 Live from Las Vegas, Nevada, it's theCUBE at IBM Interconnect 2015. Brought to you by headline sponsor, IBM. Okay, welcome back everyone. We are live here in Las Vegas at IBM Interconnect. This is theCUBE, SiliconANGLE's flagship program. We go out to the event and expect a significant noise. I'm John Furrier, Mike Oz. Dave Vellante, our next guest is Don Bully, a VP of cloud services with IBM. Welcome to theCUBE. Thank you. So cloud is hot, cloud services are everywhere. So cloud seems to be sprinkled in all the different users in IBM. It's the hottest topic here at Interconnect, of the three shows. The top trending conferences we've been tracking over the past, you know, three months. And now on site on the show for the past three days has been cloud, internet of things, and big data. So where does cloud services fit in there? Explain what your role is in the organization and what you're focused on. How does it connect to the new way, the new IBM, the new way IBM is doing business? Right, so as you said, cloud is everywhere. So that's certainly the case for pretty much all the units within IBM. I'm from the systems and middleware unit. And what we talk about with cloud services is really moving our traditional middleware closer to the cloud, deploying in different types of cloud, deploying as a service, and also integration between clouds. So the integration for clouds really is that kind of bridge between the system of record where most of our middleware runs today and systems of engagement where you see things like Bluemix and public cloud via software. So that glue, if you will, between those two is really where the cloud services is. Is systems of engagement a product or a concept? It's a concept, it's an approach. Framework, mindset. Well it is, and there's a little bit of a speed element to it as well. So when we talk about systems of record, typically think of your bank data, your airline reservation, these are things you don't want to get screwed up and so you have to have them. Database middleware, the full hardened solution, like a data lake. Correct, so you go into the systems of engagement and now speed becomes the essence, right? So how fast? More dynamic. Yes, how iterative. And speed is typically combined with context and analytics and insight. So that combination just sort of runs at a very fast pace. Like a data ocean. It could be a data ocean. Like a data lake, as you get that in there. That's our debate all week. Do you like that term, data ocean? I don't dislike that term. Or do you prefer a data lake? No, what do you think about it? A data lake? I hear a data lake more often than I hear a data ocean. Yes, we hear it too. But we sort of question data lake. We think data ocean is maybe a better option. Well we don't think the music's exclusive, but data lake to me reminds me of batch. There's no speed boats on the lakes, there's no waves, no whales. So yeah, ocean certainly gives a bigger connotation to it. I'll agree with that. Really management becomes the issue though, because if you don't then it becomes the data swamp, right? And nobody wants that. So that's really the real key to. But real time, the concepts of in streaming, the stuff that's going on in cloud, in software now, dynamic, agile, you're seeing all kinds of stuff going on in the software life cycle, and then how software has to be more adaptive, more intelligent. That's the cloud promise, right? That's like the new middleware. The new middleware is changing, pass in cloud. Things of that nature. So that's a complex area that's kind of in flux with innovation right now. Can you share your perspective on this dynamic area? Because there is new waves of data coming in. That's uncontrollable. Retail has Black Friday, there's surges, there's predictive analytics to get forecast on what's going on and that changing sea of data, if you will. Yeah, so services and data both. I mean, I think what we find in the middleware team in particular is you have to reimagine some of these things. The core elements are still there. So people still need run times, they still need databases. But when they put them into a cloud context and a cloud native type context as well, the management system's different. You pointed out the volatility of kind of the Black Friday could occur at any time. That tends to be the profile of most mobile apps. So you need to really rethink how some of these things end up getting delivered. But at the end of the day, people still have run times. They had run times in the system of record. They have run times in the systems of engagement. The difference I think is, again, the speed and velocity, the tool set you use. So we talk about DevOps in both, but the speed and cycle of DevOps in a traditional environment versus a cloud tends to be an order of magnitude of difference, right? So that's kind of the marrying of those things. So you talked about, Don, middleware being sort of a linchpin of portability across different types of systems, systems of record, systems of engagement. But is it also not as well the linchpin for hybrid, cloud, on-premise, public? Yeah, so the portability is key. At the end of the day, that's really what people want is choice and flexibility. So if I decide to deploy a workload here today, I may decide to go to another cloud tomorrow. I may decide to pull it back for different reasons as well. So that element of portability is there. There's also what I said before about the different speeds with which certain clouds work, right? So again, if I'm talking about a traditional environment, that's going to have a set of processes against it in the new world with mobile, that's going to iterate much faster. And we use integration as kind of the buffer between those two. So if I want to connect back to my traditional database, maybe my mainframe data, that's where most of the interesting stuff still lives. So my mobile app is really with a presence that people see on the engaging side, but anything interesting has to flow back. So how do you manage those areas between the two clouds, so to speak? So that element of integration in addition to portability. So interesting, let's define interesting. Interesting is value, high value transactions. Yep, at the end of the day, it's still about transactions, right? It's the systems of record that document all this stuff. Yeah, we can abstract as much as we want. Tom, Rosemilia, all this says, the infrastructure still matters. I mean, you still have to run it somewhere, right? We may not talk about it a lot, but it matters when it comes down to people deploying real workloads. It's the same with systems of record. That's where the interesting data is. You still, at the end of the day, want to close business, make new business models and new revenue, and those are going to be transactions. So can we unpack the arguments middleware portfolio a little bit? Sure. And sort of take us through the audience. So yeah, so we have a pretty wide-ranging set of capabilities there, right? So in our WebSphere portfolio, for example, we have run times, messaging, integration, in particular, tend to be kind of the major themes. And all those things have a presence in both a system of engagement and a system of records. So it's not just one or the other, but how do you blend the two? When you look at parts of our portfolio that deal with things like performance monitoring and really the management of IT systems, those also have presence in both traditional as well as more cloud-native type use cases. Because at the end of the day, if you're managing something, you want to have insight end-to-end on that transaction, that flow, even if it starts out in the cloud with mobile and ends up back at the mainframe, you still want to end up with good visibility across all of it. So we talked about the infrastructure as a service layer, the PAS layer, or the SaaS layer. Are you obviously contributing to the PAS piece? What's different about the cloud piece and the systems of engagement? I mean, sometimes I ask myself, okay, did companies like IBM just take what they had, cloud wash it, and say, oh, this is part of our cloud portfolio. I'm inferring the answer is no. You had to make some changes to the deep tech. So talk about what those are, give us some sort of detail there. So yeah, so as you said, systems of engagement, which primarily driven by things like mobile are kind of the what? Cloud becomes sort of the how, right? So how do you get that done? What kind of systems and infrastructure and platform do you need underneath that? When you talk about our middleware, the kind of the reimagining of it for some of these new use cases in systems of engagement, typically what we end up doing is pretty deep changes to how it's managed, how it's deployed, how it behaves in an operational environment. So again, traditional examples of middleware, you typically had tight control over things like scaling and availability and those characteristics. You move to the cloud, you start to see people do a lot more with automation, with policy, and so that management model is now much more automatic. So if a spike comes in because of a Black Friday example, how do I then scale up all those resources? I don't want you to send me an email and have it happen like two days later. I want it done now, and I want you also to then take it back down again when everything comes back down. And that's not the way the traditional middleware has been developed. So there's a deep engineering problem there that gets solved as we move forward. So you can still use the core of the capability and so our customers get kind of the flexibility of understanding what the middleware was in both environments, but the management model around it tends to move towards that accelerated, automatic, policy-based, policy-driven kind of model. So you built in a lot of automation, elasticity, ground scaling, and that's all programmatic now as opposed to you sending me an alert. Right, and there's an operator there doing something with it. So the set of API, the entries and exits sort of evolved, did you talk about that? Yeah, so pretty much anything in the middleware space these days is instrumented with APIs, with usually analytics as well. That's another characteristic we see as you move to the cloud, right? We can retrofit that in the systems of record, but in the systems of engagement, that's an expectation that you're going to have. Heavy API, heavy analytics support so that you can in fact drive everything to that end goal of I can be hands-off, I can just operate the way it needs to operate when I get these spikes. It's just a different model and people are starting to get their head around that and we have to make sure the middleware actually can make that model work. So you're talking with PointSource, one of the customers, XIVM'ers, so it's going to get the perspective, but they're doing very well. They're a big shop, they do a lot of mobile first stuff. Talk about the issues around cloud services that you're involved in, you're in the middle between old, traditional, and stable software. Right, and we don't like old so much, stable software, especially with my gray hair, I don't like old, old, old, old, old. But mobile really comes down to- Proven software, let's go with proven software. Trusted in proven, so. Experienced, reliable, proven software. Now, the unreliable cutting edge cloud, okay? I'm killing myself, I'm digging a big hole. I'm going to stop, my hole's getting bigger. Okay, APIs, APIs, notification, security issues, I mean they're everywhere, right? So this is the hardened issue, right? So the old stuff is hardened and reliable, so you've got to move to a new world, speed, agile. What's the bridge, what's the connection there? So when we start talking about mobile apps in particular, we start talking about composition as opposed to just raw creation, right? So you don't just take out a blank piece of paper and start building something these days. And in fact, you don't want to because each of those components you have to build yourself then have to have all of these availability characteristics, scalability characteristics, and that's hard stuff. So if you can, you want to sort of pick it off of a palette of, hey, here's a data service, here's a notification service, and behind that are built in all the qualities of service. So in the new world, that's what those things end up doing. In the existing system of the record, we use integration as kind of that buffer. So you don't want to expose your end point from the system of the record. It's a thoughtful process. You got to think that through. You got to think that through and you want flow control, right? So most of these systems of record were not built for an environment where it was heavily volatile, mobile-based kind of applications. So you've got to protect them from coming and getting too much load on them, and that's another kind of function of the integration and the API piece. So how do I flow and do the flows and limit when necessary so that I don't flood the backend? You bring up a good point. We had Inhi Chusan on earlier today and we had talked about the Hadoop market. She says it should be growing much faster. Just to tie this off with your point is that, the open data platform has been, there's a lot of controversy around it. The purists are saying, why do we need this? What's the benefit to customers? We don't need it. Some EMC, you guys saying, how has it hurt customers? And the point is, what you just talked about the integration is not something that a pure play company like Cloudera thinks about because they only have one product, right? And their customers aren't as comprehensive as, say, IBM. Yeah, it's easy if you can just sort of clean the slate and not worry about anything else that's ever gone in the past, right? They're monolithic in their solution, open source in their approach. Okay, I buy the open source, but like you have to go to Mark and deal with integration, security, legacy. And we do a lot with open source. So open source isn't bad and we use it extensively. It's awesome. Most of what we do, especially in the systems of engagement space, is based in some form on open source. But to your point. IBM's betting on open source. Correct. Like everybody else. But nothing gets done that's interesting in the mobile space and even the analytics space without incorporating some of that system of record data and services. So it's really the connection back. That's table stakes. The integration with system of record legacy, that's called legacy kind of blanket it over. That's legit stuff. That's important. You call it running your business on data. Think about your banking app, your airline reservation app. It's not very interesting if all it's doing is showing you nothing from the backend system of record. That's what everybody wants. Because you brought up data swamp earlier in our little data ocean kind of riffing around. It's kind of having some fun with that data lake versus data ocean, which is a fun conversation. Because it's provocative. It brings up some new things. You mentioned data swamp. What is a data swamp in your mind? And how do people, how does that happen? I mean, it's just like, I planned for a lake, got a swamp or ocean turned. I'm like, what does that mean? And what does it come from? And what's some of the triggers for that? Yeah. So I mean, I think it comes down to management and flexibility, right? So at the end of the day, what everybody's looking for out of the data lake kind of concept is, how do I get everything into one place so that there isn't 15 copies of things going when I don't need to, that I can't find things because they're not connected together. So that's the good side. The bad side of course is when the management system behind that breaks down, then you can't find stuff or stuff gets lost or stuff isn't optimized for where it needs to be. So this notion of tiering as well. As you put everything into one big pool, you don't want it all on the most expensive storage media if you can avoid it. You'd like to be able to sort of scale stuff back, maybe all the way back to tape to be able to archive and store and get the right price performance for what you need. And so if you're doing the data lake concept, in my opinion, correctly, then you really need that robust management system to kind of move data around where it needs to be. Yeah, and that's not an easy task. If the bigger you are, unlike the mainframe business wheeler yesterday, the more workloads it takes, the better it gets. So data is kind of like the opposite, right? The more data you come in and you're not paying attention. Correct. And you get swamped with, you know. So it becomes a policy-based thing as well. We were talking about using automation and policies. A lot of these, you know, tiering technology and the management practices have to be the same way, right? I have to have policies for how I do that and how I interact with that data as it grows. And it turns out we have some really good, deep technologies there in terms of, you know, our storage portfolio that allows us to do that. Okay, so that's not sort of middleware that's. Not middleware, but it is systems. Reusable. It is systems, right? It is systems, yeah. And there's a cloud, back to my own role, right? I mean, cloud storage is a big deal and being able to, in fact, you know, integrate between existing, you know, storage, you know, things you have in an on-prem or even a system of record with some of these new object and file-based storage, you know, media they might have in the cloud. It's a big problem, right? How should customers think about this? I talked earlier about the, I'm a simple guy. So the infrastructure service, platform was a service, software was a service. How should customers think about consuming that? Should they think about consuming a solution, a set of services? Should I think about, you know, feature sets within each of those layers? How is IBM going to market? So I think it comes down to, from a personal perspective, the amount of control you think you need to exert, you know, on the system, right? So if you want the hands-off approach, then the software as a service model is kind of the, you know, here's a subscription, you know, on board very quickly and, you know, we manage everything for you or your provider does. As you go down to the next level with platform, you get more control, right? Now I can talk about what kind of middleware I want to deploy, maybe, you know, use policies and automation to scale it out, but I still have control. I may be even building applications in that particular model myself. The infrastructure layer, of course, now, you take on all of the software, you know, parts of that. And so if you have, you know, a very specific need to control everything about that software stack, then, you know, the infrastructure approach might be the right one for you. So it really depends on the workload, you know, especially when you get into specific workloads with performance characteristics, some of what we can do with software, especially with bare metal, gives us options there in terms of things that you wouldn't normally consider as viable cloud workloads suddenly become possible because you don't have to deal with, you know, the overhead and penalty of things like virtualization. So you can do some very deep optimizations, but, you know, you can also have, you know, the hands-off approach, right? So it's really a question of which level you'd like to get into. So I'm curious as done as to what the customer conversations are like, and I'm sure they're in all three of those categories, but let's take a big back, you know, the sweet spot of IDMs, you know, customer base, they're tech-savvy, they're oftentimes very much leading edge, they're cloud paranoid for certain core applications. What's the conversation like with that class of customer banks, you know, insurance companies, you know, whatever? So I mean, if you go through the three layers here, right, the software as a service one, we tend to see those as what would have been packaged applications in the past, you know, a function within the enterprise that they can see outsourcing to some degree because it's not core to what they do from a competency perspective. That's why you see a lot with CRM in some cases with, you know, things like, you know, talent management, HR type functions, because those are things that you can imagine putting out, you know, into a SaaS type environment. As you go into the platform thing, they're looking for speed. So even though, you know, banks in particular tend to be a little more traditional, maybe a little more conservative, they do realize that when you go to this mobile model, you can't develop in a six and 12 month release cycle. It's got to be, you know, days and weeks. And so they recognize the need for things like a Bluemix environment, for example, for that rapid integration. As you get into, you know, the infrastructure layer, I hear more conversation there about, you know, setting up private clouds. How can I better manage my on-prem, you know, data center and the facilities there? So things like pure application system with patterns gives me that, you know, automation, better cycle time, better maintenance, you know, profile over the course of it. But I fundamentally haven't changed the middleware in that particular case. It's still a system of record, it always will be, but I can get a lot more out of the infrastructure by moving to a cloud model there. How much integration will occur over time between those three layers? Well, I think integration is key across all three. I mean, in fact, you know, we assume that everybody's going to be a hybrid cloud of some sort. How you decide to do it is going to vary by customer. But, you know, take that example of I outsource something to a SaaS property. I onboard with Bluemix from my platform as a service and let's say I use PureApp inside my enterprise to manage that. All three of those environments still need to talk to each other. My SaaS app isn't going to be an island. So if I outsource HR, the rest of my ERP systems are still going to be back here. So how do I integrate there? So you're, again, at the heart of that integration is really your middleware on a huge portfolio. So this doesn't happen overnight. So on a scale of one to 10, how integrated is that massive portfolio today? Actually, it's pretty well integrated. We've done a lot of work over the years, even within the middleware group to truly come up with those connections and those integration points. SOA did a lot for us in that regard in terms of being able to do a service layer and service approach around this stuff. We hear a lot of APIs now. API is kind of the new buzzword. It's really a form of integration to some degree. And it's exposing those assets from maybe a traditional system or SOA environment out into the internet or to mobile apps or to other clouds. So the terminology changes a little bit and some of the protocols do too. I mean, it's a little simpler now with some of what people do with APIs, usually using technologies like REST, for example. But fundamentally, it's still about service orientation. It's still about exposing these things in kind of a controlled way. And that gives us a way to bridge between these different ways. And how about the SAS portfolio from SAS to SAS being able to exchange data? Is that something that you guys are... It's a big deal. So we have a technology called Cast Iron that does exactly that. And it's for, we call it application to application integration. Your example is perfectly valid, but the more common example, quite frankly, is that I'm using, say, SAP as my enterprise application. And I decided to start using Salesforce for one function. How do I then connect and sync that data between those two? Mixing and matching. Classic mixing and matching. But as you go forward, we assume more people will have more SAS properties and then you have the SAS problem and then the SAS test on-prem. That's more the norm, having mix and matches that take on a long tail distribution, big, heavy, committed to niche-y apps. And you've got, it comes down to, if it's something that you don't think is a core competency for your business to manage, maintain, et cetera, then presumably somebody whose business is solely to do that is going to do it better, right? And if you can get your head around that model and deal with, you know, typically it comes with a little less customization. So, you know, as you go to the SAS model, it tends to be a little more simplified. But as long as you can work within those boundaries, then it tends to be a really good option in terms of managing those things. So what's your plans for this year? Talk about your organization, your chart, your mission. I mean, you're in an interesting spot. You're like dealmaker internally, you're brokering services, you're at your hands in a lot of pies, you're cooking many kitchens. Is that true? I mean, as that would be. It does feel that way. I mean, you're a glule. I mean, there's a lot of action going on around you, right? So, true. So, you know, from my perspective, the name of the game here is going to need to make sure that, you know, the newer cloud, you know, applications and workloads, that we can capture those for IBM, whether that's capturing them with environments like BlueMix, but also with technologies we can bring from middleware. So, you know, our classic run times like Java, but we're also doing a lot with things like Node, for example, to be able to move into new spaces for new applications. So not just the existing system of record, but the new stuff, they should do that too. So I think that's a piece of it and making sure that's consistent so that a customer who engages with us here can also, you know, see, you know, some of those core elements manifest themselves. That's a sweet spot for IBM. I mean, I think it's not really talked about, it's hard to market that on the top line, mainstream, but like your skills with integration or services is really strong. IBM's a company. That's like a DNA. It's not that obvious to the people who don't understand the enterprise, but that's your core sweet spot, right? I'm going to integrate old and new or proven with cutting edge, right? That's kind of your whole purpose. What are you going to knock down this year in your business plan? What's your objective? Well, you know, from my perspective, it's all about promoting and getting the word out about APIs. So, you know, it's starting to happen. It's still early stages. How about APIs? APIs and the API economy and API management. So, you know, it's really about using APIs as kind of the gateway, if you will, from the enterprise out to these other clouds. And, you know, creating the API is a good thing, but it's frankly the easy part. It's how do you manage that API over time? How do you deal with, you know, spikes in workload? How do you deal with versioning? It's almost like electricity before grids. Things happen in APIs. They just, like, just not show up. Sometimes something bumps them off. They go off. They go on. And the JSON endpoints or whatever API it is. Right. And so we tend to use manage, analytics to manage that and look at what's going on and give insight into, you know, how is my API behaving? You know, maybe you've got one person who's, you know, really abusing that API. How do I filter that person out? You know, these are the things that you have to start thinking about because most of these APIs, when they get exposed, they don't have a predictable usage pattern, right? So, you've got, you know, mobile apps coming in. They could, you know, generate a spike in traffic at any time. The old way of capacity planning just doesn't work in kind of- Well, you're mixing and matching. So, you're mixing and matching. You get diversity, especially foreign apps, not necessarily in IBM, others. Right. APIs are kind of a nice little interconnection. I get that. What is getting the word out me? Like, community, open source, wiggle system? I think all the above, right? So, you know, first is making sure that our clients get the message that this is, you know, a way to get this done. That this integration space in cloud is, you know, all about connecting things in APIs. Reliable and safe. Right. We expect to do things with open source. You know, we're starting to use things like Swagger, if you're familiar with that, in the API space, a definition. It allows people to kind of, you know, coherent way talk about how this API behaves, you know, the parameters to it, the documentation of it. And, you know, do that in a way that's kind of reusable over a course of different things. So, we expect to do a bunch of things in that space. But, you know, the big deal right now is to make sure people are aware that they need this. A new way. People don't necessarily know that this is a problem. The DevOps guys are all over it because they're born in the cloud, they're playing with APIs. But that's the norm now. What's more than it will be? Going more API mix and match. Correct, correct. Okay, well, appreciate you coming on theCUBE. Anything else you'd like to share at the crowd? Give them a taste of the vibe. That's our question of the day is, what's the vibe of the show right now? I mean, it's new. Is three shows in one, is it working from your standpoint? I think it is, you know. I think it's been great to see, you know, I've gotten a lot of feedback on, you know, just the notion of cloud and mobile and how that really is reshaping and re, you know, causing people to rethink how they really want to, you know, deal with not just their new systems, but also their existing systems, right? And so most of my conversations this week and the vibe I get is, people get it now. You know, this is different. This is something I need to do. We're not, you know, we're not in a conversation now where we have to convince anybody that, you know, this is a big deal. Everybody comes to the table fully acknowledging that and now they want advice on, you know, how do I make that happen, right? What are the best ways to do that? Don't believe the VP of cloud services of important role, the glue, the integration that's a big, you know, unspoken differentiator and important. If the API economy is here, notifications, APIs, mobile, it's all evolving in the cloud. Cloud is the engine of innovation. This is theCUBE. We sent all of our content to cloud. It's on siliconangle.tv, it's on interconnect go. That's an API driven site, by the way. We rolled out a new, have you seen the interconnect go site? All APIs, you know, there's a lot of APIs in there, all coming in from the crowd data. Go to interconnectgo.com, you want to check out all the keynotes on the main tent, theCUBE developer action and see our APIs in action. See our APIs in action all from the crowd, trending hashtags, trending stories, see what's the conversation, join the conversation and of course get all the access information from IBM, log in, require. This is theCUBE, we'll be right back after this short break.