Developer Platform Product Management - Development Unfiltered: From code, to cloud to comedy. : Performance

Our offer still stands and you know where to find us...

stevemar — Mon, 17 Aug 2009 19:35:00 GMT

I talked about the results of a benchmarking study performed by Greg Leake a couple months back on my blog. Greg took a hard look at Windows Server 2008 and the .NET Framework 3.5 in comparison to IBM’s Power6 and WebSphere 7 from both a cost and performance standpoint. The results show that customers can save up to 81% in total system costs by running applications on Windows Server 2008 rather than IBM WebSphere 7 on Power6/AIX. The study also shows that Power6 customers who switch to Windows Server 2008, but continue to use WebSphere, could save up to 66%. What’s more, these cost cutting measures do not come at the sacrifice of performance; customers could see an increase in performance up to 57% by employing .NET Framework and/or Windows Server 2008.

Why do I bring this up again? Well, we’re not done putting our platform to the test! Recently, we discovered that IBM quietly published their own benchmarking results to attempt to invalidate our original results. This provides us with yet another opportunity to make some comparisons, and bring them out to our collective customers in a public forum. The outcome - we stand by our previously published results and my earlier invitation stands to have IBM to meet us in an independent lab to perform additional testing of the .NET StockTrader and WSTest benchmark workloads and pricing analysis of the middle tier application servers tested in our benchmark report.

I guess the only thing I’m left wondering about is who will love Windows next?

...And we'll prove it!

stevemar — Fri, 01 May 2009 23:31:00 GMT

Yesterday I blogged about some recent findings regarding both system cost and performance when comparing Windows Server 2008 on an HP Blade Server against AIX on a POWER 570/POWER6 based server. As I stated, the tests showed that WebSphere loved running on Windows…to the tune of 66% cost savings and with better performance.

We encourage customers, third parties and IBM to check out the findings, run the sample apps and let us know what they find. IBM had a thought, but not about the results, just the fine print. Let me be clear: we stand behind our testing methodology and the results it generated. To prove it, we’ll pay for the bake-off. Here is our open letter to IBM:

Dear Ron,

I know you are just as interested in saving customers money and improving performance as we are which is why our recent tests caught your eye. To demonstrate my confidence in the numbers we have produced, I’d like to propose having a third party re-run the benchmark tests and publish the results. To make it easy, I’ll put my money where my mouth is and fund it. Are you in?

Hope to hear back soon!

Thanks,

steve

WebSphere Loves Windows…Who Knew?

stevemar — Thu, 30 Apr 2009 15:55:00 GMT

For years I’ve heard people say that Microsoft technology is great for mid range solutions. I can’t tell you how many CIOs have told me things like “we use WinTel for department level applications but the big iron apps run on Unix.” There’s a general belief in the industry that the further you get into the datacenter, the less Microsoft technology you tend to find. While license numbers, data from IDC and other major third parties paint a different story, we decided it was time to put ourselves to the test against one of the biggest players in the space – AIX running on optimized IBM Power6 with WebSphere.

About a year ago, I blogged about some .NET / Windows Server benchmark testing results produced by Greg Leake. After taking some well deserved time off, Greg went back to the lab, expanded his testing and agreed to help get to the bottom of the “back office” debate. For the first time, the results include IBM hardware (POWER6) which allows us to evaluate price / performance using typical customer configurations. In an era of cost cutting and the need to squeeze as much optimization out of systems as possible, we think customers will find this information very interesting.

Let’s start with costs… Greg’s findings demonstrate that customers save up to 81% in total system costs by running applications on Microsoft .NET and Windows Server 2008 vs. IBM WebSphere 7 on POWER6/AIX. The study also showed that customers who run their IBM WebSphere 7 applications on Windows Server 2008 and Hewlett Packard/Intel 64-bit blade servers can save up to 66% in total system costs when compared to running the same applications on IBM WebSphere 7 on an IBM Power6/AIX platform. These results illuminate significant savings for businesses of all sizes, and particularly speak to the value of the Windows Server platform. Our hope is that people are able to make use of these findings to get more for their money, either by making new investments or by maximizing assets they are already using. Who knew WebSphere and Windows Server were such a match?

Most folks tend to buy into the potential for cost savings but speculate that they give up performance to get it. The study also found that for the hardware configurations tested the Microsoft .NET Framework on Windows Server 2008 handles 57% more load than WebSphere 7 running on IBM Power6/AIX. We also found that Windows Server 2008 running WebSphere 7 handles 37% more load than IBM WebSphere 7 running on POWER6.

So, what do these findings really mean and why am I sharing them with you today? A few important things to know:

1) Windows Server and the .NET Framework continue to be a powerful, leading combination for application development, deployment and management.

2) Customers who have made a bet on WebSphere can improve performance and reduce costs by running WebSphere on Windows.

Please visit http://www.wholoveswindows.com/websphere/ to read more about these results and check out the .NET StockTrader downloads. We think the .NET StockTrader is a GREAT example of how small bits of innovation can have BIG impact. Don’t take our word for it – these sample applications and guidelines are available to anyone. Instructions on how to replicate the testing we conducted are also available here, and I strongly encourage customers, and all third parties, to conduct tests for themselves. Let us know what you think