 We're back in Vegas, I'm John Furrier with my co-host. I'm Dave Vellante at wikibon.org. We're here with McLeod Glass. And we're going to talk about the changing nature of the server design, the IO stack, the big problem. Dynamic workload. Talking about here. It was interesting to note that the announcement package that you guys had and the messaging that you had started with this whole data tsunami, that's a change in the server world. And you had to really rethink the way in which you were approaching workload management and system architectures. Talk about that. Yeah, yeah. So I mean, when we looked at this generation of servers, one of the things that we had to go address was this gap between, if you look at compute and memory and how that has grown in terms of performance with Moore's law and how storage has continued to lag in performance behind that. And so some of what we looked at was optimizing around solid state. Also drawing out the intelligence within that system and being able to help close that gap with the intelligence. And actually, the intelligence is actually workload aware. And with that workload awareness, we're able to actually go in and optimize performance and efficiency within the system with that intelligence. For example, if you look at virtualization, typically that would look like a random data within the system itself. But we have the ability to look into those virtual machines and actually see sequential data, reorganize that data, and actually use it and resequence it, and actually optimize performance based on the actual type of data workload that's going on. So what's the efficiency advantages with the Gen 8 platform now that it's solid state optimized? Yeah, so part of it is the ability, if you look back previous generations to go drive some of the performance numbers from a workload standpoint, you had to put as many as 2,000 spindles in that system. Today, with solid state optimizations, you're able to cut that down to around 100 spindles. So you think about the power and space savings from that. And the one example that we use, we were able to go from four racks of servers and storage down to one rack. That's about 88% power savings and a 75% space savings to drive a 50% higher performance in that solution. What's the trade-off there? Is it sacrificing capacity? I got to pay more because there's less spinning disk, more flash. What's that equation look like? Yeah, yeah. So typically, customers would have to trade off. If they want performance, they've got to trade off capacity. And it's very expensive when you talk about solid state. One of the great things about the foundation that we've laid down, we're able to deliver performance gains today, but we've also put a really good foundation down for the future in terms of what we're going to be able to deliver. And that is a balanced architecture with the ability to kind of right size or cut, eliminate over provisioning of expensive flash. So for the amount of hot data that you have in your system, you'll be able to optimize the right amount of solid state media. And then when you've got high capacity, you'll be able to put that storage onto that rotational media and then also establish a new economic and new cost paradigm in terms of being able to do that. Today, it's very, very expensive to try and drive that. So talk about the software that's doing that automated management and movement. Yeah. So it's based on our SmartArray technology. We delivered the first SmartArray back in 1989. And we're using that same technology and continuing to develop that to basically deliver this. And that's where the intelligence resides. So that's 1989. 1989. Autorade? Goes all the way back to our first servers that we delivered in 1989. I was there at the time. I worked at HP in 1989. Yes. Yeah. And it's a great legacy and heritage. When you talk about storage, we've delivered advanced data guarding, which was kind of the precursor to RAID 6 in the industry. And now we've got some cool stuff, like advanced data mirroring, that we're announcing as part of this, which is actually a three-drive mirror and is a much more resilient and performance-driven solution. You don't have to make the trade-offs. You get performance plus resiliency. So this isn't kind of an off-the-wall question relative to this conversation, but I want to ask it anyway because it's related to this. Is that there's a slew of, or some say, bubble in the startup ecosystem around Solid State. A lot of companies coming out with innovation around RAID, this version of that. How do you guys see that emerging startup field? Do you see any opportunity for the guys coming out of there? What innovations do you guys see that are coming out of the market? Because you guys have a lot of IP in storage, you mentioned. What are you guys seeing in the marketplace? Well, I mean, I definitely, I mean, we wouldn't have driven at this optimization around Solid State if we didn't anticipate that this is going to continue to grow. It's just, in order to close that gap, as I mentioned before, of the compute and memory, being able to close that gap to the storage performance, Solid State's really the way that you have to go. Is it play at the server? Is it move more into the network side? Is it a caching layer? Is it just a caching layer? You see more, more strategic use? Well, I think HP is uniquely positioned in the fact that we own the RAID stack inside the server and all the way out to the external storage. You talk all the way out to three par and we have all of the IP necessary to drive some very unique solutions and being able to drive the caching close to the compute, local to the server, as well as being able to take advantage of the shared storage. I mean, data these days is about volume, velocity, hot data, active data, fast data, as we say. And we coined that term in Sapphire. Was that changing some of the architecture? Is this, you guys just business as usual? Yeah, I mean, I think it's just an extension of what we already do. If you look at what SmartArray was built on, we've been caching in terms of within that controller for a long time, and it's an extension of those technologies that are gonna get us there. So you guys were quoting, I saw some of the numbers in the announcement, like a 6x improvement on the SmartArray, but was that on a RAID 6? Well, can you just detail up that, Yeah, so that's a RAID 6. So comparison of solid state to solid state. If you look at a system level, we're talking about in the previous generation products running about 75,000 IOPS, going up to 500,000 IOPS with this new server. So how do you, how does the data protection schema change at all, or does it, with this sort of new flash-based architecture? I mean, I think that, obviously there's some on some of the resiliency today is based on the fact that you had to have RAID for rotational media and you had failures. One of the things with solid state still today, you still have a probability of failure because of wearout. One of the great things that HP has is this SSD wear gauge, which actually can predictively let you know that you've got, that your drive is going to reach it's end of life and let you know ahead of time so you can actually go replace that drive. And that's another HP unique technology that we're really excited about. You can automate that by getting some robotics into your data center, having a little arm going and pull it out. Yeah, you guys are doing a great job. We really appreciate you coming on theCUBE. Thanks so much. I'm Claude Glass, the product marketing for the Industry Standards Service. Thanks so much. Great job, great announcement. Congratulations. Claude, great to meet you. Yeah, yeah, great to meet you. Thanks for coming on theCUBE. We'll be right back in one minute. Okay.