 Okay, we're back here live in Las Vegas for IBM, IOD. This is theCUBE, our flagship program. We go out to the events, district of Central Illinois. I'm John Furrier, the founder of SiliconANGLE. I'm joined by my co-host Dave Vellante. We'd love to get the experts and the authoritative figures on theCUBE, CEOs, entrepreneurs, but more importantly, we'd love to get the analysts on here because analysts will tell it how it is. Judith Hurwitz, president and CEO of Hurwitz & Associates. Welcome back to theCUBE, Cube alumni. Getting Dave, doing theCUBE with Dave is fun for me because we can just ping-pong back and forth and pass, shoot, score. But when we have another analyst on, when you've been on, when Ray Wang's been on, when Merv's been on, it's been really awesome. So thanks for coming on. My pleasure. Okay, so give us the scoop. What's the take? Give us the quick first inning overview of today for IBM. So it's interesting because what I'm seeing is these elements starting to be brought together, hardware, software. A lot of the issues around analytics, visualization, big data, predictive analytics, cognitive computing, pure data, all of them. I really like the blue announcement. The Neo? The columnar database announcement with blue. So I, you know, also bringing in, you know, analytics from a manageability standpoint. So you've got people, you know, out of the formatively organizations, CEL, whatever, CIA, no, not CIA. NSA? What is NSA? I forget what they're calling it now. No, talk about cognitive computing. What's going on there? What do you mean by that? So cognitive computing brings together big data, machine learning, predictive analytics. And it starts to take all these things together so you're not just looking at what question to ask, but you're looking at patterns and you're seeing sort of what's similar, what's associated to what, what's the context. And that starts to get very powerful as a solution in various industries. So that's gonna be sort of the next stage beyond big data. And I think that that's gonna be extremely important. It builds on it. Yeah, I was gonna ask you, is that a solution? And that's exactly what you're saying. It's bringing all these, because it feels like we got a lot of point products going on and a lot of little innovations popping up all over the place. Sounds like cognitive computing brings it all together. Yeah, cognitive computing does because it really then allows you to start looking at it from a knowledge base. So whether it's in, you know, healthcare or engineering or, you know, manufacturing, to start putting the pieces together to understand next-back action, what this means, what it's telling you about what you should be doing in the future, just for, you know, diagnosing illnesses because there's, you know, part of this is when you look at answers to questions like what's causing this disease, you bring your biases into it. It could be a disease that you've never seen or hasn't, you know, occurred in your country, but your patient may have been to some country that you have no idea what diseases are rampant there. So it really sort of brings in a combination of automation combined with, you know, knowledge coming from lots of different sources. So you guys wrote the book, The Big Data for Dummies book, and of course, things are moving so fast here, but when you wrote that book, it was starting to become clear that all the big players were gonna get involved, whether it was IBM, Oracle, HP, they've all got their big day EMC, they've all got their big data plays. What have you been tracking now? What's sort of changed since you wrote the book and what do you see going forward? So, you know, it's interesting because in some ways, a lot has changed, in some ways nothing has changed because, you know, when you look at something like Big Data, you can get, you can look at all the concepts, but how do you actually implement that? What does that actually mean to real customers? Because, you know, as industry analysts and people, you know, who watch this industry, you assume that, well, everybody's already down the maturity curve and it's not true. A lot of companies are either just getting started or, and what's happening a lot, the companies that are doing a lot of work in this, they're doing this in silos, they're doing in projects and business units. So now we actually have the issue, and this is one thing that I'm tracking, is now all of these silos, we've already had a lot of data silos, but now we have Big Data silos. And there's also the issue of when you go and grab all these sources and you use Hadoop and you go get all of these Big Data sources, they're not clean and you start to bring them in and you have one version of the truth. So you might have one department that's done a Big Data analysis who has now taken some line of business data and some of these data from third-party sources that bring them together. This is what they're doing their business plan on. You've got another unit that's taken a different approach. So now you've got multiple versions of the truth and is, they were talking this morning about veracity and that's actually going to, I think, rise as a much more important issue. So Big Data sources by their very nature are dirty. If it's social media data, some of the results could be that company planting a lot of information about isn't this service grade, we're so happy. How much of what's out there is somebody who's disgruntled and is just saying things that may not be true but they just, they're mad. They're just putting out data. So how much of that data is good? How much of it is junk? How much really is meaningful to what customers, let's say, are really saying about your products? So yes, you want to grab as much data as possible but that's why we talk a lot about going from big to small. It's the small data that's really important to you, not the, the big data is just your starting point. So you see an information quality problem growing up there. Oh yeah, absolutely. What's the answer? Is that a process issue? Is that a technology issue? Combination? Well, it's a combination. It's been able to sort of start off with a lot of data sources but understanding that they're not clean. And then culling from that, where are my patterns? Where are my anomalies? And when you get to that set of data, then you want to look at context. Is this, is there actually a correlation here or is it just something that, you know, I noticed that people who buy toilet paper also buy blue pens. They're, you know, it seems to show up. Is there a correlation? Is it making any sense? No, it just, it's just, it just happens. So you have to make sure that what you're looking at as an association is a correlation actually is meaningful or is it just sort of one of those things that happened? And that happens a lot. So in thinking about how information quality problems were solved in traditional BI and enterprise data warehouse worlds, I mean, ERP obviously helped a lot. But even with ERP, sometimes you had certain systems that are more up to date than other systems but at least there was knowledge of that discrepancy. It feels like with big data and particularly with social data, you know, you hear about info streams. I mean, this stuff is happening in real time. How are organizations going to solve that data quality problem when everything is so fuzzy? Do we have to change the mental model of what is quality? Well, no, no, I don't think you change the mental model. It's when you get to that subset of data that is meaningful either because there's a pattern there or something that doesn't match the pattern, then you have to apply data cleansing and tools that look at master data, that look at governance requirements. I think the other thing that's really important here is that when people are looking at huge amounts of data and they may be taking data that's been masked where there are regulations that say you can't show social security numbers. A lot of times when data is being moved into, for, you know, ingested for big data analysis, it unmasks the data. So all of a sudden this data in flight is now unmasks and in terms of governance, that's something that people have to pay attention to. The governance question came up, truth about iteration. You know, it sounds a lot like the agile startup, but you know, one of the customers on here said, hey, you know what? The idea of pulling back, getting it all together, reviewing it and then unleashing it is the old way and that data quality always creeps into it no matter how good you clean the data. So there needs to be a process of getting it out there and then managing the data kind of in real time. Well, managing the data in real time and making sure that that masking doesn't go away as data's in flight because it- So what do customers do and what's their choices? I mean, do they have any options? Is it simply a vendor issue? Is it an open source technology? What's that look like? I think it's a vendor issue. I think you have to look- Just pick the right vendor. Yeah, I think you have to look for vendors that are aware of that issue. You know, when you use open source, for example, you have to make sure that when you use that, you use it in combination with a vendor that understands the governance issues that you're required to follow. Is this a tempest in a teapot or is this going to be the way it's going to go? No, it's going to go this way. You know, you can't put the big data genie back in the box, you know? That's it. Data entry. Well, we heard the data genie. Although we want three wishes, though. Okay, number one, make the big data go away. We heard in New York last week from one of the CTOs. Once the data tap gets turned on, you can't turn it off. No, you can't. And what happens is then people want more data. And where that will actually lead us is to data visualization because, you know, the more data, you know, you start to collect, and then, you know, now it's not just this little pot of data that you want to look at. It's now, it's actually really big. And actually my first experience was this, was an infinity pharmaceutical in the Boston area. They actually, we did some work with them years ago, where what they were doing is they would take huge amounts of data about molecules that might have an affinity for each other. And they used the data visualization tool to see where these clusters of affinities were. If you didn't do that, there was no way that they could go look at everything. Oh, does this one like this one? So that's a way of using visualization to in effect control the fill rate of the data lake or data ocean. Right, right, exactly. So visualization I think will become the front end to big data. So, you know, everybody talks about putting data, big data analytics in the hands of business users beyond just the power analyst. Isn't visualization sort of the path to doing that? Yeah, I think visualization is, I also think that we will have to get to areas of abstraction. Because if, you know, I always, when people say, we're gonna need 300 million data scientists by the year, you know, 2019. If we actually do, this industry will have failed. Yeah. Really, because if we don't have layers of abstraction that abstracts the complexity, we will never have enough data scientists. So you'll know that the industry's really made progress in terms of maturation when you don't need so many data. Yeah, you're saying data science doesn't scale. It doesn't. Data scientists. Well, we know. You know, when the automobiles first came out, the biggest problem that people pointed to, will we ever have enough show first? Really? That was gonna be the limiting factor and how many automobiles would be bought. Another great soundbite. So again, but again, last week to bring the data point from BigWolf Data NYCR event we had last night, which Strata had an event kind of with us as well, same time, was this estimated about 20, 200,000 data scientists out in the global market, about 200,000, give or take a few here or there. And two million analysts. And that number is understated. So I think this notion of knowledge worker we were riffing earlier was, that was the Nirvana of the 90s, the knowledge worker. Now, you can shake your phone and get analytics. So the iPhone and the Edge device. And I think, you know, and there are definitely issues there, I knew somebody working for a company that is doing one of the social media campaign for scientific data. So they went to Google Analytics because these were science types. And they said, look, our Google Analytics score didn't go up, how come? You know, and it's, you know, this is not, you know, it's Google Analytics is not a science. It's sort of working. There's a lot of gamification going on. People are trying to game this. They are trying to game the system. And they just changed the algorithm then everyone who's gaming it, I get it. So how come, you know, if we, now that we have, you know, 14 tweets, how come our score didn't go up? You know, so it's sort of a funny world right now. Yeah, it's kind of like market share data. Right, exactly. That's never game, does it? No, never. Never, I didn't think so. How about this idea of a chief data officer? That's come up a bunch of times today. We heard it at the MIT conference. Do you see that as a role that is going to emerge? And will that role be reporting into IT? Will it be distinct from IT? What's your take? I actually think there will be part of IT. But it'll be more of a committee between the, because, you know, it's an IT issue in terms of data cleanliness, getting the data sources in, manipulating them, abstracting them the security. But it's not just an IT issue. It really is a combination of what IT does and what data the business needs. And sometimes I won't even know what they need. You know, if you know exactly what question to ask, you probably don't need big data. Because, you know, your universe is small enough to know exactly where the data is, you know exactly what to ask. But if you're in an ever-changing market or a very competitive market, you know that there are a lot of sources and it's figuring out what new sources you need to bring in. So the idea of the chief data officer, I think it's too, it's just a pipe dream. So how do you say it's a pipe dream? Too early or it's inferred, it's inferred? Requirements? It's not just one person. What does a chief data officer do? Let me go see if I can find some more data for you. You know, no, you can't do that with the data. What's their job? Kind of like hiring social media consultants in 2009. Well, they don't know what to do because it hasn't been done yet. In a regulated industry, like financial services and government and health care, the job is to cover your butt. But then, you know, that's the, you know, chief security officers is not, isn't that their role? So how is that done? Only in part, right? Because there's data quality issues, there's data governance, information management, certainly security, data science. You could expand that role, that's a. But again, is that one, is that one? Well I think what she's saying is the use cases are not defined, so you're essentially, it's a moving train. And what you're saying is that you can limit the scope, like a chief security officer say, your job is to protect the assets. Yeah, but how can you isolate that? Okay, I'm only concerned with the security of the data. It's a slippery slope. I'm concerned with the, making sure I have the right business data. Some, so, you know, maybe it's a team of people. I think it's too big. It's depending on the industry and the individual. Dave Laverde talked about today, who owns the data? Is it IT or the business unit? It's a squabbling mess. He said nobody owns the data. Everybody owns the data. No, I mean, and it becomes a political issue. I mean, you know, if you look at, you know, the life sizes industry, it's been political for decades. Wait a minute, that data's mine. Oh, but I'm trying to, no, you can't have it. It's my data. It's from my clinical trial, you can't have it. Yeah, so you'll never get to a single version of the truth without some kind of standard. Well, you'll never get to a single version of the truth anyway. That's true. Yeah, I mean, we're further away than we've ever been from a single version of the truth. Yeah, but nonetheless. And it's going to get worse before it gets better because we have so many silos. How about privacy? I want to ask you about privacy, because particularly in financial services, you hear about the three big use cases are risk, which really credit risk, marketing, let's sell some stuff and fraud, you know, fraud detection. And particularly, you know, the credit risk is interesting because you now confer from data race, religion, sex, obviously. I mean, so many different things that you're not supposed to use to make a decision about credit worthiness and there's some law that says you can't do that. Is this going to be self-policing? Will organizations actually be trusted to do that? I mean. You know, I think it's the Wild West right now. And I don't think we've solved that yet. And I think, you know, I think people will go, I mean, look at Facebook with some of the issues they've had. So some of it will be the marketplace and the consumers pushing back. So I think it'll sort itself out, but right now it's going to be the Wild West because you can't, you know, you can figure out just about everything. So what's some of the stuff that Hurwitz is working on, some of the cool things that you guys are doing? Well, we are doing a new study on predictive analytics. We came out with our Victory Index predictive analytics two years ago and we're updating it now. So that's going to be exciting to see what customers are, we look at it from a variety of perspectives, what customers are saying, you know, how these vendors are presenting themselves to the customers, what the value they're offering and, you know, what's their vision for the future. So we'll be redoing that. We're doing a smaller study, you know, quick study on maturity in big data. Looking at it, we actually just came out with a new paper about securing the cloud that we just did and, you know, sponsored by IBM. We talked to a bunch of customers about how they are using, you know, security in context with really customer facing cloud services. What do you make of IBM's, you know, cloud play? They obviously made a big move with soft layer. Yeah, I think soft layers is really, we'll put them in a very strong position. It's a very strong offering because it allows you both to use multi-tenancy and to be on bare metal. So it's a very strong platform. Yeah, management, bare metal, but now, so, and it was a big hole that they had to... It was a big hole. But some have said, including John Furrier, wow, it's really just hosting. I was talking about soft layer earlier. I'm not big on soft layer. So you're not sold on soft layer? No, not at all. I'm pretty positive. Judith's pretty positive. Why are you not sold on it? Yeah, I'm positive on it. I mean... I'm not negative negative in it, but it just doesn't feel $900 million for essentially a hosting company. I don't have all the data, so I'm just cheap at the moon here, but... You have to look at what do you mean by hosting? A cloud solution is hosting. I mean, depending on the automation. I trust you guys will talk about it more. Well, it's self-service, because it can be on-demand pricing. But if you look at those cloud-like attributes, that's what I look at. It smells like the cloud. It quacks like a crowd out. It does quack like a cloud duck. You'll agree with that, right? See, I didn't realize that quacking was a technical term in the cloud. Yeah, but I'll take your word. We're going deep here. But it doesn't smell like software-defined data center, okay? You think about virtualization. You're talking about commoditization, horizontally scalable, integrated stacks, like we see the success with Amazon's happening. Do you really think there's a software-defined data center right now? No, but we do a lot of research looking at software-defined networks. I think Amazon is. Looking at virtualization. Do you think Amazon is not? No, I don't. I don't. I think that basically it's a bunch of instances. They do have a certain amount of manageability, but I'm not sure that if you think about the way a lot of people use Amazon services is because it's cheap, and I say that in quotes, depending on- It's easy. It's easy. It's easy for a developer. So they're winning the developer market. And it's cool. They're losing the SLA market. They're losing developers a lot. No, they're winning the developer market. Well, have you ever read their SLA? Yeah, yeah. It has three words in it. No problem. Well, some they don't have SLAs on some of the high-end computing. Yeah, but you know what, though? I challenge anybody to go find a better SLA for a public cloud. I mean, they all stink. No, there is- On paper. Go ask Sarah Mark what their public SLA is. They can't even get it. No, the SLAs for these companies are written by a lawyer to protect that company. They have nothing to do with protecting- With the actual service level. With how the customer's protected. Right, it's an indication of how much risk the lawyers are willing to let the company take. So to me, it's largely meaningless. It's not an indication of what the actual service level is. It's not an SLA. I agree. It's legalese to protect them against anyone saying, I'm down, it's your fault. They're basically meaningless because they can't, you know- But right now- Expose themselves. Right now the conversation is just soft layer Amazon. Amazon, from a developer standpoint, has a lot of seamless capabilities, very scalable. A lot of the stack that they have is integrated in. But from a developer, software developer, it's a dream environment. From an IT role and get out as an enterprise, whole different ballgames. There's no purpose built hardware on Amazon, than I know of. It's, I'm calling it the iPhone of the data center. That is truly, Oracle will never be that. Although Larry Ellison claims he's the iPhone of the data center. I will maintain that Amazon is wonderful for picking up services, doing things fast. They have done a wonderful job on- Test dev, just like VMware was five years ago. Right, right. Do you not see it evolving beyond that? Amazon, I would not put it past them to evolve. But today I also think that if you want to deploy something really big, if you're offering services to your customers for a fee, then it's going to be very expensive. Oh, I agree with that. Rental from Amazon is always going to be more expensive. Right, and it looks cheap at the outset. The other thing that companies are contending with is that there are bits of Amazon all over their enterprises. And then how do you do governance on that? Because they don't even know where it is. And they don't know what's running on it. So Amazon would disagree with that. So I used to have that, I've said that before, and Amazon would say we have the transparency to be able to let you know. In fact, you know what they said? They said that before you deploy an S3 instance, you have to declare where you want it to go to reside. Yeah. Now, it can't reside in Germany. So in Germany, they allow them to reside in Ireland. That's not what I'm saying. I'm saying something different. So I am the CIO. Right. And there are 85 divisions in my company. I get it, yep. Yeah, it's a nightmare. And a bunch of them are, I'm doing a pilot project. Turns out that it's going to be something that's really strategic IP. It's out there. Yeah, and it's different. And I'm the CIO. And it's outside the scope of your governance criteria. Right. And it's not necessarily bad or good. It's just different. But then it comes inside my governance. And the CIOs don't know it's out there. But that's not Amazon. That's any cloud. Do you agree? It depends on the cloud and how it's managed. Now the difference is, The question is, Now help me with this. The difference is IBM would say, we'll write the SLA however you want us to write the SLA. We'll work with you hand in hand to change that. And I think, I think Amazon would say, here's our SLA. Take it or leave it. Yeah, that's the difference. And it depends what your use case is. If your use case is, I'm doing pilot projects. I'm trying out things. And it's fine because, and then when I'm putting it in production, I may move it inside because I'm going to actually, something like a State Street or a big corporation that doesn't have direct customers, but they provide services to companies who then provide it to their customers. Do you want to have that overhead of paying, in addition to your own infrastructure that you're going to support, then you're going to pay a fee to Amazon and then you- I agree with that. That's where Amazon's challenging. When you want to bring it back on premise, Amazon's not going to help you facilitate that in a way that maybe IBM would or another on-premise provider. Their bet is keep it in the cloud. I think that we're still early with the cloud. What do you think about IBM's social business manifesto and their positioning that they're coming out with? I mean, they've been down this row before. We love Lotus Notes, all that. They've had a lot of efforts going on, unified communications going back, many, many years. They're not new to the technical concept, but if you look at their overall strategy and what they're doing from a software perspective, it's not anymore these separate independent standalone products. It's really infusing social media into big data solutions, into commerce, into whatever the particular solution is. So I think it's a much better approach for them than trying to compete with the small social media companies. It makes a lot more sense. They seem to be bundling it in, but I mean, they have a great vision. I mean, IBM- I think the vision's good. They have a lot of experience internally. The whole unified communication thing just seemed like a legacy PBX market. They seemed like they were always incrementally bolting on to voice over IP. Yeah, and I think they had trouble with that. I do think connections has actually gotten some, some traction in the market. I'm starting to see, there are a bunch of companies in the collaboration space that are trying to be aggregators of different platforms, bringing them together, adding governance or whatever to it. And they're starting, so they're telling me that connections are starting to pop up as an important platform. Yeah, how about Cloud Brokerage Service? I don't know if you saw the service mesh. CSC bought service mesh. Right, right. Do you expect more of those types of deals? Is that a key acquisition? Are you mean from a broker standpoint? Yeah, I think so, but it'll probably be a few more years before that sort of the norm. But I think it has potential there. Have you been tracking the whole pivotal thing, what I call the misfit toys, that they ultimately get a great financial move and stick it under pivotal and get 100 million from GE and all of a sudden they get a billion dollar valuation? I mean, a brilliant financial move, great financial move. Largely Green Plum still, right? In terms of revenues. I mean, Green Plum has a good reputation, whether you can take Green Plum and all of the open source stuff, Cloud Foundry, and Gemfire, and Gemfire, and Sequelfire, and bring it all together and have something cohesive. I guess I'm still a little skeptical, maybe more than a little skeptical. Yeah, ways to go, one dot oh is not there yet. Okay, guys, great, we're on time here, we're getting the hook here. So this has been a great conversation. Final word to Judith, what should the folks expect to hear for the next day or two here at the IOD? What are you expecting IBM to talk about and what do you think that's gonna connect into into the marketplace? Well, I think that we're seeing a lot of it now. It's bringing a lot of threads together, not as many small, little products. We're announcing version 4.6.2 up. You're seeing a lot more solutions-focused, a lot more customer-centric, outcome-centric. So I like the fact that the messaging is much more around customer-facing, outcome-facing. So I think that you'll hear a lot more of that. Bringing together governance, manageability, predictive data, beginning of discussions around cognition, which we'll see a lot more of. Okay, the analysts break it down, the horses on the track, horses for courses, as Dave would say. This is theCUBE, love the analyst action. Of course, we'll always have our commentary and opinion here in theCUBE. We'll be right back with our next guest after this short break. This is theCUBE live from Las Vegas, IBM's exclusive coverage of IOD, hashtag IBM IOD. This is theCUBE. We'll be right back.