Burning question: Blades or rack?

* When should you buy blades and when should you stick with rackmount systems?

Blade servers are no longer on the bleeding edge, but there continues to be a lot of debate about when and where the slimmed-down servers make sense. It seems they’re starting to make sense in a lot more situations. Last week, I wrote about Sun rolling out a bunch of new blades, which it hopes will help it gain a footing in the growing market. Today, HP and IBM lead in blade server sales.

But when should you buy blades and when should you stick with rackmount systems? Tom Henderson and Rand Dvorak of Network World’s Lab Alliance tackled that question head-on recently. They conducted a real-world test that pitted IBM’s blade servers – the HS21 and HS21 XM, to be exact – against IBM’s rackmount servers – the x3550 and the x3650.

All the servers used quad-core Xeons, but you can get more specifics about the configurations by reading the full report here (free registration required).

Henderson and Dvorak came away with some interesting findings, one being that blade servers are more economical when it comes to power and cooling. Overall, blades and rackmount servers performed equally, Henderson and Dvorak said, but noted that newer, faster CPUs likely will show up first in typical rack servers vs. blades, where engineering time creates a technology lag.

Another good point in the article is that while blades offer a number of benefits, they also result in vendor lock-in since blades must fit in a specific chassis and be fitted with specific components. Also, while blades may be easier to service, some applications, such as those that require more onboard storage, run better on rack servers.

What are your thoughts when it comes to blades vs. rack servers? Read the review and let me know.

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.

Copyright © 2007 IDG Communications, Inc.

IT Salary Survey: The results are in