 Live from Las Vegas, it's theCUBE, covering Dell Technologies World 2018. Brought to you by Dell EMC and its ecosystem partners. Welcome back to theCUBE. We are live on day two in Las Vegas at the Sands Expo Center at Dell Technologies World. I'm Lisa Martin with Stu Miniman and we are welcoming to theCUBE for the first time distinguished engineer and VMAX product group CTO, Adnan Sahin. Adnan, it's great to have you on theCUBE. Thank you, great to be here. So big announcements going on at the event this week. Talk to us about the Modern Data Center. Saw that press release this morning. What does the Modern Data Center, how does Dell Technologies define it and how are you seeing and helping customers implement it? And again, it goes into like when you have running an application, you need compute, storage and network. So that really have to have a modern infrastructure to cover all those bases. So that's what really doing the keynote we heard today from Jeff Clark on our compute capabilities, new servers, as well as new storage offering from Dell EMC, the PowerMax. All right, Adnan, PowerMax, let's start there. So one of the flagship announcements here at the show, building on the VMAX, which of course builds on the symmetric history. Why don't you give us a little overview and then we'll dig into some questions I have. So yeah, we've been really following the industry trends. So we introduced VMAX All Flash a couple years back and then we're also looking at the industry trends and what we realized is the industry is transitioning in terms of media interface from SaaS connected drives into NVMe and PCIe connected drives. And the main driver for that one is this two-fold. One is reducing the latency. And with NVMe you can get much leaner software layers that really gives you lower latency. And the other one is the media transition that changes from an end-based non-volatile memories technology into newer and emerging low latency, ultra low latency technology. So with NVMe we can get both at the same time. All right, so Adnan, we remember back when it was EMC at the time, it came with the flash technologies and everybody's doing flash. Now anybody in the storage industry, NVMe, NVMe over fabric, everyone's talking about it. PowerMax Bob up on stage. Bob, you presented it though. Many, many years working on this said, there's a big difference between just having it and really being able to utilize it. So without going too deep, explain to us some architectural things that have to happen from in a hardware and software standpoint to take advantage of this transition. Sure, so VMAX and PowerMax is really scale out multi-controller architecture. Therefore we need to have persistent storage accessible through multiple controllers, at least two. So in order to really be highly resilient and high available to system, we need multi-ported, dual-ported drives available to us. One of the things that we spend quite a lot of time is to really make sure that dual-ported NVMe drives are ready for our highly demanding enterprise resilient storage system. So we spend a lot of time improving drive quality as well as our software to handle all the NVMe related concerns. So at the event, the theme, Make It Real, yesterday Michael Dell kicked things off and talked about these four transformative elements where customers need to transform to be successful. Digital, IT, security, workforce. With what you just talked about and the new enhancements to some of the technologies, how is that helping customers make their digital transformations effective so that they can deliver differentiated products? I know you just came from a customer meeting. Maybe give us an example of what you're seeing applied out here. So the important thing is the applications. The customers have been using databases, many different variants. There's been some traditional databases and new and emerging databases. Main driver is, of course, availability and resilience, but at the same time, operational simplicity because we have simplified our user interface and overall user experience significantly over years so that with fewer people they can manage many, many larger capacities of the systems. And then with latency is another important aspect of application experience. If you lower the latency, either through caching or lower latency media, you give better experience to end users and therefore they can do more with what they have as the infrastructure. And on any specific use cases or verticals where you're finding that, especially the NVMe offering is going to be most helpful at day one? I mean, I think any traditional application that has, let's say, like journals' available benefit, but on top of that, if you're running large queries of random IO access storage, they will get lower latency out of NVMe-based systems. They could be real-time analytics, for example. You can get ultra low latency from the back end. And also if you're using some of the database, data warehouse type of applications, you can get massive bandwidth out of VMAX and PowerMax systems that helps you process more in shorter time. Yeah, so analytics as a use case, attach the storages, of course, really interesting one, heavily growing. One of the other interesting things about PowerMax is, I think it was discussed, it's predictive analytics inside. So, you know, I think back to, I mean, disclosure, I worked at EMC for 10 years. We think of intelligent storage was something we've been talking about for a long time. Explain what's different about this generation of analytics and predictive compared to previous storage innovations. So, we have the infrastructure to keep track of workloads at very fine address granularity. So, we keep track of access types and access sizes in as small as five megabytes a piece. So, in a larger system, that could be 40 million data sets for a 200 terabyte system. And once we have that data, we can analyze and we have some linear regression time series analysis that we can predict whether an active address space will remain active or whether a cold address space will remain cold. And based on that, we can make decisions. Previously, we were able to use those decisions for tearing between hard drives and flash drives. More recently, we are using it for data reduction technologies, for compression, for example, or deduplication. If data set is highly active, we don't necessarily compress them because they will be updated frequently. Therefore, the CPUs used will not be effective. So, and then taking to the next level when storage class memory becomes available, we will be able to use the media based on their strengths. So, if for storage class memory low latency, we can place read heavy and write heavy workloads into storage class memory. So, giving customers presumably the ability to take data, use it as a catalyst in many different lines of the business to combine it, recombine it, and be able to use the analytics that are built in, it sounds like, to not just get insights they can take action on, but actually act on them. Correct. Give us an example of a customer that's maybe doing that to be able to deliver a differentiated product or service to their customer. So, one of the important features that we are introducing that's available both in PowerMax and VMAX systems is service levels. So, that's very relevant to most or all of our customers, because for example, if you are a service provider, customer service provider, even though all the data resides on very fast NAND flash media, they can still provide differentiated performance to their own tenants. For example, if a tenant is paying a certain amount, they may get silver or bronze service level. They may not see the full benefit of flash at that service level, but when they are upsold into a higher performance level or service level with a simple change in Unisphere, for example, they can get flash response time right away. So, it's basically changes and simplifies their business models and makes it more predictable for them. Another one is also the prioritization. They can also set priorities for applications. As long as high priority service level gets it is response time, expected target response time, everybody will enjoy low response time. But if the higher priority group or application does not meet its targets, then we start to increase response time of lower priority applications to give more resources for higher priority applications. So, that's really a way that customers can capitalize on this feature. All right, Adnan, I wonder if you can give us a little bit, dig into NVMe, NVMe over Paverick, and you talked about storage class memory, specifically looking at availability, maturity, and what's kind of the pricing considerations for these that we could expect kind of today and through the next kind of 12 months? Yeah, so NVMe, as like the interface that drives themselves at this day, they may be at a premium compared to SaaS, but the expectation when we talk to industry leaders and vendors, there will be crossover expected very soon. So, that's really the positioning that we just want to be in this market, get a product out, and then really be ready for when that crossover happens. In terms of storage class memory, again, it comes at a premium, but then we, using our intelligence, if we can direct most of the IOs to this premium storage media, then we can let customers enjoy the benefits of that extra premium that they would endure and they would have to pay. But over time, just remember early days of flash when the first flash came out, it was very expensive at the time, but over time, it became more and more prevalent. So, our expectation is storage class type of memory. Over time, we'll follow a similar path and it will become very possible in the near future that we will see all storage class memory systems coming out of vendors. All right, and how about the NVMeover fabric? NVMeover fabrics, we are looking definitely, we have plans for NVMeover fabrics. Of course, standards are still evolving and also for enterprise customers, there's concerns around multipathing support and maturity of that. We are working with standards bodies and other vendors on improving that aspect. Okay, so, there's one thing about this transition that's a little different than most. It has an impact on the application. So, where is Dell getting involved or how are you working with your customers? So, you talked about getting ready for that storage class memory. This is not just, we've been using SCSI for a long time. So, how do we get ready as an industry? What's Dell's positioning in that discussion of application? I mean, Dell teams various across the board are participating in standards bodies and with the industry thought leaders on really getting to common standards-based solutions. I think that's one direction that we're going after with this. Anything on the application side, though? Is that more on the pivotal VMware side? Of course, VMware, we have very deep discussions with VMware on NVMeover fabrics and how we can handle work with them more efficiently. So, then when we kicked off the segment, we talked about it being the first Dell Technologies World and indicator of the absorption of the EMC Federation. We are over your shoulders, the Dell EMC Partner Program. What are some of the feedback that you're hearing from partners, technology partners who are collaborating? You mentioned VMware. What's some of the feedback that you're hearing at the event in terms of what you've announced and how do your partners help influence the design of these leading technologies? That is great excitement. I mean, we've been working with them, listening to them, learning from them, and I think overall, everybody's excited with the new product and we are also, there's a group very excited with NVMeo we've been working for for a while and we are happy to be able to release the product today. All right, Adnan, one of the other product lines that there were a bunch of announcements around was around the Xtreme IO and the X2. Could you just help us make sure we understand positioning today of things like VMAX and PowerMAX and the Xtreme IO family? Yep, and each product platform has strengths and if customers are happy with what they're using, they should continue with the same product line. I think that's really makes it easier for everyone. And Xtreme IO, I believe, they announced a remote application as well so it's great. From a foundational perspective, what are these technologies going to be able to do to enable enterprises to start taking advantage and realizing the possibilities of emerging technologies like machine learning, artificial intelligence, IOT? Yeah, I mean, I think important part is if you look at all those things, what is really needed is ultra low latency, high bandwidth capabilities from storage because you have massive compute capability. Sometimes customers use in-memory applications as well and we need to be close to compute as close as possible to the memory. It's not always possible, but we want to get to be there. We have significant value add to be clear. For example, we have local and remote replication capabilities. If you're running any of those applications in a machine critical environment, you want to make sure that you have local replication capability as well as remote replication, disaster recovery, business continuous models built around it. And what we have with our infrastructure is to really give customers that type of machine critical. If you cannot take any outage in this day and age with the applications. Yeah, Adnan, I got to talk to Jeff Clark earlier today on theCUBE and he talked about the engineering culture. So, from the EMC side, I'm curious if, working with your, now the Dell team, you've got that whole server team, has that changed some of the processes there? How does that impact both the development and the viewpoint from the engineering team? There's much, very clear, much better communication. We've been talking to the server team very, very easily and very frequently actually, just to make sure that, for example, we understand their challenges and then the type of solutions they come up with on the server side and how we can apply on our storage. And the same from our side, we give feedback on our experiences on the storage to them. And not only with the server side, but also across different portfolio components on our storage business unit as well. So, last question, customers that are here, maybe in the early stages of transformation and are looking for best practices, where do we start? Do we start with transforming IT to make it into a profit center? What are your recommendations? Can you repeat, I can hear the last one, the IT? Yeah, what are your recommendations for customers that might be at the very beginning of their transformation journey? What do you recommend? Where do they start in terms of going, hey, we've got our business leaders recognize IT should become a part of our business strategy. It shouldn't be a cost center, it should be a profit center. How do you recommend they start these conversations with Dell EMC, Dell Technologies to get- They just need to talk to their representative about business need and application needs, right? So, we have a large portfolio of products available to our customers. And again, on the high end with resilient storage, with remote replication capabilities that might be VMAX on the mid-range, it could be either Unity or Storage Center. And on the server side, again, similar types of options available. They just need to talk about their application needs, virtualization needs, storage needs, hyper-converged versus traditional block storage versus file storage connectivity. Those make all the difference. And I think our field people have experienced in really helping customers out. Well done, thanks so much for stopping by and sharing with us what's new with the technologies. We appreciate your time. Thank you. We want to thank you for watching theCUBE. I'm Lisa Martin with Stu Miniman. We are here live at day two of Dell Technologies World from Vegas, stick around. We'll be right back after a short break.