Expand +



Detail and Insight Coming Together

Thorsten Franz on what SAP HANA and "the convergence of OLTP and OLAP” mean for ABAP developers and the business

by Thorsten Franz

January 13, 2014



Thorsten Franz, an SAP HANA Distinguished Engineer and SAP Mentor, took a moment between SAP TechEd sessions to speak to us about HANA, ABAP and founding his new venture, Operatics, which, he very publicly announced last fall, was betting his livelihood on the “convergence of OLTP and OLAP.”

For those who find this phrase as comprehensible as “the convergence  of QSTF and ABZD” (as Thorsten joked with me), Thorsten unpacked that phrase, with new implications for everyone: What DOES the merging of daily transactions, records and invoices with high-speed analytics mean for end users? What does it mean for developers and IT?

The full podcast is above (and the transcript below) – but here a couple quick snippets from Thorsten on why this “convergence” is so critical:

You have different sets of users … One team makes the makes the day to day decisions, speaks with the customers, and processes documents. And the other people – analysts, controllers  look at the web reporting and controlling tools and hopefully derive insights from the sum total and big picture of the data.

But very frequently there’s this wall between them. So the insights the controlling people derive from working with the data don’t go back into the operational business, and the people out in the field don’t make better decisions based on the data that is available.

This is the divide between transactional and analytical processing of data, and hopefully that’s going away right now. The convergence of these things is the process that is going on in the SAP world… 

Thorsten also shared advice on how this divide between transactions and analytics is now being bridged, the challenges of reporting that come with speed, where developers should be looking for their next new skills, and the implications of HANA for ABAP developers and IT generally:    

OLTP and OLAP coming together –  that means decision making, action, coming together with insight, and arriving at better actions through better insight. And having speed in your system, which you can achieve with ABAP on HANA.

But just speeding up transactions doesn’t get you the whole way. You need to change user interfaces, you need to bring analytical data - and any data that is ready for insights - into the UI where people are making operational decisions. … We need to learn to do the new things that are possible with HANA...

This spans every aspect of development and design. It starts at the moment you speak with end users … and you need to do that in a different way, so you don’t get the same old answers that you have been getting all along.  

For more from Thorsten on the big opportunities for ABAP developers right now and his advice for you,  take a listen to the full podcast (above). As a long-time ABAPer, developer, and architect, Thorsten sees this as a rare moment of opportunity for developers. Where will the "convergence of OLTP and OLAP" take your organization?

Transcript of the interview:

Kristine: This is Kristine Erickson, I’m at SAP TechED 2013 in Las Vegas, and I’m joined today by Thorsten Franz. Thorsten is an SAP mentor, the founder of Operatics, and a HANA distinguished engineer. He’s joining me today to talk about OLAP development, HANA, and his new venture. Thanks for joining me, Thorsten.

Thorsten: Well thanks for having me, Kristine.

Kristine: Back when you sort of announced your new venture, the foundation of this is the convergence of online transaction processing, and online analytic processing. So, OLTP and OLAP. Tell us, in business terms, what that really means.

Thorsten: So, traditionally, in IT, there’s been this huge divide between two very different kinds of things you can do with data. One of them is you can have the day-to-day operational business: where decisions are made, where invoices are accepted or rejected, where things are booked or not booked, and orders are placed or not placed, and so on. That is the OLTP, the transactional processing, and from a technical perspective, it means that usually, one database record at a time is being worked on, and the rest of the database records or in the entire system are not even looked at. You just look at one record in a sequence of millions of them. And the other side of the coin is the analytical processing, where the individual record is utterly insignificant and all you’re interested in are sum totals. You’re looking at the big picture there, and not the individual record. So you’ve got systems that are designed for one purpose and entirely different systems that are designed for the other purpose, and one system can only do this one thing, and the other system can only do this other thing.

But this divide goes further. You’ve got, originally the same data, coming from the same world, and the same real-world transactions, but you’ve got it in two separate systems and two separate formats, and two separate technical environments, you’ve got separate types of experts and if they want to make a change that impacts both sides, you need to talk to both teams, and they need to work at the interface where one system and the other system connect with each other. You’ve also very separate users and groups of users for those, and frequently they’re in different teams and even different locations so the one people make the day-to-day decisions, and speak with the customers and as I said, process documents, and the other people, analysts, controllers, look at the work with the reporting and controlling tools, and derive—hopefully—derive insights from the sum totals and the big picture of the data.

But very frequently there’s this wall between those, and the divide so the insights that the controlling people derive from work with the data doesn’t go back to the operational business. And the people out in the field don’t make better decisions based on the data that is available. So this is the divide that’s existing between the transactional and the analytical processing of data and hopefully, that’s going away right now. And the convergence of these things, this is the process that’s going on in the SAP world, that we’re tearing down those walls and bridging that gap at every level that I’ve just been talking about.

So eventually we’ll end up with not two siloes of data but having one set of data in one place that can be used for both purposes. And hopefully more and more by the same people, in the same place, in the moment of making the decision, what to do with that particular document you’re working on. And so, that it will lead to better business decisions being made on the spot, and insight being derived at the moment of conducting the operational business, and also getting rid of all those redundancies at the technical level. No longer having to have multiple systems, multiple administrators, multiple technical teams, that’s going to reduce the cost of ownership of those systems. This is why I think there’s a really compelling business case for making that divide go away.

Kristine: And then for, you know, the technical folks who are working on this, what difference will they see in this scenario?

Thorsten: First of all, there’s new tools, new programming languages, new IDs, new technical standards to get acquainted with. And it’s always a good idea to stay on top of developments and when something is obviously here to stay, it makes sense to learn it. And even in the early days when it’s not clear whether it’s here to stay, people can make that decision if they’re gonna take that risk of being among the early people who adopt that technology. And maybe have an edge, a competitive edge, being early adopters. This is how I decided to go about it, being working with SAP HANA and with the technology surrounding that very early on, because to me it made sense very early on.

And now I’m, personally I’m in the fortunate position that I have a longer track record working with these things than many other people, but it’s not too late to join that, and I think it’s a very good idea to look into those new skill sets and learn about them, but what I think is probably even more important is to look at all thought patterns for example, look at, you know, what is an analytical system really all about? It used to have a particular scope, it used to have a thing that it does do, and other things where you could rightfully say ‘We don’t do that.’ And that is no longer entirely valid, because users are coming to you with demands and if you say ‘We don’t do this’ because you historically don’t do this, you’re not giving them what they need.

So, experts in the field of development and analytics and so on, need to re-think the scopes and the classical boundaries of their particular fields as a merging of these areas is going on. And it’s going on in my experience, faster in the minds of users than it’s going on in the minds of technical experts. So I see more and more of users say, you know, I want to have an operational capability in this analytical application, or want to have an analytical dashboard, and pie chart diagrams and so on, this transactional user interface, and we all have to be able to deliver on that.

One thing where I’ve experienced people being stuck in a previous model, its one issue is real-time. When we look at analytical applications, when we look at sum totals as opposed to individual records, then we’re used to having well-aged data, so you used to have a rhythm that once or twice a month, or once per week, or in some cases once per day, you would get this truckload of data, and it would take half a day or so—

Kristine: Or in some cases, I’m hearing a week, you know, they started on Monday so they can get the report on Friday.

Thorsten: Yes. So, you expedite that from one system to another system, it’s a long haul, and then once that has arrived at the doors of the analytical system, it needs to be processed from level to level and that can take another few hours, and usually you don’t update that time slice frequently, but it’s frozen in the system. And there’s some good to that, because if you print out a report and send it to the CEO on Monday, and then the CEO gets to read it and becomes very excited or anxious about it, and asks his department heads to explain those figures, they better be able to get a printout of the same data and not some entirely different data so nobody knows what they’re talking about and when you have the capability to have real-time data this moment, while the operational system knows at this very moment, the printout that the department heads get is current but it will not be the same as the printout that the CEO is talking about. And people need to acknowledge that and think about whether that’s desirable. Then there’s the people who try to reconcile that somehow and say ‘Can we do both and be able to have the current data but also be able to look at historical data and snapshots and simply be aware of what we’re doing.’

Kristine: And also it sounds like it’s re-defining what historical data is, historical data might be a couple minutes ago, or an hour ago, not last week.

Thorsten: Yeah. Yes, that’s very true.

Kristine: What are you most excited about in your new venture, from a technology perspective?

Thorsten: I think we’re kind of entering a renaissance period in IT. It’s a time where old paradigms are really fundamentally re-thought, and where people go back to the drawing board and they’re ready and prepared to do that. But it’s like everybody is popping their head out of the water and looking around and seeing exciting things that they haven’t been able to see before and seeing new directions to go to and it’s really really great to be there at that time, and to be able to brainstorm with people and help them work out their new directions. This is really fun.

I love to brainstorm, I love to be confronted with challenge, conceptual challenge, and help people arrive at a satisfying solution and currently there’s so much opportunity to do this because people are really taking the time to say, ‘Let’s sit down and think about what’s possible and really re-draw the map’ and that happens only very rarely, in every domain in the world. And it’s always the people who are there when that happens are always very lucky people, in my book. And this is the time for IT and that’s the part that I personally find extremely exciting about it.

Kristine: Let’s talk about the fact that you were named a HANA Distinguished Engineer. That sounds like it plays into this idea of being at the right place at the right time in an exciting period, especially for SAP right now.

Thorsten: Yeah that was a great honor, that honor came very shortly after I had announced that I was starting my own business, betting my livelihood on the convergence of OLTP and OLAP, as I titled the blog post in which I told that story. If you look at the HANA Distinguished Engineers program from SAP’s perspective, SAP doesn’t select HANA Distinguished Engineers. There’s a panel of mostly independent people who select the HANA Distinguished Engineers, and the HANA Distinguished Engineers try to be realistic and while they’re subject matter experts, they will tell you whether you’ve got a nail and you need a hammer, or whether this is a screw and hammer might not be the right solution to this tool.

Kristine: Does OLTP come in together with OLAP? Is that the equivalent of ABAP working with ABAP on HANA?

Thorsten: Not quite. OLTP and OLAP coming together, that means decision-making, action, coming together with insight and arriving at better actions through better insight and through insights being available at the moment of action and having speed in your system, which you can achieve with ABAP on HANA, is part of that, but just speeding up transactions doesn’t get you there the whole way. There’s other things you need to take into consideration, you need to change user interfaces, you need to bring the analytical data and any data that is relevant for insights into the user experience, into the user interface where people are making operational decisions, so just knowing how to do the same things faster or how to do the same things on HANA is not enough.

We need to learn how to do the new things that become possible with HANA being around and that spends every aspect of the development and design process, it even, it starts at the moment where you speak with end users and try to get requirements out of them that help you build good software, you need to do that in a different way, you need to pose them different questions if you don’t want to get the same old answers that you’ve been getting all along. And you need to listen in a new way to them, with a fresh and open mind. You also need to learn new tools and get the best out of everything that’s available in order to be able to build those new things.

Technically, everything’s becoming a lot more heterogeneous at the moment, so you’re building solutions out of more different building blocks than you’re used to at the moment, if you want to create new types of solutions, and that means you’ve got to broaden your skills. So OLTP and OLAP coming together, action and insight coming together, the detail and the big picture coming together; and a favorite image of mine here is the whole world in a grain of sand—the big picture, the sum total of all the records in the database, on the one hand, and the tiniest detail on the other hand. Bring that into the same UI, into the same business process, and making both of those perspectives available to one user at the same moment so they can actually not only have them both on the screen but have that in their head and be able to intellectually and mentally work with them both, that goes way beyond being able to do SQL script and native HANA SQL. That takes a new kind of thinking.

Kristine: Well thank you very much, Thorsten, for speaking with us today, it’s been a pleasure speaking with you, and enjoy the rest of your TechED!

Thorsten: Thank you very much, it’s been a pleasure to be here!

An email has been sent to:

More from SAPinsider


Please log in to post a comment.

No comments have been submitted on this article. Be the first to comment!