What’s the intersection between the Big Data stack and real-time capabilities? The convergence is long overdue, says Trestata founder Abhishek Mehta, who discusses the market juncture on a recent NewsDesk segment (full video below).

Mehta first describes how real-time analytics gradually transformed from a nice-to-have to a must-have in the past 12 months.

“The capabilities you need in analytics to solve some business problems need real time ability, as well as real time capabilities to not just query and play with the insight, but also find answers to some topics,” Mehta says.  “Real-time and its interception with Big Data is a must-have technology piece in any Big Data stack you may choose to adopt.”

The need for real-time analytics brings up a new question: what exactly constitutes real-time? Mehta calls this “the multi-billion dollar question,” and answers it by saying that the definition depends on the specific business problem that you’re trying to address.

On one hand, a high-frequency trading algorithm that processes transactions in real-time has to maintain sub-millisecond performance in order to live up to its name. On the other, an application that fulfills a different role might operate in 24-hour cycles and still carry out its task just as efficiently as the former solution.  In the context of data analytics, real-time describes the ability to deliver results as fast as possible.

Mehta goes on to say that the new Big Data stack can be divided into three distinct buckets: real-time query, real-time insight, and real-time action.  The first category has received a lot of attention recently thanks to the launch of Greenplum’s HAWQ, Taze and Impala, but it’s the latter two buckets that got Mehta interested. Open-source platforms such as Project Spark and Twitter’s Storm are disrupting the way in which insights can be extracted and translated into value.