Sramana Mitra: The other thing that is vertical specific and what lots of companies are working on is the heuristics of these analytics. The applications you get out of the big data analytics depend on what you search for, what heuristics you build your models around, and what kind of learning algorithms are on it. There are a variety of things that are very domain specific. That is an area that is active but also underserved.
SC: At the end of the day, whatever that machine learning algorithm, that is the benefit of the opportunity. If the customers who were in early are able to create algorithms in analytics that give them a leg up in the industry, since it is early, that is great. In many cases the vertical solutions will come in and begin to enable the next wave and the later adopters to take part. But right now I would argue that a good bit of the next three years will consist of chasing opportunities, understanding what those repeatable patterns are, and operationalizing at scale. Entrepreneurs who are focused on unlocking some of the more valuable ones there is a great opportunity to go to business.
SM: Do you have any other thoughts you want to share?
SC: One of the things I put out on the role at Hortonworks and in general, from an enterprise Hadoop perspective, enabling a technology like this so that it integrates well with what enterprises have is important. In large part [it means] demystifying and making Hadoop enterprise consumable is to integrate it well with a lot of these schools that people are familiar with, doing it in a way that speaks similar languages and maybe evolve existing skills, whether you are a SQL developer or a classic database administrator. There are elements of Hadoop which, if thought through properly, you can approach properly and in a very familiar way, and not as something like, “This is something new and maybe I am going to be out of a job.”
Our focus is to make sure it integrates well so that it gets towed in to the enterprise. Partners like Teradata, SAP, or Microsoft help us do that by integrating it into their databases. Once that common platform is in place, it gives a stable platform for these other vertical apps and others to flourish on top of it. There is a lot of exciting work ahead.
SM: There is a tremendous skill shortage in the industry right now.
SC: Exactly. You can train people in the new technologies, but with technologies like Apache Hive, which provides and SQL interface on top of Hadoop data, the more standard SQL it can be, the better standard performing it can be and the better integrated it can be with some of the classic tools like Business Objects or Tableau, that are used for reporting and interacting with that data. That enables the data to get to the user in a familiar way and breaks down the barrier to entry. More needs to be done, but we are making great strides in the open source community and also in the traditional vendor ecosystem.
SM: Thank you, Shaun.
SC: Thank you.