 Live from Las Vegas, Nevada, it's theCUBE at IBM Edge 2014. Brought to you by IBM. Now, here are your hosts, John Furrier and Dave Vellante. Okay, welcome back. We're live here in Las Vegas for IBM Edge. I'm John Furrier, the founder of Silicon Amgen. Dave Vellante, co-founder of wikibon.org. It's theCUBE. Our next guest is Suame Koshalakata. Nice, got that one, good you with BNY Mellon. Thanks for joining us. So first of all, share with us your take on IBM Edge. Why is it so important? What's going on here? What's the show about? Well, first of all, thanks for having me. It's a great time to be at a conference like IBM Edge, where you're learning a lot about some of the new announcements that IBM is making. But to kind of take a step back, in general, it's good to be in the infrastructure business right now. It's good to be in the technology business, particularly infrastructure, because of the velocity of innovation. And if you see some of the companies like IBM, what they're doing, we're really happy to see the roadmap that you have and the announcement that you're making as much as the innovation that's happening in industry. Are you impressed with the portfolio refresh? Yeah, absolutely. I think the whole strategy and how IBM is investing the dollars behind the strategy on the software-defined data center and some of the things that Jamie talked about, having that advanced compression that will help the enterprises. What unique thing would you point out, share with your peers out there around that they're doing the strategy with software-defined that's unique and compelling for IBM? What is the point to that lever, the differentiator? See, IBM had a lot of research that was already in their tucks before. And what I found it unique is, the moment IBM decided to do software-defined storage and they were able to quickly say elastic storage and they were able to take some existing ingredients, such as the GPFS file system and use that single name space and quickly launch that as a product. So to me, IBM's ability to take the research that they have and offer that into products is my biggest takeaway. So Swami, I wonder if we could talk a little bit about what we're talking about off camera. This notion of hyperscale. John and I talk all the time in theCUBE, but if you want to know what's going to happen in the enterprise, just look at what Facebook and Google and Amazon are doing. And if you go back to what they were doing five, six, seven years ago, you're seeing a lot of the concepts seep now and bleed into the enterprise. So there's a large financial services organization. How much do you guys track those trends? Can you bring some of those concepts in? What learnings are you bringing in from the hyperscale crowd? That's a good question. We do look at what's happening in the Google and the Facebooks and the eBay space. We really like the way they're using technology to solve a specific problem. You know, unlike a Google or a Facebook where they have to do one thing right, where they have to do the storage, they have to do the search, they have to do the search at scale. But large enterprises, we have to do many things at scale. So for us, we're observing what Googles and the Facebooks are doing and how they're using infrastructure to deliver the amount of things that they're doing. You know, half a billion users on Facebook or how many members that they are. It's amazing to see how they are applying and we're actually able to say what aspect of what they do in the data centers that we can apply in companies like us, in enterprises. The first lesson that we have done is we have a lot of developers and we work on lots of applications at the same time. What we have done is we standardized on how our developers use their development environment into one single stack. You know, much like when you go with Google or Facebook, you get their APIs and their stack. We standardize on how we develop our applications. Now that we standardize on how we develop our applications, we can deploy our applications at scale in our data centers. Now, one of the differences that we've noted between the sort of hyperscale crowd and the traditional enterprise crowd is many, not the least of which is the complexity of your application portfolio compared to theirs, but the hyperscale guys will spend engineering time to save money. The traditional enterprise is very much labor intensive in IT and they will oftentimes spend money on a solution to save time. Is that a mindset that will shift? In other words, will you apply those engineering resources like a Google or a Facebook, or will you look to a vendor like IBM with the PhDs running around to build those products? Will that change? We are actually shifting gates in our enterprise where we're building everything with open source and like I mentioned with that single stack. What that would give us is with that single stack, if we engineer that single stack very, very well and put all of our efforts in engineering that stack that can scale, our developers can write to that stack and by that definition, what they build will scale. So gone are the days where we would watch things happen and go back to big vendors like IBM and HP to give us a solution. No, those days are gone. We are engineering and the revolution in open source community is allowing enterprises to do that. The guy at the other bank said to me the other day, I asked him, what do you want from storage? He said, I want an open source object store that works. You feel similarly and why? We want everything open source because the speed at which that we would want to innovate, we can innovate if it's an open source technology as opposed to working with a vendor, being part of their product roadmap, working with the product team is probably more time consuming than adopting an open source technology that has so much movement and riding the wave of other people's requirements as well. So you see open source as a main spring of innovation, that's what I'm hearing. And is a speed enabler? You think it's faster? Oh, absolutely. And clearly you can discern, I presume, different companies, different vendors, suppliers have different attitudes toward open source, right? Maybe talk about that a little bit specific to IBM, what's your sense as to IBM's commitment to open source and from a customer standpoint, what's your reaction to that? Right, see, the core engine of the technology can be open source, but you need an ecosystem around that. So you could have a turbo engine, but you still need leather seats. You know, I think we look for vendors like IBM to kind of provide that management suite around it so that we can actually manage the complexity of our environment when you deploy this open source technology at scale. What open source tech are you leveraging for your speed and delivery around the value chain, the IT, all the things you're doing? Right, so the single stack environment that I talked about, beginning with spring source, Java runtime, spring framework, the way we use our continuous build environment, the way we manage our dependencies, all of that is being done with the open source tools and now we are looking at the open stack, you know, open stack swift implementation is what we're aggressively deploying and piloting inside the corporation. So definitely open stack is the next big thing that we're looking beyond the application runtime. You say single stack environment, you mean versioning control and that stuff built into it? Absolutely, and the DevOps layer is built into it as well. So in respect of your Java developer, taking your application from the development environment all the way to the production is fully automated with open source technologies. So Swami, we always talk about the 70, 30 mix, 70%. It's spent on keeping the lights on, 30% on innovation. I like to look at it a little differently, maybe breaking it down with when you're running the business, you've got pieces that grow the business and you've got pieces that transform the business. So my question to you specifically is how do you fund new innovation? Where do you find the money to do that? Yeah, that's a good question. So from an infrastructure perspective, IT budgets are not growing. We're either flat or down. In fact, we're able to take costs out because of the price drop and the adoption of the commodity and the adoption of open source. What we're able to do is when we see a price drop in a particular technology or when we adopt the commodity, we're able to give some money back to the business and we're able to put some money back in the innovation and infrastructure that we're able to fund. This cloud environment that I'm talking about, that single stack cloud-enabled environment that we have in BNMLM was actually funded from infrastructure from our own run rate. It was not a top-down initiative where we basically said, let's do this. We really drove the run rate reduction and use that to fund. And I think there's so much innovation that's happening right now that for the next four, five years, I don't need to go back to the business and say, give me money. I can actually use my run rate and save the money and give some to the business and put some money back in the infrastructure. So you'll have a top-down budgeting process every year that says, okay, your budget's flatter down. There it is. Go figure it out. So they're essentially running like any business. If you're under budget, you can spend the money any way you want. Right. Let me give you an example. What software-defined data centers do really take away the notion that the enterprise applications were used to, which is, this is my server. This is my network. This is my storage. We're taking the my out of our language. We're basically saying that everything is virtual. Everything is automated. Everything is elastic. We will give it to you when you want it so that you can be more smart about how you want to use it. So a lot of the money that we are spending today in the enterprise is everything is over-deployed. You want something. You expect a lot of growth in the next four years. We deploy the infrastructure today, assuming that that much growth is going to be there in the next four years. We're kind of turning the things around where if you program to the single stack and everything is software-defined, we can provide infrastructure when you need it as you need it and we can shrink our entire infrastructure down. Now you're speaking on a panel tomorrow. I think David Floyer, our CTO at Wikibon is also on the panel. You guys talking about, I think you're talking about XIV. But so where's XIV fit in all this sort of software-defined vision and elastic vision that you're talking about? So we have XIV in our environment, part of our dual vendor strategy. And then we like the management and simplicity of it. Part of what we like about XIV is the multi-tenancy that XIV provides. It's very simple to manage, which is one thing that is very attractive for us. But more importantly, in enterprises we have different types of workload. At five o'clock after the trading is done, we run reporting. And after reporting is done, we do statements to our customers. So what XIV, because of its high multi-tenancy, it'll allow us to have multiple workloads on the same machine without having to sacrifice the IOPS for having multiple tenants on the same box. And you're using a quality of service capability? Yes, we do. How much do you really push that? You pin bandwidth, latency, or bandwidth IOPS and capacity to an application? Pretty much on the IOPS side. And you're sealing that? So most the IOPS, okay. And capacity, obviously. Okay, and then you charge back? We record all of our costs. We give all the costs back to the business, so we do record all the costs. And we, in the infrastructure, are measured based on how we are reducing the unit costs down for our customers when we charge back. All right, Swami, we have to leave it there. So, John, take it away. Okay, thanks for coming on. I think you really appreciate it. I mean, this is a great session. My brain's full right now. I've been just taking, tweeting away those sound bites. I totally love the integrated stack, single stack. Really, open source is the driver. It's awesome. I think the issue is speed. I think that's one of the things that people aren't always that clear about. And the aha moment for many is actually it's faster. Right, absolutely. And so that's the real innovation. We're just talking to folks that's on two chats right now. A lot of stuff happening on Twitter, around IBM Edge, IBM Innovator, when the next open conference is coming. This is theCUBE. Extracting the data, sharing that with you. We'll be right back with our next guest after this short break.