Thought Leaders in Big Data: Interview with Robert Youngjohns, SVP and GM at HP Autonomy (Part 5)

Sramana Mitra: Once the architecture of the world is much more mature, you are coming in at the heels of a platform as a service trend. So, you basically provide a platform as a service, provide the heuristic layer on top of that, etc. We were talking about business models earlier. You can charge a per user fee, for example. Then the vertical folks jump at their piece of it.

Robert Youngjohns: Maybe I am a little old fashioned – I have been in the industry for a long time. The way I would use this is as an operating system for human information. If you think about an operating system you abstract common functionalities, access to data, etc. which is why I formulated that term, because another important part of it is how we can actually see in to all the various data sources. We built a very substantial group of connectors, which allow people to break in to files and understand what is going on in those files. On a simple level it could be an Excel or a PDF, but on a more complex level it could be a video image or an audio signature. We create an industry around what we are trying to do here, because I think people underestimate the ability to actually get at the data. You talk about big data today and everybody thinks Hadoop solves every problem.

SM: I had this conversation multiple times already. The storage layer is not where the complexity lies. It is more in the machine learning algorithms and in how you process the data.

RY: That is exactly what we think. If you want to use Hadoop here, good, we will work with it. We will create connectors, take out data from the sources of records, dump them into Hadoop, and then we will put the machine learning layer on top of that. We are fine with that model. In fact, in HP Software we are working on a project called Hayward, which is Hadoop. We are trying to produce a platform there, which just happens to have Hadoop as its big data source. At the same time we are exposing application development capabilities to others through this API layer. So, the idea is that as an application developer you should be able to think there is a service layer that has all these exposed services in it, which range from unstructured data through to structured data sources. That is something we will be talking about a lot more within the next few months.

SM: A number of players in your space are thinking exactly the same way. They have expertise in a couple of verticals, and then they are going to expose their platform as a service and develop their networks around it.

RY: I think that is a workable model. I keep telling my team: “When you see a model that has been in the industry for a while and it seems to work, that is probably because it works.” That is not a suggestion to make things completely different. If that is a model that works, we should do it. But the way we are trying to pull it together is that combo of  Hadoop as your big data lake, Vertica as your structured big data engine, and Autonomy as your unstructured big data engine. Our own application portfolio will sit on top of that, and it includes things like workflow management for legal, our archiving product, eDiscovery products, etc. There is a plethora of applications that one by one look a bit dull, but when you build them in that way you can go to an enterprise and say that you actually have a good end to end story, and also one that allows you to plug in other vertical solutions that you need.

SM: That is interesting. I would like to hear from you again once you are ready to work with the startup community.

RY: We would be happy to do that.

SM: It was very nice talking to you.

RY: Thank you for your time.

This segment is part 5 in the series : Thought Leaders in Big Data: Interview with Robert Youngjohns, SVP and GM at HP Autonomy
1 2 3 4 5