The rise of cloud computing is going to stoke demand for servers, according to a new forecast from IDC. For me, the critical point is that public cloud providers like Google don't buy servers, they build them. And the design decisions they make -- constructing sparsely configured but powerful scale-out servers -- will feed back into the enterprise market.
The rise of cloud computing is going to stoke demand for servers, according to a new forecast from IDC. For me, the critical point is that public cloud providers like Google don't buy servers, they build them. And the design decisions they make -- constructing sparsely configured but powerful scale-out servers -- will feed back into the enterprise market.If you want to check out the IDC forecast, go here. The raw numbers of which they speak are as follows: "Server revenue for public cloud computing will grow from $582 million in 2009 to $718 million in 2014. Server revenue for the much larger private cloud market will grow from $7.3 billion to $11.8 billion in the same time period."
When you consider that the overall server market was around $44 billion in 2009, the qualitative takeaway is that the public-cloud portion amounts to a spring rain shower. However, I believe the influence of this sector will be disproportionately large. Here's why:
The server market for public cloud vendors such as Google and Amazon is a hidden market. These guys don't go out and buy in bulk from Dell or HP. They roll their own, and the own that they roll are highly, highly, highly efficient designs. The strip their blades of all extraneous components and outfit them with nothing but the bare necessities. So if taking out a USB port saves $5 and a couple of watts of power, that's how the system is configured.
What's also not obvious to the public is that the Googles of the world exert influence on the server design arena right at the source. Namely, they go directly to Intel and AMD and tell them stuff they'd like to see in terms of TDP (processor power dissipation), power management, price/performance, etc. Since a Google can buy upwards of 10,000 of a processor SKU, Intel and AMD listen.
So who cares if Google builds cheap, stripped down servers and attaches them in their data centers with Velcro strips? Well, I admit that right now there isn't a clear connection between Google's servers and what a Tier 1 server maker sells to its enterprise customers. (It should be obvious, but in case it's not, I'm talking about the x86 server space, not RISC/Itanium.)
However, as the server technologies applied by the public cloud vendors becomes more visible, it will influence buyers. For one thing, they'll say, 'Hey, we want these things.' Thus it will become harder, over time, for the Tier 1's to sell quite as many of their heavily configured conventional servers.
More correctly, I should say, it will force them to continue to push the value proposition up from the server level, per se, into the higher rung of the data-center play. As in, how can you offer an enterprise something configured to meet its total compute, storage, and networking needs. And also connect it together, manage it seemlessly, support heavy use of virtualization, and yada, yada, yada.
Fortunately for the likes of IBM, HP, Dell, Oracle/Sun and Cisco, that's where they excel anyway. Which is good, because that's where the competitive center of gravity in the server space will be moving, if it hasn't already.
Google in the Enterprise SurveyThere's no doubt Google has made headway into businesses: Just 28 percent discourage or ban use of its productivity products, and 69 percent cite Google Apps' good or excellent mobility. But progress could still stall: 59 percent of nonusers distrust the security of Google's cloud. Its data privacy is an open question, and 37 percent worry about integration.
InformationWeek Must Reads Oct. 21, 2014InformationWeek's new Must Reads is a compendium of our best recent coverage of digital strategy. Learn why you should learn to embrace DevOps, how to avoid roadblocks for digital projects, what the five steps to API management are, and more.