From MRI image processing to water pollution abatement, customers continue to find new and intriguing uses for SAP's new technology. In this interview, Dr. Berg discusses some of the more unique SAP HANA use cases and projects he's worked on.
DH: Hello and welcome. This is Dave Hannon with SAPinsider. Joining me today is Dr. Berg and we’re going to be talking about the many ways to leverage SAP HANA. Dr. Berg has some very interesting examples he’s going to share with us today. He’s also going to be presenting at the Reporting and Analytics 2013 conference in November in Orlando. Welcome Dr. Berg and thank you for joining us today.
BB: Hi, you’re welcome.
DH: I want to start our discussion today about the breadth of options for leveraging SAP HANA. Tell me a little bit about how you’ve seen them expand in a relatively short amount of time since its release.
BB: Actually this has been a very big success. I think that most people started off with HANA a couple years ago with the HANA standalone database or building their own stuff inside of it and then of course we’ve got HANA BW, and back in January we got HANA being able to run on ERP as well and then there’s the whole auto slew of products. You have rapid deployments solutions, virtual marts, pre-delivered content from SAP. We even have, beyond the rapid deployments solutions, we also have rapid marts – the old BusinessObjects rapid marts that are now running on HANA.
I think that it’s gone in a very, very short time period it’s gone from being kind of small and sort of proprietary to being a wide solution for all these tools and technologies.
DH: OK. I know you’ve been working with companies directly that are implementing SAP HANA in a variety of different ways, some with BW, some without. Maybe you could walk us through a couple of those more intriguing examples of how companies are using HANA today.
BB: Yes. I’m involved in quite a few HANA projects right now and it’s been very very different. We have one company that I am working with in Houston that basically went live this summer and they’ve got ERP on HANA and they chose not to have a data warehouse. Their data warehouse is going to be analytical, calculation, attribute views and then exposed into BusinessObjects. So they don’t even have a data warehouse.
Then we have another company, a very large one that I’m working with, they looked at their current BW system and said, “You know what, I don’t like the way we defined things. It’s eighteen, it’s basically twelve to fifteen months of junk and then we have ten years of legacy stuff.” So basically put in a new HANA system including BW on HANA and then reconfigured and greenfielded the whole implementation and said we’ll leave the old BW alone and as we start doing moving piece by piece over to the new HANA box.
And then I have another client who said no, no that way is too expensive. What we will do here is basically we will export and import to a database, which will migrate the whole database system to HANA and then we’re going to take advantage of all the new layers-scalable architecture, so basically we’re moving layers inside the box and then optimizing it for HANA performance. So we can find all sorts or varieties of how the implement.
I’ve got a company in Colorado, and with that company we’re brownfield. That’s a new name I never heard before, but it basically is taking BW on HANA system, raise it up and then only migrate the pieces that you want and then redesign the new stuff.
So we have like four different scenarios here that we didn’t fully expect that when we saw when HANA first came out.
DH: Interesting. Were there any technical challenges that any of these companies ran into that you think other companies can benefit from hearing about?
BB: Yeah, I have one of the largest HANA implementations in the world we’ve been working on for about eight months and one of the things we ran into was just pure export. Exporting the old database and then importing the database into HANA. We sat around for eighteen days. It took about eighteen days, this is about 40 terabytes of raw data and it took eighteen days. We had a team of eight to ten people sitting there looking at the system.
Later on we were able to go in and add more indexes, bitmap indexes, we were able to split up the jobs. It wasn’t just straight out of the box and we got some second run tests we got it down to seven days. On the third run test we got it down to five days and we think we have even more improvements. But that was very surprising. Most slow systems go very fast and very easy, no complexity at all. Yet when we get into those tens of terabytes, I think you need to be a little more careful. Don’t just use the regular run books. That was a technical challenge that we ran into.
DH: Ok, ok, good. How about down the road, do you see this sort of trend continuing, the breadth of use for SAP HANA expanding even more and in the future as companies learn more about it and get a little more creative with it?
BB: Yes, it expands way beyond SAP. I’m working with the University of Pennsylvania on the medical field and what we’re doing there is basically using MRIs. So we’re taking 4D MRIs at seven teslas, which is an extremely high resolution about seven times more resolution than a normal MRI and then we’re able to process it. That normally takes processing time of about an hour after we’ve taken the pictures in 4D. Then we have to spend about an hour processing it. We’re now able to get that down under two minutes. That’s a totally new revolutionary use. We don’t use it for reporting. We don’t use it for transaction. We’re using it to generate an extreme high-end imaging in the medical field.
BB: I think we have a lot more use cases, facial recognition software, to detect people for security services. There are data mappings. There are war rooms, pollution tracking. I know this, the Russians are looking at … I gave a speech in Moscow this summer and the Russians are looking at the Spektr-R telescope where they’re basically generating 1,000 times higher resolution than the Hubble and they haven’t settled on how just yet, but they’re looking for a much faster data performance. So in the scientific field there’s a screaming need for high-end computing that doesn’t cost as much as grid computers and cluster computers and super computers and that’s a totally untapped market where I think HANA will play a big role in the future.
DH: Well, so a lot of us tend to think of HANA as a complex technology. These examples would sort of lead one to believe that it’s more flexible than we might think.
BB: Yes. The fear factor is they also put HANA Studio inside of them so you can build your own raw material. Last week I was up at the University of Wisconsin. I did a one day HANA workshop for the university alliance, the SAP University Alliance and when I worked with them, in basically one day people can build tables using data services, loading the tables, putting Business Explorer on top of it and we do it in one single day in a workshop. It’s very very accessible technology with extreme high end of freedom. Don’t think about it as a proprietary lock system. It could also be used as a total development platform for any apps that you’re interested in.
DH: Mmm, hmm. Ok. Lastly, based on the work that you’ve done, do you have any advice for companies that might think HANA’s not necessarily a good fit for their business today?
BB: Yeah, I hear that quite a bit and I think there’s a miscommunication here or misperceptions that HANA’s going to be in the millions and millions and millions. Sometimes it is because we’re looking at a very high-end boxes, but for most mid sizes to small companies we’re not talking about millions here. You’re talking about a few hundred thousand and sometimes, you can actually get a box right now for some smaller vendors for thirty to forty thousand dollars up to about a hundred a hundred and fifty thousand.
So, I think when I present those actuals, I usually have three quotes with me from three vendors. They are three different ones and I show them to my clients that here’s what you really will be paying. Most of the reaction I’m getting from them is, “Wow, it’s that inexpensive?” It’s not inexpensive, it’s just less expensive, but I think there’s a miscommunication here that people have perceptions that this is always in the millions of dollars, which absolutely it is not.
DH: Ok. Great. Great. If you want to learn more about SAP HANA from Dr. Berg he is the co-author of SAP HANA: An Introduction from SAP Press. He’s also going to be at the Reporting and Analytics 2013 conference in Orlando in November. Dr. Berg, thank you very much for joining us today.
BB: You’re welcome.