This is a transcription of an interview with Craig Nunes, VP of HP Storage Worldwide, and Sean Kinney, HP Director of Marketing, discussing HP’s August 23 Federated Storage and V-Class storage announcements with Wikibon Chief Analyst David Vellante. The interview, which was recorded in Wikibon’s Marlborough, Mass., office studio just before the announcement, is now available on SiliconAngle.
DV: We are here in Wikibon’s new studio in Marlborough, Mass., inside the Cube. We’re here with Craig Nunes and Sean Kinney of HP, welcome guys. A very thin Craig Nunes, you’re looking good, thin provisioning, marketing all the way. So we’re here to talk about the new HP announcement, August 23rd. I know you guys have been busy. It feels like we were just at HP Discover yesterday. It was a great event. You guys have recovered from that for the most part. We had the Cube there, of course.
So summer’s always been a big time for 3PAR announcements. Here we are again. Sean, you have a variety of backgrounds, and you joined HP relatively recently, right?
SK: Over a year ago.
DV: So welcome guys. So tell us a little about what was announced today and we can dig into it a little bit.
CN: Sure. So we announced a couple of things. One is kind of building on what we talked about in June, we introduced our converged storage architecture. And what we talked about converged storage is really being Storage Without Boundaries and driving storage much closer to your applications, and breaking down the boundaries between server, storage, and networking.
We’re taking that to the next level today in announcing Peer-Based Storage Federation for HP storage. And what that is really doing is delivering on that storage Without Boundaries theme. Really allowing our customers to really turn on data and keep that up and running wherever it might land, wherever they might need it to be. So Peer-Based Storage Federation is one big part of our announcement.
The other big part is a platform extension to our HP 3PAR Family called the P10,000 V-Class. The HP 3PAR V-Class is a platform that delivers new scalability for your virtual machines, it delivers untouchable efficiency in terms of some of the new enhancements we’ve driven into silicon and software, and we view it with the Peer-Based Federation capability and everything the platform has brought to bear over the last several years as the benchmark for federated Tier 1 storage for client data centers.
DV: Sean, we were talking off camera about thin provisioining. You were not 3PAR Classic, not part of the acquisition team.
SK: No.
DV: So thin provisioning is now part of the portfolio, and you see a lot of companies that have announced thin provisioning. As a quasi-outsider to 3PAR, what has thin provisioning meant to the insider at HP, and talk a little about thin provisioning, and your contention was it’s different from the other stuff out there.
SK: Yes, it’s really core to the 3PAR value proposition because it’s built into the architecture, and as Craig was saying, we’re now in our fourth generation ASIK, which includes the 3PAR buzz word Thin Built In. the difference is that it is built in, it’s basic to the architecture, versus where its bolted on and is called a feature, my guess is it’s not always that well used. With 3PAR almost every customer uses it because that’s kind of why they bought it. And we guarantee, we say give us a chance and we guarantee that you will save at least 505 in your storage or we’ll actually buy the rest. That’s actually a little conservative; we usually do better. But where it’s a bolt-on to check a box in the RFI it’s probably not being used, and I question whether all the other arrays and their customers are getting anywhere close to the benefits 3PAR does.
DV: We’ve quantified some of that on Wikibon. If you search 3PAR on Wikibon we actually did a customer survey where we actually pulled data – it was only metadata, no customer data – out of the arrays and Data FLoyer did a detailed analysis. It was pretty impressive. There’s no doubt about it, you guys pioneered that whole space.
So now you’re getting into this whole federated storage area. Talk about what you mean by “federated storage” and then we’ll talk about what’s different about you guys.
CN: So here’s the background. We’ve come a long way in highly virtualized storage platforms. The benefits that one can realize on a highly virtualized storage platform from tiering and management, etc. Very clear, right. But what our customers still struggle with is in this new data center where you’re deploying large-scale virtualization, you’re deploying your cloud-based architectures, you run into a couple of different situations. One is that data center is very unpredictable in its workloads. You’ve got a lot of diverse and changing workloads. So part of what we hear from our customers is “I sometimes find I’ve got a workload trapped on a set of resources, and I’d really like to get it to a set of available resources somewhere in my data center, maybe even in another data center, but I can’t take the downtime to do that.
The other thing we find is when it comes time to take advantage of storage technology refresh, lifecycle asset management for storage, again in a cloud data center or a virtual data center where you’ve got many applications consolidated, the ability to take you’re multiple applications down, your multiple tenants down, and do that refresh – very painful, very difficult exercise. And then the final thing, and it kind of builds on the point you guys were building on with thin provisioning, we have loads of thin provisioning customers, and what they see as an opportunity is they’ve gotten high utilization results from thin provisioning, they actually see and opportunity whereby they can kind of share free capacity resources across their data center with technologies that go beyond in-the-box virtualization. And that’s squarely where we’re aiming storage federation.
DV: Okay, so it’s actually a collection of independent arrays that you’re managing as one, and you call it a peer-to-peer, right?
CN: So I would describe storage federation as the following: it is distributed volume management across peer-based storage systems – a federation – interacting in native in-band communications between them. They’re relying on no external devices to actually do any of the functionality we’re talking about. Versus virtualization – think of virtualization as sort of a hierarchical technology, a SAN-based virtualization appliance virtualizing arrays beneath it, or within a storage array a storage controller virtualizing disks beneath it. What we’re talking about is a per-based technology, and the advantage that brings you is you don’t have an additional layer of equipment, an additional management point, additional failure point in your architecture. Keep it simple, keep it more efficient, etc.
DV: Okay. And you talk a little bit about the use cases and the problems that it’s solving. It’s not heterogeneous. We’re not talking about an appliance that goes in and managed other storage. So what problem does it solve?
CN: Sean’s got a great take on this. At the end of the day what we’re aiming at is solving the problems of unpredictability and in your cloud-based architecture, and tackling that for our customers in a better way than anyone’s every been able to do before.
DV: So Sean when you talk about cloud-based architecture’s – that’s a term you used previously – what are we talking about? Are we talking about private clouds, or are we talking about cloud service providers?
SK: C – all of the above. So cloud service providers are some of 3PAR’s biggest and best customers, but corporate It is moving that way to, and they are saying, “Hey, I like the flexibility because I have a data center with unpredictable workloads. I need an array or I need an architecture with multi-tenancy built in.” Again, not bolted on. So taking that as a model they say, “But I don’t want to outsource because it’s too much risk or lack of control or my CIO doesn’t want the data to leave the building.” So we’re moving towards this cloud service model and the good thing is the service providers are leading the way and now corporate IT is following. And it’s not nearly as risky as it used to be because other companies have built successful on it, and on 3PAR.
DV: Is the gap closing between cloud service providers and internal IT, or are the cloud service providers going so fast because there’s so much demand that they’re actually innovating faster.
SK: I think they’re innovating on a business model. I would say on the technology corporate IT is catching up. The biggest difference for me is still charge back. It’s the cloud service providers’ business.
DV: Yes, we figure maybe 155 of the people out there are doing charge back. So if you’re not doing charge backs is your private cloud really a cloud? Well really by the technical definition I suppose not. I think more and more people are going to start doing that because they want to see that visibility.
SK: And we’re making it easier. The technology that supports the business also has to enable the business. Now with federation and peer motion as you can add more storage, and as part of the software suite within 3PAR to be able to automatically rebalance the system with the right choice of services at the right price at the right time, not only within the array but across arrays, a lot of that human management and the game of “who’s got the best spreadsheet” is gone.
DV: We love managing by spreadsheet. So the V-Class, Craig, is another arrow in the quiver of the 3PAR portfolio, right? Talk a little about the V-Class and the details there.
CN: So if you step back on what folks are after in this new data center:
- They’re trying to handle this unpredictable multi-tenancy, all those mixed workloads and changing workloads.
- They’re trying to deal with security and quality of service.
- They’re trying to maintain persistent access to data that’s maybe moving around their environment.
- They’re thinking about pay-as-you-go models and keeping capacity utilization as high as absolutely possible.
- And they’re driving on self-configuring, self-provisioning, self-tiering systems to keep the operational costs down.
So you’ve got a set of criteria that frankly is showing the cracks in the architectures of 20 years ago that we’ve kind of been working towards. And with the new V-Class, it’s our latest extension to the family that’s built on that. We view the V-Class as a benchmark for folks who are trying to deliver mission-critical storage capabilities in this new data center. The V-Class extends the line about 50% higher in overall connectivity to I/O and disk. It supports some workloads as much as three-times faster. From a virtual machine perspective it really drives up the scalability of VMs in your environment, and it does all of that with a whole new level of efficiency, built into that Gen 4 ASIK. You get ??? conversion faster, you get a more granular reclamation capability, even your remote copy or DR between arrays get zero detection, gets thin capabilities beyond the thin-aware capability built into it before.
SK: It’s built on a more granular thin provisioning. We used to do it on 128 MB block sizes. Now its 16 K. so whatever that math is – 10,000 times more efficient? So we’re sucking out all that extra, unused, stranded space and giving it back to the general storage pool.
DV: Thin.
SK: Very, very thin.
DV: For the 3PAR platform, what’s the primary competitor that you see in the marketplace?
SK: I would say traditional thinking. The approach of applications, servers, and storage basically in silos in the environment.
DV: As opposed to an infrastructure that’s going to support multiple applications ….
SK: If you have predictable workloads, you can buy predictable storage. If you have unpredictable workloads you need multitenant architected from the ground up storage. That’s 3PAR. So it’s that shift.
DV: Now I know you’ve got some Left Hand announcements coming up. We’re not going to talk about that today. But I want to ask you where EVA fits. A lot of people in the Wikibon community have asked us questions about EVA, what’s the roadmap look like. Can you talk to us a little bit about EVA, what the fit is, and what you’re doing there?
SK: Absolutely. Back in June at HP Discover we announced we are now shipping our fifth-generation EVA platform. We now have over 100,000 units installed in the field. For customers who like EVA, love EVA, we’re going to continue to invest in it, make it better, and we’ve added thin provisioning. You will see more announcements coming from us early in the fall this year, and then into 2012. EVA’s not going away any time soon. We have a loyal customer base; it’s still, according to third-party analysts, 20% easier to use than other traditional mid-tier architectures, and it’s a successful business for us with loyal customers. So we’re not going to walk away from it.
DV: So Craig, do you miss being a $200 M public company? How does it feel being back part of HP? How/s it going?
CN: Ah no, to your first question. So 3PAR at HP. What 3PAR’s done for HP is, I hope, pretty visible. 3PAR is an anchor storage platform, it is a basis for our path to the cloud and cloud systems. It is a part of our best-of-breed virtualization stack and virtual systems. It is a part of our mission-critical metro and continental cluster offerings. It has very rapidly become a fixture across the business.
What HP has done for 3PAR is give it the avenue to really grow and touch far more customers than a $200 M company was able to do. Thousands and thousands of channel partners, thousands and thousands of sales people, all making calls with 3PAR and having a conversation that to be honest is a conversation, especially for our EVA customers, feels like a big brother. It’s a virtualized platform with a lot of ease-of-use benefits, and feels like a perfect fit in the family. So from that perspective the growth we’re seeing is no surprise at all to any of us.
DV: That obviously was a huge acquisition. We covered it very closely and are very high on it. I’ve said that a better use of cash many times than R&D is acquiring a successful company. So good luck with that.
We’re going to be at VMworld and have you guys on, and maybe have more discussion about federation, maybe dig into it a little deeper. That would be great. So we’ll see you back in Vegas. So Sean and Craig thanks for coming on the Cube and talking about the announcement. Look for coverage on Wikibon.org. We’ll be covering this announcement like a blanket like we always do. And good luck with it.