Thought Leaders in Big Data: Eldad Farkash, CTO of Sisense (Part 4)

Sramana Mitra: There is one area that I think still has some openings. Some of this is very large opportunities but a lot of it is also small to mid-scale opportunities of these vertical solutions that take advantage of the more mature infrastructure layer underneath. But then, you need a serious amount of domain knowledge on top to be able to actually do meaningful things with that infrastructure. Are you in agreement with that statement?

Eldad Farkash: Yes, I do agree with that. I think there is a lot of potential for innovation on how to simplify those huge projects. I also think that there is a fundamental problem with trying to oversimplify complex systems. I think when we look at how scale out and scale up systems grew for the last years, their first intention wasn’t to be simple. They grew out of Google and Yahoo. For those companies, tying together 10,000 nodes is something achievable. The whole idea was that they are the only companies who can do it.

If we’re talking about startups who want to go out there and want to change the Big Data landscape, I would say try to focus on speed boats rather than battle ships. I have quite some experience with enterprise startups. It’s always impossible to try to beat those huge players in their own game. You need to come up with something that is very different in a dramatic way. Do we need 100 nodes or would it be possible to trim down those nodes into something that you can put in your living room? Hardware and technology trends definitely provide potential for startups to come up with tools to simplify those huge infrastructure projects.

Of course, just on the software layer, there is always opportunity. I’ve seen some amazing companies trying to really put those clusters in a box and make it very simple for someone who comes from the conventional IT to be able to manage those clusters. I really think this will be a really strong trend next year. With new cloud infrastructure coming up everyday—Raskspace versus Amazon versus OpenStack versus IBM—how do you manage those infrastructures as a startup? What do you invest in and how do you move your software from one cloud vendor to another as you grow and as your model changes? Definitely, there’s tremendous potential in software-only solutions that simplify those huge clusters.

This segment is part 4 in the series : Thought Leaders in Big Data: Eldad Farkash, CTO of Sisense
1 2 3 4 5