SAP HANA can help you dramatically accelerate analytics, business processes, predictive analysis, and sentiment data processing – all on a single in-memory computing platform. In-memory computing is attaining a lot of hype as a new solution for organizations looking to become more agile, and SAP has been banking on their product HANA to become the new foundation of this movement. But the practical CIO will have reason to question the speed and the economics of SAP’s suite of tools in comparison to other platforms, as well as the entire in-memory database on which it is built. We ask SAP’s CTO if HANA can really hold up to such scrutiny.
- Dr. Chris Nokkentved, CTO, HP Enterprise Application Services
The misunderstood beauty of In-Memory Computing is that it really does allow you to do things that were previously impossible with disc-based databases. This is most true in SAP’s HANA, a platform that touts itself as much more than simply an accelerator. SAP is betting the farm on HANA, and they’ve developed it from the ground up to not only be “infinitely scalable,” but to be able to handle existing and innovative new applications. But SAP’s customers are currently struggling to determine just where a use case for this computing paradigm might lie. While some are speeding up existing transactional issues exponentially, others are using HANA to chart new territory, white-boarding ideas and having IT map it for them in a matter of days and adjusting figures in a matter of seconds. The truth is that HANA and In-Memory Computing are not panaceas, nor are they made to run on everything for everyone. Certain applications within certain industries are legacy systems that simply cannot be transferred so simply, but SAP’s supporting partner HP says that you have to look at the cloud platform underneath the application to see the full perspective as to how it might impact the business.
Yet in terms of accomplishing things that could not be accomplished before, HP and SAP challenge their customers and those looking to go the route of In-Memory to start thinking about their IT architecture as an enabler for business capabilities and as a driver of business value. The scaling up process could be fast or it could be major in terms of introducing entirely new tasks and skills to staff, being careful to avoid underestimating the large amount of customization that exists within their current application.
What HANA really allows you to do is go through simulations and what if scenarios when crunching numbers in real time that otherwise would take hours before you could get your job done and physically cannot be completed with a disc-based model because there simply is not a server big enough in the world. SAP hopes that HANA and In Memory really change the way a company looks at their business processes and the type of value they can now bring to their customers. In that light, In-Memory Computing has become a paradigm that can no longer be ignored.
Sanjog Aul: Welcome listeners, this is Sanjog Aul your host, and the topic for today’s conversation is “Can SAP HANA Hold Up To Scrutiny?” And I have with me Dr. Jeffrey Word, who’s the Vice President of Product Strategy for Databases and Technology with SAP. And we have Dr. Chris Nøkkentved, who’s the Chief Technology Officer of Enterprise Application services with HP. So, In-Memory Computing is finally breaking into the mainstream with all of the major players taking a stake in its evolution. With a myriad of solutions being offered in the market place today, the question now is where can the IT practitioner start? And we’ve invited Chris and Jeff here today to help explain how IT leaders can identify what the real value of an In-Memory solution like SAP HANA as implemented by HP could be for their organization. So Jeff, starting with you, what makes SAP HANA a viable solution and what are some of HANA’s core differentiators?
Jeffrey Word: What we see today in the In-Memory is really what SAP has been pioneering since the mid 2000s is deriving out of the best practices of what we’ve learned from operating the world’s most mission-critical complex applications the over the last 40 years to design a new architectural layer for the database industry. Coming from SAP’s perspective of an applications company, we really decided to build the next generation data layer for applications to not only power innovation but to overcome a lot of the deficiencies of the previous architecture of disc-based databases. And because of that, now you see all of the other database vendors kind of following SAP’s lead into this new architecture for In-Memory databases. And the really critical thing about SAP is it was built from this ground up from the first line of code to be a disruptive new innovation in the database world.
“HANA was really designed from the very first line of code to be a mission-critical, infinitely scalable, bet your company on it database to run not only your existing enterprise applications but all kinds of innovative new applications that you can build now in this world of no constraint.”
This isn’t something that was warmed over or bolted on as an accelerator; this really is a natively developed In-Memory database that does both transactional overlaps, oral TP and analytical both equally well and is designed to do both at massively, massively scaled volumes. So that really is the biggest differentiator for what people see in SAP HANA compared to the other In-Memory architectures out there. HANA was really designed from the very first line of code to be a mission critical, infinitely scalable, bet your company on it database to run not only your existing enterprise applications but all kinds of innovative new applications that you can build now in this world of no constraint.
Sanjog: So how can interested parties identify use cases where In-Memory will be valuable to them?
Jeff: Well that actually is the hardest thing about In-Memory technology; it is a little bit different under the hood than what people are familiar with in current databases. Really where we see most of our customers are struggling is, they know that this is going to be the next big architecture and they know that this is a journey they need to start to transition towards, but they often times have a difficult time figuring out where that initial value use case is going to be. And so what we see is customers are trying HANA in a lot of different ways. The most exciting for me at least are the opportunities to do things that have never been done before, like the really innovative ideas that companies have that are impossible to do in the architecture of a regular disc-based database. So that really is a really cool thing, where you take an idea from the white board to execution with this new technology to really drive a massive amount of business value and do things that have never been before.
“This isn’t something that was warmed over or bolted on as an accelerator; this really is a natively developed In-Memory database that does both transactional overlaps, oral TP and analytical both equally well and is designed to do both at massively, massively scaled volumes.”
But we also see people doing things where they’re speeding up existing transactional issues that they’ve got, where it might take four, five hours for a backs run and now they can do it in a couple of seconds, to traditional data warehousing types of use cases where things that take them a week from doing the transaction to actually seeing the reports, now they can do in a matter of seconds. Those types of more traditional use cases that are bound by the constraints of the disc-based architecture, and really HANA can come in and eliminate all of that latency in there. But like I said, the really amazing stuff is where people are starting to see this ability to take in innovative new idea and bring it out in just a week or two to a massive scale of users and do things that have never been done before.
Sanjog: So Chris, based on Jeff’s response, it looks like it can literally fix anything and everything. Is this truly a panacea?
Chris Nøkkentved: Well it’s not. I think In-Memory Computing has certainly a very important role to play in the evolution of enterprise application computing as we’re moving forward as it is not just applications that are affected but also the platform. But there are many other applications within a very large company, like manufacturing execution systems for example, all legacy applications or even specialized analytics that contain an enormous amount of business rules and complex interactions that cannot just directly be transferred over to the new platform. I’m looking at it very often from a heavily customized business processes area. Some processes are extremely stable; if is something really stable and it’s working, why necessarily move it if there is no business value and impact on it? So you have to look at the whole platform of HANA, not just the application, but also the cloud computing that comes underneath it through the HP offerings as a perspective of the impact to the business.
“You have to look at the whole platform of HANA, not just the application, but also the cloud computing that comes underneath it through the HP offerings as a perspective of the impact to the business.”
So if there’s no direct impact to business, why then move it to HANA? I have seen other use cases where we can see a resurgence of interesting response to supply networks for example. We can do things we couldn’t do before like sales and operations planning, demand signal management, much more advanced and fine grain responsive management. As well as in the customer and channel excellence area, we can do trade promotion effectiveness and sales pipeline analysis and predictive segmentation, stuff that were not even possible in the previous platforms being added by SAP on top of this platform, enabling it via the In-Memory construct.
Jeff: I completely agree. SAP is not saying or promoting that HANA is the solution for everything or that every application should run on HANA. It’s more of an issue of technically they all could, it’s speaks pure SQL and has all of the different platform capabilities you expect from a regular database for an application. So technically they could all run on that, but we as well promote the idea to our customers that the business value from this new technology as an enabler is really where you have to focus for these use cases. And some of them are about speed and execution, but there’s way more than that; it’s about really figuring out what that business value is, and how it’s going to transform your business, and if it’s not a really compelling case today, that’s fine.
There are a lot of other opportunities where you can get this into your architecture that are much more compelling and a much bigger business value. So it’s really an issue of not just can you do it but what should you do, and that’s really where we see people focusing their efforts right now as really identifying those initial use cases where it’s going to have maximum impact.
Sanjog: Is there a one size fit all solution available to broad stroke it across multiple industries and feel that life is good?
Chris: No, no there is no such thing, and I would even suggest that this is nearly impossible. I would imagine that SAP as we can see from portfolio developments are exceedingly focusing on industry, industry solutions and verticalized solutions.
So you will see even more applications align to some segments of an industry, for example the dairy industry versus the food industry within food. They have very different requirements and you will see more and more applications that were not possible before being developed for these particular sub-segments of a particular industry. So no it isn’t one size fits all and it certainly will be very diverse in its application as we’re moving forward. And people will also use the platform for new things we couldn’t do before, such as predictive segmentation or the ability to unite all data in a sales and operations planning exercise. This was literally not possible before to a standard application; now it’s possible because of the speed, of the ability of the platform to assemble an enormous amount of data and continues to give you the updated real time. So, no, no one size fit all.
Sanjog: How can organizations going this route assess their readiness to take this on and create the most value for themselves?
Jeff: Again it all starts with the business case, and the business case really defines the scope of what technology or what disruption is going to be imposed upon the architecture that they have today. So in many cases there is no preparation. We literally can come in and do a design thinking workshop with the customer on Monday and they can be alive on our system on Friday. So they can be stand alone, they can be running in the cloud et cetera, so those from a smaller use case perspective can be completely non-disruptive and there really isn’t much preparation. On the other side, this can involve a full blown migration of your entire enterprise architecture and we have a lot of companies that are going through that process right now. Thousands of people are going through trainings and getting skilled up to be much more valuable in this new world because they’re going to be performing slightly different tasks than they did before. DBAs are not going to be just regular guys doing backups and defrags and things like that. They’re going to become much more of data architects and really look at things holistically.
“Companies need to start thinking about their IT architecture as an enabler for business capabilities and driver of business value.”
So it really can go from very, very small to very, very, very big but the most important thing is these companies need to start thinking about their IT architecture as an enabler for business capabilities and driver of business value. And that really is across the board, small and big, something that everyone needs to look at, we often advise people that the best way to look at it is to think big but start small.
Sanjog: What due diligence do you advocate that an organization perform upfront to avoid common pitfalls and issues? Can you give us an inventory of some of the common pitfalls and issues an organization may face?
Chris: A very common pitfall is the customers underestimate the amount of customization that exists within their current application. So it’s not just migrating data from one database to another. There’s a lot of integrated functionality between applications that also needs to be moved over. And this is from what I have seen so far, one of the most underestimated efforts as we are moving applications over to the HANA platform. So this also brings me to the whole point around due diligence where we suggest to customers to conduct significant house cleaning for major applications.
It’s very important to re-document the current solutions brought from a technical landscape perspective, from a custom code perspective and also from a business process perspective. One of the major advantages of the new platform is that SAP is proving new functionalities and enables new business processes that can come in and add to your current pallet and inventory.
“A very common pitfall is the customers underestimate the amount of customization that exists within their current application. So it’s not just migrating data from one database to another. There’s a lot of integrated functionality between applications that also needs to be moved over. And this is from what I have seen so far, one of the most underestimated efforts as we are moving applications over to the HANA platform.”
So from that perspective, the re-documentation of the enterprise architecture is an important exercise, especially if you start involving solutions and system integrators into the effort. If you look at it from that angle, there is also misunderstanding where HANA really plays and where and how people should build a business case around HANA. For example should this business case be on agility or on speed? Jeff mentioned a couple of examples: we know from production planning and MRP that going from a six hour MRP to a six minute MRP means that this particular manufacturing company can actually create many more different types of components within the same planning period, given the setup of a manufacturing machinery. But literally, the business case behind the movement is also one of the very underestimated components of this transformation and a very significant pivotal part of the due diligence.
Sanjog: What are some success stories or case studies you can point to?
Jeff: Some of the ones that pop to mind, and we have thousands of these at this point live running on HANA, some of them fall under the category of doing the same thing faster, better and cheaper than they have before, like Chris mentioned, doing MRP instead of a six hour run it does six minutes. We actually have hundreds of people that have done 10,000 times faster reports and transactions. Now those are terribly transformational from a business on face value, but when you actually look under hood, it transforms fundamentally the way that the user interacts with the system. Instead of hitting enter at 5:00 that night as I leave the office and then getting the answer somewhere around noon the next day and then they can get their job done, if they can get an answer within a second or two, it turns the system into a conversation piece, and it also allows them to do simulations and what ifs, and if I change that variable, if I move this out, if I push that around, it allows you to really start to understand the impact when they are not waiting around for that system.
“We can now do a bunch of magical things that literally with data could never be done before and do them on the fly and then real time. And that really changes the way that the company starts to look at the business processes that they execute and the types of value they bring to their customers.”
So even though it might on the surface those use cases might seem like they’re just faster, it really does transform the way that you get value out of those applications. And on the other side you have a lot of them that are involved in doing things that literally cannot be done today. We have lots of customers that are doing stuff that it just physically can’t be done; there is not a system big enough on the planet. HP doesn’t make a server big enough that you could run these things because of the way it was architected in the old world with the disc-based database and just struggling to get data off that disk. And so now that we do have everything entirely in RAM all the time and we have all these engines inside HANA that can execute these incredibly complex algorithms on that data inside the database, we can now do a bunch of magical things that literally with data could never be done before and do them on the fly and then real time. And that really changes the way that the company starts to look at the business processes that they execute and the types of value they bring to their customers.
“Instead of hitting enter at 5:00 that night as I leave the office and then getting the answer somewhere around noon the next day and then they can get their job done, if they can get an answer within a second or two, it turns the system into a conversation piece, and it also allows them to do simulations and what ifs…It allows you to really start to understand the impact when they are not waiting around for that system.”
My favorite one in the world is actually MKI Industry or Mitsubishi Knowledge Industries in Tokyo. What they do is they take personal genome maps of people for cancer treatment and come up with a personalized cancer treatment for you based on your own personal genome and the specific type of cancer you’ve got along with the histories of what other people with that type of cancer have been treated by. And it used to take them several weeks and hundreds of thousands of dollars to generate this personalized plan.
And now by putting HANA in there and adopting that application to use HANA’s capabilities under the hood, they can actually do it less than 20 minutes while you sitting in the doctor’s office and do it for less than $1000. That is a totally transformational thing that they can now provide a service that they could never do in the past, and that really actually affects peoples’ quality of life pretty significantly when you have that ability. And so those type of things are the ones like I said that personally excite me the most, but it might not be a business value that is associated with it. It might be much more than that, like a quality of life value. So people really need to look at that and try and understand where those use cases are that this technology can add value upfront in incremental ways, but then also at the same time look for those really transformational opportunities that you can do things that have never been able to be done before because it was too impossible with the technology or too costly to try and attempt.
Sanjog: Chris, what are some of the case studies or success that you feel most proud of?
Chris: Yeah, I’m extremely proud of some of the work that we are doing on Volkswagen and HANA because VW and HANA are the most widespread users of the platform right now, and I can see that companies have started moving away from pure analytics and started thinking about using or achieving a business goal with the analytics. We have an example in China of a food company that improved their day sales outstanding five to seven times, and that had actually a significant impact on their stock price because they were actually able to report on this improvement. They also allowed them to have real visibility of their data and also make decisions in a more timely and rational manner.
I have another example from the US of a company that was able to use the COPA functionality, the profitability analysis for trade promotions, running in nine minutes instead of 28 hours, and that meant that they could make many more fine grain decisions in the way they should allocate their stock trade funds for maximum impacts with the products that they were launching into the market. So now instead of waiting throughout the weekend, they could actually launch in on Monday and still throughout the week make small adjustments as the shipments were actually arriving at the retail areas. And finally another one is a very large beverage company where we improved the data load times of the external data into the actual database seven to 10 times. So they were able to load point of sale data and make actual decisions on the data within the day.
“All these things sound a little mundane, but in the end they have a significant impact on the company’s relationship with its supply network and with its customers and retailers.”
All these things sound a little mundane, but in the end they have a significant impact on the company’s relationship with its supply network, with its downstream customers and retailers as well as the ability to understand how they can move and optimize the overall supply network around them.
Sanjog: The creation of new applications for innovative processes and/or business models is very crucial for any cutting edge organization. So how do you see SAP HANA helping to improve and accelerate the creation of these applications?
Jeff: Oh, that’s a great one, and that’s actually where we have some of most incredible use cases for these new things that literally can never be done. We have over 3,000 startups that have joined SAP HANA and based their algorithm and idea on HANA to do something that’s never been done before because now they have the capabilities of this new platform and that really is the difference is once companies understand the capabilities of these new platform and really kind of the paradigm shift in programming approach from the old world of doing things in a disc-based database, all of a sudden there is a flood of new ideas. And one of the great things that I’ve seen and I know that Chris sees this a lot as well is see customers or companies out there that have an idea, some business leader has an idea that he thinks or she thinks is going to be very transformation for the company, and they draw what they want it to look like on a white board, and the IT guys sit there and can translate that using the tools and the power of SAP HANA into a solution for them in just a couple of days. When they start to see that then and you watch their first reactions to that kind of evolution in just a few days from an idea on a white board to actual business value at their fingertips, that is a truly amazing and transformational thing. So as companies start to experiment with that, HANA is much more about just doing the things that you today faster, better and cheaper, it does that obviously extremely well. But it’s about this power of being able to extend those mundane things as Chris said, in to incredibly new, cool, sexy things because they were impossible to do in the old world.
“When they start to see that then and you watch their first reactions to that kind of evolution in just a few days from an idea on a white board to actual business value at their fingertips, that is a truly amazing and transformational thing.”
Chris: We are actually surrounded by Generation-Y and youngsters who are used to leveraging systems of engagement, mobility devices and smart phones, so their interaction with transactional systems are changing. They want to be much more flexible, and also while in the midst of a particular interaction with the transactional systems, start asking questions. How can I reduce customer product return rates? What was the biggest bottom line impact of that particular trade promotion? What happens when there are bills outstanding with this particular customer? Who is the best customer profile for loyalty reward?
So you see people starting to combine a multitude of requests and questions around a particular decision process of, for example, allocating an order and a discount to a particular customer. Now with HANA, as Jeff mentioned, there are a number of drivers that enable us to build, for example to decide this events team processor and the mobility front end, much more demand event driven applications that were not possible before. And there’s a lot of opportunity to build light weight applications that can live in all kinds of interfaces, on iPad, iPhones, Android and of course Microsoft setups. And the most important thing is the event driven nature of the new applications. Typically, before we had a transactional sequence of things, you had to complete the purchase acquisition, then you had to go to purchase order to group them, now we see many of things being able to be done based on events. So when purchase orders or acquisitions for example reach a certain level, then you can actually group them immediately, the system can do that in a way that it was not possible before. It was too heavy logistically and programmatically to build this kind of logic within the systems; now it’s possible. So I think the key focus here is the ability to build light weight, event driven applications that were not possible before.
“The key focus here is the ability to build light weight, event driven applications that were not possible before.”
Sanjog: What investments is HP making towards helping organization in getting the most value out of the SAP HANA platform?
Chris: That’s a really good question given that the HP and SAP has had a very long partnership and relationship and we have working collaboratively around the SAP HANA platform. We have done considerable investments in trying to integrate technologies on the hardware level to enable the platform to function at its optimum. We are continuing to make these investments to not only improve the platform, but also improve the opportunity to virtualize the platform and expand the capabilities.
“We have done considerable investments in trying to integrate technologies on the hardware level to enable the platform to function at its optimum. We are continuing to make these investments to not only improve the platform, but also improve the opportunity to virtualize the platform and expand the capabilities.”
We are very soon are going to announce the HANA Hawk Box, that will enable you to create a scale up infrastructure up to 24 terabytes, and it can then be scaled up to nearly an endless setup that enables us to build really large scale real time data repositories. On the software group side, HP has invested considerable money in enhancing the Cloud Service Automation Software that enable HANA to live in a private or virtual private cloud and enable applications to work at their optimum as well as deployed faster within a multitude of cloud constructs within HP’s Hybrid cloud construct.
The final point that I want to make is services. We are specializing a lot of our platform and application life cycle services for SAP HANA, and we have been collaborating with SAP on solution manager and a number of other platform specific applications like landscape virtualization management to enable us to move applications around in a multitude of constructs, for example, between a public cloud for testing and a private cloud for execution and rollout and deployment. We have built a considerable organization, a HANA COE that focuses on assessments and accelerates migration. And we are contributing already in to SAP’s rapid diplomacy solutions portfolio, by building a lot of RDSs ourselves and working also to build rapid deployment appliances. So there’s a broad pallet of investments across a number of areas from hardware, software, services that range from implementation acceleration services to operational acceleration.
Sanjog: What is your message to the enterprise IT leaders who wish to embark on this journey with In-Memory? What things they should do, what should they watch out for and what’s the future for them? What should they do to go about making the most value of this new computing paradigm?
Chris: HP views SAP HANA as a very central environment for the future of all SAP centric eco-systems out there and companies and enterprises. We view it as a single environment where you can combine and marry analytics, transactions and planning in a flexible and seamless interaction that will allow the organizations to move away from how the platform works, and rather focus on what are the business outcomes of the initiatives that they are driving with the business and what is the impact of these initiatives to the business.
Jeff: The key thing the people need to understand is this is not something that can be ignored at this point. We are way past the point of market acceptance on this, and what people need to understand is SAP is the world’s largest business application provider. We’ve bet the entire future of our company on HANA and that’s not hyperbole or over-statement in any way; that’s the absolute truth. All of the other data base vendors have added In-Memory capabilities to theirs and they are following our lead. So this is not something that is going to go away and it’s not something that the people can ignore and so I want to ask, all of the listeners, this is something you need to think about. It is a transformation, and it is a transition from an older disc based architecture to a new In-Memory based architecture.
“There are some pitfalls, it’s not a panacea, and what they need to do is really stop and think about the business value. And at each step along this transition, really identify what the business value is going to be and only do things that bring significant business value in each one of those steps.”
There are some pitfalls, it’s not a panacea, and what they need to do is really stop and think about the business value. And at each step along this transition, really identify what the business value is going to be and only do things that bring significant business value in each one of those steps. They need to plan out what that road map is, for getting on to this new architecture, bit by bit by bit. But as Chris has said and I repeat, this is a holistic approach. It is not a piece by piece thing in the long term. It’s a piece by piece thing as we progress towards that over the next few years, but SAP believes that most of our customers are going to have HANA as the core backbone of their enterprise in the next few years.
All of our applications have ported to it, and we are developing hundreds of new applications and innovations that are going to be powered by it. So, it’s something that needs to be thought about seriously and start planning on today. But I want to stress again that business value is the driver for this and so each step along that path, make sure that you clearly identify the business value, and you make sure you achieve that before moving on to the next step.