Is Storage Performance Predictability when building VMWare Virtual Desktop (VDI) Storage Clouds important? This can also apply to Citrix and Microsoft Windows Hyper-V Virtual Desktop Systems.
Here is yet another great example of why I just love my job. Last week at our Xiotech National Sales Meeting we heard from a net-new educational customer out in the western US. They recently piloted a VDI project with great success. One of the biggest hurdles they were running into, and I would bet other storage cloud (or VDI specific) providers are as well, is performance predictability. This predictability is very important. Too often we see customer focus on the capacity side of the house and forget that performance can be extremely important (VDI boot storm anyone?). Rob Peglar wrote a great blog post called “Performance Still Matters” over at the Xiotech.com blog site. When you are done reading this blog, head over to it and check it out 🙂
So, VDI cloud architects should make sure that the solution they design today will meet the requirements of the project over the next 12 months, 24 months and beyond. To make matters worse, they need to consider what happens if the cloud is 20% utilized or if/when it becomes wildly successful and utilization is closer to 90% to 95%. The last thing you want to do is have to add more spindles ($$$) or turn to expensive SSD ($$$$$$$$$) to solve an issue that should have never happened in the first place.
So, let’s assume you already read my riveting, game changing piece on “Performance Starved Applications” (PSA). VDI is ONE OF THOSE PSA’s!!! Why is this important? If you are looking at traditional storage (Clariion, EVA, Compellent Storage Center, Xiotech Mag3D, NetApp FAS) arrays it’s important to know that once you get to about 75% utilization performance drops like my bank account did last week while I was in Vegas. Like a freaking hammer!! That’s just HORRIBLE (utilization and my bank account). Again you might ask why that’s important? Well I have three kids and a wife, who went back in to college, so funds are not where they should be at…..oh wait (ADD moment) I’m sure you meant horrible about performance dropping and not my bank account. So, what does performance predictability really mean? How important would it be to know that every time you added an intelligent storage element (Xiotech Emprise 5000 – 3U) with certain DataPacs you could support 225 to 250 simultaneous VDI instances (just as an example) including boot storms? This would give you an incredible ability to zero in on the costs associated with the storage part of your VDI deployment. This is especially true when moving from a pilot program into a full production roll out. For instance, if you pilot 250 VDI instances, but you know that you will eventually need support for 1000, you can start off with one Emprise 5000 and grow it to a total of four elements. Down the road, if you grow further than 1000 you fully understand the storage costs associated with that growth, because it is PREDICTABLE.
What could this mean to your environment? It means if you are looking at traditional arrays, be prepared to pay for capacity that you will probably never use without a severe hit to performance. What could that mean for the average end user? That means their desktop boots slowly, their applications slow down and your helpdesk phone rings off the hook!! So, performance predictability is crucial when designing scalable VDI solutions and when cost management (financial performance predictability) is every bit as critical.
So if you are looking at VDI or even building a VDI Storage Cloud then performance predictability would be a great foundation on which to build those solutions. The best storage solution to build your application on is the Xiotech Emprise 5000.
Another great article Tommy! It is amazing that people (and vendors) will push performance to the side so they can sell a new buzzword…as you’ve proved, you need the building blocks to make the most of your environments, TP et al won’t do you any good if you can’t fill your storage past 60-70% or you can’t handle more than a handful of VM’s at boot.
Thanks JMO !! I appreciate the feedback. I agree, performance and capacity utilization is pushed to the side in favor of “bells and whistles”. Great, you want to use TP then reserve 10% to 20% of your capacity for “Margin Calls” and alert thresholds. You want to put your data on the outer tracks to solve the vendor’s architecture performance issue, just reserve another 10% to 20%. Eventually you are lucky if you can use 60% of the capacity you purchased!!
Whats interesting is VSphere has a great Thin Provisioning feature that really negates the need to run it on the storage array. Don’t get me wrong, we have that feature as well, but my opinion is letting the OS own it is a better way to go. First and foremost it has ZERO impact on reserve capacity on the array.
Thanks again for the feedback !!
With NetApp’s dedup, during a boot storm the deduped data is resident in cache, greatly reducing the required back end disk IOs.
Thanks TechMute for the question !!
That would work – would you need PAM to make that happen? Also, I’m all for cache fixing the ills of legacy storage systems (NetApp/CX/CML), but with ISE we don’t need it. Our ISE/Emprise 5000 is 3U and can support about 250 VDI boots at once. It should be noted, we can support a LOT more VDI clients then that – probably closer to 500/600 but if you size to the bootstorm – 250 is a great number for 3U of space.
*Note: I am not a NetApp Guru*
To my knowledge, no, you would not need PAM. In my understanding, PAM is like a giant Tier 3/Tier4 cache. As cached blocks “expire” from cache, they trickle down into PAM.
I need to shoot an email to Peglar. So far, I can’t find much tech doc except for the marketing whitepapers on why XIOTECH is different wrt to performance, etc. I don’t trust “FM.”
(Where FM == flipping magic)
Hey TechMute – no problem on the NetApp part – for all I know I’m totally wrong. As far as Peglar goes – he would be a great resource. If you really want to understand how our magic works then check out the following RAGS (Redundancy Allocation Grid System) patent:
It’s about as technical as you want to go 🙂 Actually, I think I’ll work on a RAGS (Secret Sauce) Blog to help better explain a little how we-do-what-we-do.
In the mean time – did you read this whitepaper:
http://xiotech.com/pdf/analyst_reports/Xiotech-Building-Autonomic-High-Performance-SANs-via-Intelligent-Storage-Elements.pdf ? I think it might be buried in on our site so you may have missed it.
The good news is we have vendor neutral validation of our performance claims. Storage Performance Council is a 3rd party organization that requires full disclosure on their test results. Notice that we aren’t “short stroken” our drives to get these results. No outter track tricks etc.
Click to access ae00001_Xiotech_Emprise-5000-146GB_SPC1E-ExecSumm.pdf
Pingback: How Xiotech can help you build resilient, scalable storage clouds and turn your IT department into a profit center « StorageTexan's Blog
Pingback: VMworld 2010 « Pjselin's Blog