Storage Performance Transcript

SearchStorageChannel.com has posted the transcript of our podcast with Tony Asaro of Enterprise Strategy Group. The podcast itself is worth a listen, but if you prefer to read the details, point your browser here:

Transcript of Storage Performance Podcast with Tony Asaro of Enterprise Strategy Group

Enjoy!

Comments

[zax]

"Gary Orenstein: If I'm a storage administrator how do I know if I have a storage performance problem? Where can I check and are there tools that I can use to evaluate parts of my environment?

Tony Asaro: All you have to do is either look at your emails or answer your phone. Probably one of your users is calling you up and screaming at you. That's not really a joke; that's the clearest indication. It's all about user and customer satisfaction and one of the clearest ways to measure, without getting into the technology or anything, is to see if people are complaining. You may not even have a performance problem even though you think it, but as long as your customers are happy, you don't have a performance problem. I would say that most users have very little tolerance for slow performance, so you will get people complaining to you if there is a performance problem.

The other answer is that, yes, there are storage management tools that allow you to analyze performance and then you can troubleshoot and things of that nature. Some are better than others. Some will be able to give you performance analysis from the application all the way to the disk drive. Some of them will only give you performance on what is going on in the storage system itself, which is very limited. So the notion is that if you can get a view holistically, because performance is a chain, there are a lot of variables involved in it, from the application all the way to storage. "
----------------------------------------------
LOL...Is Tony Asaro sharing a new revelation with us here? You mean to tell me that there is a performance problem, when the users tell us?

Most Enterprise shops are becoming more proactive in diagnosing performance problems. If the user is the first to know, then there is a problem. It would be more useful to see something like this:

If your latency is over XXms for a period of X minutes, then you should look for possible performance problems. Or if your controller CPU's are over XX% for a period X minutes, then you should look for possible performance issus.

I realize that these numbers may vary depending on the vendor and application, but some generic best practices would be of more assitance then what was provided.