Listen in as Ovum's Madan Sheina reviews a recent Ovum report on enterprise data warehousing trends. Topics discussed include:
- The methodology Ovum used for conducting the research
- What defines an analytic database
- The value of in-memory databases
- The drivers behind the need for an analytic database
- The role of Hadoop in analytic databases going forward
View the video, and read the edited transcript of this conversation with Madan Sheina here:
Ken Murphy, SAPinsider: Hi, this is Ken Murphy with SAPinsider, I’m reporting live from SAPinsider’s BI and HANA 2014 event, and I’m pleased to be joined by Madan Sheina, who’s a lead analyst for Ovum Information Management Software. Madan, welcome.
Madan Sheina: Thanks. Good to be here.
Ken: I wanted to ask you about a report that you authored earlier this year on decision matrix, selecting the right analytical database, and HANA placed pretty well in this report. Could you just tell me a little bit about the methodology behind this report and how you, you know, chose, the criteria you used to choose the best rankings?
Madan: Sure. Ken, the report really started about a year ago when we looked at the market and we saw a specific trend in the market for enterprises that had invested heavily in enterprise data warehousing and SQL-based technologies. And we’re now really looking for smarter ways to analyze data, but also smarter infrastructure for processing data as well. So it was important really to get the definitions right, because when we’re doing a comparative report, we want to make sure that we’re comparing apples to apples and not apples to oranges, and there are a lot of different database technologies out there, you know we came up with a definition and we came up with inclusion and exclusion criteria, and the key tenets of the definition was, what we term as an analytic database, which is what the report is all about, has to do processing differently, or more optimally than traditional OLAP type of analytics.
And the other key tenets are it has to scale and it has to be high-performance, and it has to be delivered in a very simple way that is either pre-configured as a hardware/software type appliance, deployable in the cloud, etc., so we came up with the criteria, and we came up with three real categories, what we term as advanced SQL, really driven by different, new and different, you know, innovative types of infrastructure, which we identified you know, as being around MPP, columnar databases and in-memory databases as well. So we came up with a list of eight vendors, a shortlist of eight vendors, I wouldn’t say it’s an exhaustive list, there were some vendors that did not want to participate because they didn’t agree with our definition, but that’s our job as analysts, to set our own frameworks.
Ken: So of those eight, I know HANA placed pretty well. You termed it a market leader, one of a few, is that correct?
Madan: It did indeed, I mean, the actual methodology was based on how we evaluate it was—there were three elements to it, the first was a technology assessment, and this is really a way for us to differentiate key functionality between the products that we evaluated, based on criteria such as data access, high-performance analytics, scalability, availability, systems management—there were eight key dimensions and we really did, we went to the vendors and we mapped discrete functionality, and I’ll tell you, I think it took us about six months to devise this framework, and there are about 300 discrete elements that we analyze and focus on. It’s a big commitment and effort on our part, equally on the part of vendors as well. The second area we looked at was market execution, how well the vendors actually—how innovative they were, how interoperable their solutions were, how mature their solutions were in the market, and then the third area really related to uptake in the market in terms of, you know, revenues, and the customer base, and the geographic expansion. We mapped that out against our own sort of definition of market leaders, market challengers, and market followers. SAP I think, well, HANA, came up as one of the leaders, along with three other vendors as well.
Ken: So what then, in your estimation, set HANA apart in that category? You mentioned a few different types of things you were looking at, innovation, market—what stood out for HANA?
Madan: Well I think, you know, we were obviously impressed with the in-memory capabilities, and SAP’s invested a lot of time and resources into this, and they placed a really big bet on this as well. And we believe it is, unlike some of the other solutions where it wasn’t acquired, it was engineered from the ground up by SAP to be an in-memory, analytic database. So the engineering was sound, it was robust, we were also impressed with the scalability, and the performance. Obviously the performance capabilities was, you know, with an in-memory, is you know very important. The other aspect that really impressed us was the vision that SAP had around HANA, not just as an analytic database, but also as a hybrid OLTP database as well. And that vision is still yet to be fulfilled out, but we did definitely like the long-term vision of it being a platform.
Ken: Yeah, we heard a lot about that this morning at the keynote, the vision of it being a platform. Do you see organizations really adopting this mindset as well, just the need for, to have an analytic database and the in-memory technology or something like it?
Madan: Well, I think there’s a couple of business drivers, and technical drivers, that really reinforce that view, and the first is obviously the sheer pace of business, instant response, and the other aspect is we have more data now, with the advent of big data. They’re trying to reconcile those two together in a cost-effective platform, I think is what’s driving this going forward. Coming to your question about, you know, the use cases, it used to—let’s get one thing clear, you know, not all analytics requires real-time processing, and it used to be restricted to a niche area around financial trading, and it used to be a rich man’s game, as well. We’re now seeing that changing, you know, with the declining memory costs and even more you know, just kind of bigger use cases around big data in other areas, you know genome processing, fraud detection, which is more about instant response, this is all about processing large amounts of data quickly and getting insights, you know, quickly as well. But I think what’s maybe even lacking in the past is we have these insights, but we can’t act quick enough to take action on them. Which is what I actually liked in today’s keynote with Steve Lucas when he talked about going that extra mile and making the insights actionable as part of the process and as you know, SAP has always been about the process.
Ken: Right, so, last question: this report, the matrix that Ovum prepared was for 2013-2014, is this an annual, updated annually, how does that work?
Madan: Yes, we do plan to update it annually, once we recover from the effort. But I think it’ll, as the market changes, our, maybe our evaluation model might change as well, a little bit. We’re planning on an update next year, and what we’re probably going to be including more of is more integration with Hadoop environments, because we see Hadoop playing a key role in analytic databases going forward, not as a replacement for, but as a complementary platform and infrastructure that it can leverage.
Ken: Great. Alright, well, Madan, thanks for your insights, and thanks for joining us.
Madan: Great, pleasure.
Ken: Again, this is Ken Murphy with SAPinsider, and I’m reporting live from BI and HANA 2014, and I’ve been chatting with Madan Sheina of Ovum.