Hazelcast CEO looks to grow real-time data platform

The need for real-time event streaming data continues to grow as organizations look to use updated data for operations, data analytics and business intelligence.

Among the many vendors in the space is Hazelcast, which on July 28 released a preview of its new Viridian serverless platform that provides a pay-as-you go-approach for the vendor's real-time data platform. The Hazelcast Platform on which the serverless offering is built is a convergence of the vendor's IMDG (in memory data grid) and Jet streaming data technologies.

Hazelcast competes with real-time data platforms from multiple vendors including Confluent, which is based on Apache Kafka, as well as Aerospike's real-time database platform.

In this Q&A, Kelly Herrell, who has been CEO of the San Mateo, Calif.-based company since 2018, outlines the challenges and opportunities he sees in the market for real-time data platforms.

What have been the surprises in real-time event streaming platform development that you have seen during your tenure as Hazelcast CEO?

Kelly Herrell: I think the biggest surprise is that the market we're unlocking is quite a bit larger than I hypothesized when I joined Hazelcast.

Our origins really were in financial services and large e-commerce vendors. Now you look at what's happening and we have customers in healthcare, automotive and logistics. Why is automotive using us? Well guess what, cars are digital, and they generate a lot of data. That data is not old contextual data, that data is real-time data, and they want to do something with that data.

Kelly Herrell

The hypothesis that I had when I joined Hazelcast was that low latency in-memory data processing is valuable, but the opportunity that we saw was something bigger.

The idea that making stored data go fast with in-memory technology is good and likewise processing streaming data, that's good and valuable. But they're like two ships passing in the night, as the two types of data are completely unaware of each other.

So the technology hypothesis that we had is that if we could unify in a single high-performance processing environment, stored contextual data with current streaming data, we could enable something that would be a big opportunity. Fast insights are good, but they're not valuable unless you act on it from a business perspective.

What do you see as the intersection of event streaming data and real-time data?

Herrell: A lot of times people think real time means a certain speed. The computer science terminology for a description of real time means determinism, it means acting within the same amount of time, every single time.

The way I think about it is that all data is born in the moment, and increasingly it's being born as an event or a stream. An event data stream does tell you what is happening in the moment; however, there is no context.

The conversations that we're having with a lot of business executives now is how they can benefit from being a real-time business. So usually when somebody asks us about event streaming, we'll talk about it and we'll up-level the conversation into what the business impact can be from a real-time platform.

Organizations are under pressure to compete as the world increasingly gets digital. Think about the hyperscalers, everything they do is real time. For example, if you visit Amazon and order a package, you can get it on the same day. They're not doing that with overnight batch process, but they're acting within some window of time and they're getting strategic advantage.

What's next for Hazelcast?

Harrell: One of the things that we've been very disciplined about here is we didn't go raise gobs and gobs of money and then go spend gobs and gobs of money trying to chase growth. We've been deliberately building a sustainable, growing business with a clear path to profitability.

You should not expect any blockbuster news anytime soon about something that happens with the company; we're just driving an ever-increasing rate of growth here.