 Live from the Julia Morgan ballroom in San Francisco, extracting the signal from the noise, it's theCUBE, covering Structure 2015. Now your host, George Gilbert. This is George Gilbert. We are at the Iconic Structure 2015 conference, and we have a special guest with us today, Quinton Clark, who's the Chief Business Officer of SAP, which means he's the keeper of the kingdom in terms of the marriage of the future technologies that they are building and how that gets mapped to business problems. Quinton, welcome. Thank you so much. So SAP sort of pioneered the notion of, or brought the notion of in-memory databases into the sort of widespread consciousness. Tell us beyond the technology level, what difference that could make in the business applications that are more familiar to everyday people? Yeah, it's actually, it's true, you talk about pioneering in in-memory, in fact it really paved the way for other in-memory solutions to get some attention and some presence in the marketplace, but I would actually say the more important innovation of HANA was the re-marrying of the transactional of the analytics back into one system. If you kind of go back in database history 30, 35 years, they created data warehousing because the transactional systems just couldn't keep up, right? It was back in the day when transactions were being measured and single and double digit per second, right? Now of course we can TPCC millions of transactions in a second, right? And so that reunification, that's the big fundamental transformation. Yes, you had to get there through in-memory. Yes, you had to get there through building new data structures and new ways of encoding the information and new ways of getting the data in and out, but it is that unification that's so important. So tell us, was this unification, I mean there's the clear technical vision to say, this was a separation that was unnatural or it was natural at the time, but we can fix it now. But then what are the highest priority use cases where to bring them together we can do things we couldn't do for 30 years? Yeah, no, it's great, it's a great question. And in fact, Hannah was really born out of a question about how to re-envigor ERP. And as Hasso, a very famously Hasso Planner, our founder, was actually teaching a class, he asked the question, what would you do to really reinvent ERP? And one of the things he came up with was if you had a zero response time data system, it could really help a lot. And all of a sudden that became the focus and that became really interesting, you know, to build that data system. And so for the ERP side, things like in S4HANA, which is our new generation of ERP suite, we can do things like predictive closes and speculative closes on the financial books with the minutes and not have to wait until the financial period's over. And so for those of us who aren't as familiar with the CFO function, tell us what that means, how does that change their job? I'll just give you one sort of crisp example, right? Before S4HANA, the quarterly results, you know, the field finishes the quarter and then some period of weeks goes by until even the CFO understands what really happened and then you finally report the results, right? In today's world on S4HANA, the CFO and the head of the field organization, they collaborate on modeling different scenarios of how the quarter could close and they can do that in real time based on the real data and direct the field's energy and come back with results that are way beyond what was possible before. So it's almost like prescriptive collaboration. It's, okay, we'll iterate on a planning cycle and we think the best result is this way so we'll allocate our resources. That's exactly right. It's because the system is able to respond and perform and do prediction and planning in that fashion, right? It really is transformational for that part of the business. Okay, so, you know, our three wasn't, like Rome wasn't made in the day, yeah. So how should we expect to see S4HANA, the new generation, build out over time? Yeah, so, I mean, you're exactly right. Our three was not born a day. In fact, the transition from our two to our three, it literally took years. We're in a similar situation today and what's going to happen is our customers are first going to get onto HANA. They're going to get on the new modules as we are reducing the data set from 10, or from 100 terabytes to 10 terabytes all in one memory instance because of the power that HANA brings to the structures of ERP. Meaning you don't have to have an analytic database over here and a transactional one over here. That's part of it and the other part of it is the efficiencies are gained by being on HANA. No aggregates, no duplicate data. I can have one structure to support the entire application. Sort of like no in-memory kind of reports that are sticking around and materialized views. There's a lot of materialized views in today's and the old database implementations because that's the only way to make some of the screens perform. So ERP software, of course, supports all these different capabilities but having performance application experiences meant a lot of low level database monkeying around, I guess I would call it. And we don't need that with HANA anymore. So I'm going to potentially geek out here for a minute but R2 and R3 were stood for real time system 2 and real time system 3. But these sort of intermediate data sets really prevented it from being real time. So in addition to that collaborative closing of the quarter, can you give us some other examples where we really see a business function in real time? Of course. I mean, a good example this month we released this version of S4HANA that supports all our logistics, manufacturing, supply management, that functionality in addition to finance. And in that space, the supply optimization that can be done with the prediction capabilities there in HANA and that integrated hole between what has happened and what is currently going on has produced new features in supply management. Okay, okay, interesting. So now internet of things is one of those, I don't want to call it discontinuities but it's a big generational change. It is. But it has to do with, I mean there's orders of magnitude more volume. Like GE tells us they have 30 petabytes of sort of traditional enterprise data. They say that industrial internet's going to just swamp that. So from a commercial point of view, if you want to capture that in HANA in a new sort of configuration like stuff at the edges, pricing has to change, I imagine. But then you're also, you're going to need much beefier hardware to deal with that data and not slow it down. What happens, what do the apps look like and what do you have to do to accommodate it a lot? So you talk about that data and we're really talking about a very broad diversity of data. And one of the things that HANA has very aptly done over the last couple of years is move up the abstraction layer and have a query processor that is really able to unify different data from different data sets. And so we actually recently also introduced a technology called HANA Vora that plugs into the Spark framework that brings things like hierarchies and currency conversion, these kinds of capabilities into that Hadoop ecosystem now through Spark that integrates into HANA. So you can build a logical data model in HANA that operates over the traditional operational transactional side of data as well as all the new data sets that are coming in the IoT space, but into one model. And it's really not until the IoT data meets and mashes in and collides with the line of business data that real meaning is made out of that information. So when we go through generational shifts in applications, it usually means we can also enable new organizational or business models for SAP customers. Yeah, the applications customers. Has SAP thought much about, you know, like GE's predicts, you know, like we'll do field service, predictive maintenance for field service. Are there other examples that you're thinking about for, you know, SAP customers with S4HANA? Well those, actually if you go workload by workload, manufacturing, supply management, HR, all of these processes are going to become completely optimized by the ambient world of information around them. And so predictive maintenance actually is a great example, right, where in the manufacturing lines of course, having downtime can eat as much as 50% out of the profitability out of systems. And so combining all that industrial IoT data in with the line of business data allows that kind of solution. But it's true in upstream oil and gas where supply management flow and prediction of prices and supply needs becomes a way to optimize that business. It's true in human capital management, even doing analytics on needs for different kinds of labor and capability becomes another opportunity to optimize that's based on a lot of different sort of inputs. Does it, would it be fair to say if you're looking for what's common across these examples, your scope of analysis is greater. And, but it also sounds like that scope might go beyond the boundaries of one legal institution. It does. So how, I mean, how do you deal with, you know, I know there's a tech conference, but you know, if you want to do iterative planning with your supply chain, you know, like I too kind of promised, you know, 15 years ago and like, how would that work? Would they expose their data to their, you know, anchor customers? You know, the interesting thing is, is that most customers are willing to share their data when they believe they're getting a fair return for it. Right? And so we see cooperation amongst vendors in the supply management of the Reba or even in the travel and entertainment data, travel data in particular and concur, where the customers and the institutional entities are willing to share information when they're getting something back out of it. And so I think a big part where we're going in the business network space is helping set up those kinds of collaborative relationships where that information does end up being injected in to the overall application process. So this would be sort of like, we did the B2B exchanges, you know, and teenagers kind of set them up, you know, 15 years ago, but it sounds like we actually have the infrastructure to do that now. We don't perhaps know yet the full business model implications, but we have the technical underpinnings to coordinate across companies. And we have some incredible scale going for us as well, right? If you look at the sort of the network of providers and solutions that are available and concur, you look at the supply and consumer relationships in a Reba, or even if you look at Fieldglass, the contingent labor workforce, incredible scale of these networks now that are ripe for this kind of re-engagement on what other kind of value can come from those. So who operates those marketplaces? Will it be, you know, will it be a customer of yours operating sort of, let's say a concur marketplace, or will you be operating it and will the profits accrue to SAP? Well, I think it depends on specifically which area we're talking about, right? And I think some of those will certainly operate because it requires that there is a sort of third party, if you will, that sits in between these things to help sort of govern and run the processes. But I actually believe there's a generation of organic networks that's going to rise up, right? I've encountered some of these scenarios where even competitors want to cooperate in certain data sets to hold a supplier in check that supply all of them, right? And we needed a fostered environment where they're able to create those kinds of corporations even amongst themselves, but that always requiring a third party to be the governing body of that. Okay, slight digression. We know that traditional database pricing is kind of high relative to the volumes we have to capture now. Would SAP and HANA be accommodating the explosion in data by this logical tiering where like Spark might be the repository or the analytic engine and repository for sort of the IoT or fast data or big data and the transactional data would stay in HANA? It's true even in HANA today. So HANA at the query processor layer can speak to and include a Hadoop data whether it's through Spark or just raw Hadoop and it's overall data processing landscape. And so where you have the needs to store data in those super cheap forms, mostly data at rest, right? Data's not changing. And as well as even for the transactional, the operational data, system of record data, we've done the integration to support the IQ data structures in the HANA. So you can take data outboard out of HANA memory. This is a cyber technology, which is great technology. Outboard it out of in memory and put it on to a disc-friendly format. So I just want to touch on the competition a bit. Oracle seems to be a little more conflicted about how it wants to treat different tiers of data. You know, Oracle and NoSQL is having a tough time getting the attention of the Salesforce and the big data seems, you know, well, it takes the appliance with it. Is that changing, is that economic model or pricing model affecting how they're building their applications relative to how you're building yours? Well, I can't really speak to Oracle's design points, right? But what I do know is that, I mean, you made this point about real-time earlier, the HANA system, because it's married together, the transactional analytics side is allowing us to build our applications with new value and leveraging the in-memory data structures and leveraging the in-memory sort of revolution, if you will, to achieve that, the value's there and customers will, you know, they'll always pay for the great value. And so, it's certainly changing our architectural approach and as we build the S4HANA versions of things like financials and logistics, we're making the modifications, creating new value, and the value's really there. One last competition question. Have you heard about how Workday, you know, did their own database and it worked for HR but didn't scale for financials? Yeah, but it turns out, Davis technology's hard, right? And so, you know, I say the great news is that HANA, you know, was very much built first as a research and as an intellectual exercise to see what was possible and then it was shaped by a real application, right? In this case, the ERP application of SAP to be able to support a very broad range of workloads. Right. All right, Quentin, we have to leave it there. This is George Gilbert, Quentin Clark of SAP and we are at the Julia Morgan Ballroom in downtown San Francisco at the Iconic Structure 2015 Conference and we will be back in a few minutes.