Rich Tehrani of TMCnet Interviews WHIPTAIL’s Director of Business Solutions Matt Goldensohn at Cloud Expo 2012 New York. Tehrani asks Goldensohn what a typical Cloud Expo attendee would see at the WHIPTAIL booth and how the device accelerates virtually any application. TMC also filmed a personal demonstration of what Cloud Expo attendees were able to see during INVICTA demonstrations.
Rich Tehrani: Rich Tehrani here with TMC. We’re at Cloud Expo and it’s Cloud Expo 2012 at the Jacob Javitz Center. We’re at the booth WHIPTAIL on our program right now is Matt Goldensohn. Matt, welcome to the show.
Matt Goldensohn: Thank you, it’s been a pleasure.
Rich: Matt I’m an attendee that actually was at your show or at your booth and I went home and I want a refresher course or I want to share it with my colleagues or I missed what’s happening in the booth. Tell me what a typical attendee would see here in your booth.
Matt: Basically what we’ve been doing is walking people through what the array is, what it’s capable of and just a little bit on how it works, not in so great depth, but enough to show what we’re doing, what we’re doing that differentiates us from other venders.
Rich: You have a device, which massively accelerates virtually any application.
Matt: Absolutely. The idea behind the device is that you can consolidate performance as opposed to consolidating capacity. One of the things that we’ve been doing during the show is to demonstrate a base level performance and then adding workload on top of that and showing the overhead. If you have multiple applications that have low latency or a high bandwidth requirements, you can consolidate those kind of applications onto our platform and make them go faster.
Rich: Excellent. With that, let’s see a demo of what you’d show someone.
Matt: Sure. Basically, the array is constructed of what we refer to as a storage router, which sits on top of a series of nodes. Each of those nodes is about 1.3 up to 12 terabytes per node. We can expand up to 6 nodes or 72 terabytes total. Then, depending on which configuration and how you’ve carved the LUNs out, you can have capacity that varies with your requirement. If you have very high IO capacity requirements, you can stack your devices almost vertically across your nodes so you can leverage the interfaces on each of the nodes. What we do to demonstrate that is I’ve been running iometer in the background and that adds a consistent workload and then we’ll start 600 desktops simultaneously just to give people an idea of what we’re capable of doing.
I’m going to do that, just power them on and bring the lap…This is our performance monitor over here and as these desktops are starting behind us, you’ll see the IOPs showing up on performance monitor. In about 15 seconds all 600 desktops will be started. If you notice, these things are averaging around 6,000 IOPs with a steady workload.
Rich: That’s megabytes per seconds I see on the side.
Matt: We’re doing about 27 megabytes per second. Now, remember that the way the array is configured for the show; we can handle up to about 320,000 IOPs so we’re just barely scraping the surface. Now all the desktops have been started. Our peak IOPs was around 8,000. That should leave you somewhere in the vicinity of about 315,000 remaining IOPs that are available capacity. If you need to consolidate applications and leverage the performance capabilities of the box, you’ve still got a lot of headroom and, by the way, we started 600 desktops in about 15 seconds.
Rich: That’s awesome. What else should we know about your company?
Matt: The whole idea behind the company, (laughs) behind the products, is if you need to consolidate your applications from a performance perspective, this is the platform to do it on.
Rich: Thank you so much for your time today.
Matt: Thank you.