Network Computing logo

When the rest of the blogosphere went crazy over Cisco's UCSannouncement pronouncing it the biggest change in the data center since themicroprocessor I was skeptical to say the least. While others touted they hadmanagement software that did the admin's job for him and revolutionarynetworking that made IBM and HP blade servers look like Model T Fords. I failedto understand the excitement.

To me UCS was another player in the blade server market.After all I had just, ok a month before, been reading reports from some of thesame analysts that were now touting UCS as a breakthrough saying that Dell wasno threat to IBM, HP and Sun (OK they missed that one) because their biggestserver was a 4 socket x86. If Dell wasn't and enterprise player because theydidn't have big Unix systems or mainframes how was Cisco with just 3 models of2 socket blades that much better?

Wondering if I might be wrong I went back to the Cisco siteand found almost no details about the hardware just platitudes about how theintegrated servers, management software, networking and storage would makeeveryone's life easier. This confused me even more as UCS has no storage. Noteven the storage blades HP has running Lefthand that let you run a whole branchoffice from a single blade chassis.  EMCblogger Chuck Hollis was even saying UCS was better for not having storage asthe HP storage was something you couldn't tell from Shineola.

I even went to LULU.COM, bought, and read, Silvano Gai'sbook looking for what made UCS different.In terms of management they are shifting the paradigm fromthe blade chassis to a larger cluster and are using what looks like goodmanagement software from BMC. But HP's Insight Orchestration is pretty good asis Scalent and I could buy BMC's software to run my SuperMicro or Fujitsublades.

They didn't even do anything terribly innovativemechanically when compared to say Verari's whole rack chassis that breathesbottom to top. So what, other than the Cisco name, made UCS special?

After much reflection I finally get it.  Cisco is shipping next year's servers, nowrack mount as well as blades this year. 

UCS servers use use 10Gig CEE (OK Cisco calls it DCE) with FCoE as theirstandard networking where everyone else is 1gig standard and 10gig with orwithout FCoE as an option. They only use the latest and greatest Xeons.  Their high end servers can hold 3-4 times asmuch memory as everyone else's, a real differentiator if you want 2 socketswith over 128GB of memory. I expect most users would rather have more VMwarehosts in their clusters than 384GB of memory in one but it has a place.

In designing and pitching UCS Cisco's been very clever aboutboth positioning and partnering.  Knowingthat displacing HP, IBM or Dell as a big user's server vendor is going toinvolve months of in house testing, and politics, they didn't worry aboutbuilding servers they could sell at a competitive price in volume today.Instead they're selling the product that will be the mass market product at theend of the 6-12 month sample, test, pilot process any new server vendor isgoing to have to go through.

Other than the big memory trick, which Cisco has a patentapplication in on, HP and the rest could have made servers like UCS with 10GigELOM but they would be expensive and sell in small numbers which isn't the Delland HP business model.  They'll have themnext year when the costs come down and customer demand goes up.Cisco also knew they're champions in any enterprise were thenetwork guys so they made UCS all about fancy new network toys like 10Gig andFCoE. Notice there aren't FC options UCS is an in for a penny in for a poundplay on the network side because those are Cisco's most likely customersanyway.

Finally they cozied up to EMC, and to a lesser extent NetAppsince NetApp is the only vendor with native FCoE storage. This combines the newplayer in the server business with the only major players in the storagebusiness that aren't server makers themselves.

Clearly Cisco needed to expand into a new market to maketheir growth targets.  Servers make senseand the servers they announced make perfect sense for their position in themarket as a whole. Of course they're still not as wonderful as the first roundof hype claimed but that's how the hype cycle works.

About the Author(s)

Howard Marks

Network Computing Blogger

Howard Marks</strong>&nbsp;is founder and chief scientist at Deepstorage LLC, a storage consultancy and independent test lab based in Santa Fe, N.M. and concentrating on storage and data center networking. In more than 25 years of consulting, Marks has designed and implemented storage systems, networks, management systems and Internet strategies at organizations including American Express, J.P. Morgan, Borden Foods, U.S. Tobacco, BBDO Worldwide, Foxwoods Resort Casino and the State University of New York at Purchase. The testing at DeepStorage Labs is informed by that real world experience.</p><p>He has been a frequent contributor to <em>Network Computing</em>&nbsp;and&nbsp;<em>InformationWeek</em>&nbsp;since 1999 and a speaker at industry conferences including Comnet, PC Expo, Interop and Microsoft's TechEd since 1990. He is the author of&nbsp;<em>Networking Windows</em>&nbsp;and co-author of&nbsp;<em>Windows NT Unleashed</em>&nbsp;(Sams).</p><p>He is co-host, with Ray Lucchesi of the monthly Greybeards on Storage podcast where the voices of experience discuss the latest issues in the storage world with industry leaders.&nbsp; You can find the podcast at:

Stay informed! Sign up to get expert advice and insight delivered direct to your inbox

You May Also Like

More Insights