Comments on: Blades and Virtualization Aren’t Mutually Exclusive: Part Two, IBM Power Sizing http://blog.scottlowe.org/2009/02/07/blades-and-virtualization-arent-mutually-exclusive-part-two-ibm-power-sizing/ The weblog of an IT pro specializing in virtualization, storage, and servers Thu, 18 Mar 2010 06:21:43 +0000 http://wordpress.org/?v=abc hourly 1 By: adelp http://blog.scottlowe.org/2009/02/07/blades-and-virtualization-arent-mutually-exclusive-part-two-ibm-power-sizing/comment-page-1/#comment-43587 adelp Tue, 10 Feb 2009 00:51:45 +0000 http://blog.scottlowe.org/?p=1162#comment-43587 @Duncan and Dustin - Thank you!! @ Tony - I couldn't agree with you more. Plus, the calculators assume 100% utilization by default. This isn't real word. What about an HP Chassis and an IBM Chassis running at 50%-75%. One or the other may have power designs to run better at standard workloads than the other. Real world is the only way to compare. @DanC - Hehe - Yeah, I've had that. The HP has its share of bugs as well. For instance the results aren't very reliable if you change the utilization below 100%. @Duncan and Dustin - Thank you!!

@ Tony - I couldn’t agree with you more. Plus, the calculators assume 100% utilization by default. This isn’t real word. What about an HP Chassis and an IBM Chassis running at 50%-75%. One or the other may have power designs to run better at standard workloads than the other. Real world is the only way to compare.

@DanC - Hehe - Yeah, I’ve had that. The HP has its share of bugs as well. For instance the results aren’t very reliable if you change the utilization below 100%.

]]> By: Justin http://blog.scottlowe.org/2009/02/07/blades-and-virtualization-arent-mutually-exclusive-part-two-ibm-power-sizing/comment-page-1/#comment-43585 Justin Tue, 10 Feb 2009 00:08:15 +0000 http://blog.scottlowe.org/?p=1162#comment-43585 We are actually in the process of moving from HS21XM to IBM 3950 due to the higher densities, but before that we were VERY happy with blades. We are actually in the process of moving from HS21XM to IBM 3950 due to the higher densities, but before that we were VERY happy with blades.

]]> By: Dan C http://blog.scottlowe.org/2009/02/07/blades-and-virtualization-arent-mutually-exclusive-part-two-ibm-power-sizing/comment-page-1/#comment-43582 Dan C Mon, 09 Feb 2009 22:54:53 +0000 http://blog.scottlowe.org/?p=1162#comment-43582 You forgot to mention the most important point about the IBM Power Tool. Don't attempt to change your configuration mid-swing because it'll crash and forget all of your hard work :D You forgot to mention the most important point about the IBM Power Tool.

Don’t attempt to change your configuration mid-swing because it’ll crash and forget all of your hard work :D

]]> By: Tony http://blog.scottlowe.org/2009/02/07/blades-and-virtualization-arent-mutually-exclusive-part-two-ibm-power-sizing/comment-page-1/#comment-43577 Tony Mon, 09 Feb 2009 16:48:03 +0000 http://blog.scottlowe.org/?p=1162#comment-43577 Great blog posts, would really like to re-iterate the point you make that you can't compare power calculator numbers. IBM, HP, Dell, SUN etc all use different applications to generate the load for their measurements and may also have different power measurement methodologies. An obvious example of this is that each vendor has a report that says that they are more power efficient that the other vendors. The only way to compare power reliably at this time is to run your application in your environment and measure the input power to a full enclosure. One day maybe SPECpower_SSJ2008 will work on blades and we'll have a standard power measurement benchmark like we have for rack mounts. Great blog posts, would really like to re-iterate the point you make that you can’t compare power calculator numbers. IBM, HP, Dell, SUN etc all use different applications to generate the load for their measurements and may also have different power measurement methodologies.

An obvious example of this is that each vendor has a report that says that they are more power efficient that the other vendors.

The only way to compare power reliably at this time is to run your application in your environment and measure the input power to a full enclosure.

One day maybe SPECpower_SSJ2008 will work on blades and we’ll have a standard power measurement benchmark like we have for rack mounts.

]]> By: Dustin http://blog.scottlowe.org/2009/02/07/blades-and-virtualization-arent-mutually-exclusive-part-two-ibm-power-sizing/comment-page-1/#comment-43576 Dustin Mon, 09 Feb 2009 14:33:58 +0000 http://blog.scottlowe.org/?p=1162#comment-43576 We arrived at a very similar conclusion about a year ago and opted for the Blade Center H & HS21XM combo in numerous offices. No regrets. Keep up the great posts. We arrived at a very similar conclusion about a year ago and opted for the Blade Center H & HS21XM combo in numerous offices. No regrets. Keep up the great posts.

]]> By: Duncan http://blog.scottlowe.org/2009/02/07/blades-and-virtualization-arent-mutually-exclusive-part-two-ibm-power-sizing/comment-page-1/#comment-43568 Duncan Sun, 08 Feb 2009 09:07:19 +0000 http://blog.scottlowe.org/?p=1162#comment-43568 Cool blog posts so far, keep it up! Cool blog posts so far, keep it up!

]]>