|
Tuesday February 28, 2006
The server industry is not doomed Nicolas Carr has a blog entry asking if the server industry is doomed. Here is his first point:
"The most immediate threat is the twin trends of consolidation and virtualization. To save money, companies are merging their data centers and standardizing the applications they run throughout their businesses."
I'm going to take a whack at a counter-point. I have heard similar logic applied to CPUs. The anology may be a stretch, but I think it applies. It goes like this. AMD and Intel should not build dual core servers. If a server is twice as "fast", people will buy 1/2 has many, right? Of course not. As CPUs get faster, companies (in the end) buy more using logic simiar to the following. Sun creates virtualization technologies, such as Zones and Domains, to facilitate the very consolidation Nicolas warns of. As efficiency goes up, so does value. Value drives adoption. Adoption drives volume. Volume drives down price. Lower price results in broader applicability. Broader applicability results in more servers. Sounds a bit like the Lion King's "circle of life", doesn't it?
In the Bayer example he gives, Nicolas mentions they cut the server count in half. I have a gut feel there is more to the story. I have seen server consolidation many times in my customer base. Let's say, for the sake of argument, a generic customer has 1,335 SPARC servers. A typical customers has a few SPARCcenter 1000's., some Sparc 20's, Ultra 2's, Ultra 10s, v100's, v220's, E250's, E450's, v210's, E4500's, E3800's, 6800's, and who knows, a couple of E10K's and E25Ks. This environment was built over a decade. What companies typically do is go through a cost analysis. Simply due to Moore's law, they can collapse 32 of those old v100's on to one T2000 (work with me here, I'm pulling numbers out of the air). They could replace a dozen of the Ultra 2's and ultra 10's with a v490. If you work your way up the compute ladder, they can replace the E10K with a v890 (this is a stretch as the E10K was much more than raw compute power). Guess what. The consolidation effort results in a lot of new hardware. The customer saves money. The vendor (Sun in this case) makes money. Not all of the existing hardware goes away. It's simply doing the math with regards to the cost of support contracts of older/slower hardware versus newer/faster hardware. Some older hardware stays. Some goes. I'll let you do the math :)
The end of the consolidation effort is not the end of of IT's role. Companies want to grow ... a lot. Growth may be by acquisitions. See last paragraph on consolidation. Growth may result in new business initiatives. That means demand for more servers. This time around, though, virtualization applies. I can't tell you the number of times new initiatives are taken off the plate because of the up-front investment they require or the greater-than-18-month ROI. Virtualization enables some of those new initiatives to remain on the "go forward" list. To put it more simply, collapsing the server market from 20-25 million servers down to 10-15 million servers assumes a static marketplace. Even if it is static (which it is not), the consolidation is a business opportunity in and of itself.
To sum all of this up, efficiency drives more opportunity, not less. There is much more to Nicolas Carr's blog entry. It's a good read. In fact, the point I counter is only a small point of his entry. His greater point is about grids. However, this entry is already past being long in the tooth and I'm gettin' tired :) I am interested in hearing your thoughts on Nicolas's entry, and the logic (or lack thereof) of my response.
(2006-02-28 22:55:27.0) Permalink Comments [2]Comments are closed for this entry.
In the context of Carr's comments, what I'd imagine is that very a reliable high-bandwidth network connection is available anywhere for very cheap. That condition is only partially true today, but suppose that one day that will generally become true.
If that were the case, why would any company not in the IT services business want to run its own data center? I mean, few companies run their own power generators (some do, I admit), mainly because access to the power grid is readily available practically anywhere for a reasonable cost. Wouldn't it be so much more convenient to just be able to use a thin client to log into a remotely hosted desktop? And with standard application interfaces, such as J2EE, shouldn't a company's IT department be able to deploy an enterprise app into a remote data center's hosting environment? Many companies already do this, but I'm curious why wouldn't most follow that path (again, imagining that the above condition is already true).
If that were the case, these remote data center services could operate servers with much higher utilization, hence leading to even further consolidation of servers, and hence lower costs not only for servers, but especially for such hosting services. If hosting service costs fall faster than server costs do, then there might be an even bigger economic incentive for enterprises to outsource their data center operations, hence leading to even faster consolidation of applications to servers.
Interestingly, that's why, I think, Sun's strategy of throughput-oriented computing is going to pay off. Because data centers will want systems that they can consolidate their customers' applications into. So why it's true that there will be fewer servers sold, it will be interesting what vendors will gain market share and remain standing.
Posted by Frank Sommers on March 01, 2006 at 05:34 PM PST #
Frank, your ahead of me. That's the direction I was heading with the 2nd half of my blog in addressing his blog before I got tired and ended it short (short is relative :) ). My thoughts were also on throughput computing and the SunGrid. You just beat me to it :) Good comment.
Posted by John Clingan on March 01, 2006 at 06:11 PM PST #