Lucky day

Greg Luck announced as new Hazelcast CTO

Lucy Carey

Ehcache inventor decamps from Software AG Terracotta to Hazelcast: JAXenter gets the full story on his unfinished business in the space, and hopes for the year ahead.

The news broke today that Java caching top gun Greg Luck has now signed up for a CTO role on the Hazelcast team, upping their arsenal against space rivals Software AG Terracotta (Luck’s former employers) and Oracle Coherence. JAXenter caught up with the Ehcache inventor and Hazelcast’s Miko Matsumura last week for the full story.

JAX: What made you decide to part ways with Software AG Terracotta?

Luck: I’d been at Terracotta for four and a half years, which is quite a long time, and in fact, if you go back to when I first joined Terracotta, out of everyone who was on the executive team then, I’m actually the last person to leave. So I stayed on quite a long time.

Software AG acquired Terracotta in June 2011 and initially, Terracotta was run as a stand alone business, and we were completely self-contained. And then, probably about a year ago, a decision was made towards integrating Terracotta, basically integrating the products into Software AG. So, sales, field engineering, marketing, many functions gradually got integrated into Software AG.

What we have today is, pretty much engineering and product management are the only two functions that are left from the old Terracotta. For me, I think there are more exciting opportunities outside. And, I like startups. I’m very excited about what Hazelcast is doing, and so I thought it was time to move on.

Looking at your career to date, you seem strongly geared towards entrepreneurial activities.

Luck: Yeah, I mean, I don’t really think of myself as a ‘startup guy’, but, I have been involved in startups for about 12 years, and they’ve all been very successful. was an Australian startup that ended up IPO-ing for over a billion dollars, and I was the chief technical guy there, and solved some problems along the way. And then I sold Ehcache to Terracotta, which was a transaction, and then, twenty months later, Terracotta got acquired by Software AG. So it’s been a very interesting career, the last decade or so.

So, going to a growing company like Hazelcast is the next step in your startup story?

Luck: I’ve just finished defining the standard in this area, so for the very first time, Java architects will have, really like an architectural plank that says, “Here’s caching and in-memory data grid. Here’s how you use it.”

It should be a standard part of your architecture, and it’s totally new. We have independent software vendors like MIREX and SunGard – just two examples where large enterprise software vendors will add in JCache, which will then enable their customers to essentially just plug us in for performance and prove latency things like that. So, a standard way that you can take an application and give it a turbo boost, pretty much.

At the same time that this is happening, and the size of the market, which, according to Gartner, is growing at 40% per year, was worth $250 million in 2013, and it’s headed for a billion dollars by 2016 – while all this happening, what’s actually happening on the competitive landscape is that there’s various vendors and projects are kind of stumbling and tripping. So there’s huge opportunity for Hazelcast.

If we look at this market segment, and we see that some of the competitors that I’ve competed with over the years have lost focus, or stumble, or have personnel drift on, then there’s actually a huge opportunity for Hazelcast…which is actually executing really well, to go in and actually seize that opportunity. So I’m super excited about that.

What attracted you to Hazelcast – both personally, and in terms of potential user appeal?

Luck: I’ve met Talip (Hazelcast’s founder) – we were on a panel at last year’s JavaOne – so I was aware of Talip. Hazelcast has developed a very good reputation as a first class open source in-memory data grid.

There are so many people that I know who are actually using it, who have chosen it, that I’ve worked with over the last few years, and that has a huge endorsement to me. It’s very, very rare to have that, and this is an area in which I have expertise – I’ve written a few projects myself over the years.

So the idea of joining in and accelerating the success of Hazelcast, to me, is a very exciting opportunity.

What do you hope that Greg will bring to Hazelcast?

Matsumura: It was really for the Brisbane thing, ‘cus we’ve already got Rod Johnson from Sydney. Actually, for us, we’re just super excited, because we see Greg as one of the industry luminaries.

He’s defining the JSR JCache specification, which is part of Java 8 – he’s there on the Java Community Process, he’s making these big JavaOne speeches…to us, he’s really, in a way, the father of Java caching. And, you know, the creator of Ehcache, the world’s most popular Java cache.

So, from our view, he’s part of our agenda to bring the top talent in the world into Hazelcast and try to really advance both in terms of technology, and also position in the industry. We definitely position ourselves as the leading open source in-memory data grid, and we’d like to hold that position, perhaps get an even more exciting place in the market.

So essentially this is the Hazelcast dream team coming together?

Matsumura: Yes – that’s what we’re working on!

Greg, anything you’d like to add?

Luck: I feel like I’ve got unfinished business in this space. And, this will probably be a story for later this month – the long awaited standard in this area, I’ve worked on for the last three years, and actually finished it in December.

So hopefully that’ll go to final ballot. That creates a standard in the space which we think will rapidly expand the usage of in-memory data grid in the enterprise. There’s already indications from some ISPs that they’ll be using the standard.

By standardising an area, it makes it safer for everyone to use, it allows independent software vendors to create a pluggable way in which solutions like Hazelcast can be added in.

It’s a really good time to join, because I think the market is going to grow significantly as a result of that. So just the timing of when I joined – this thing’s going to come out, and Hazelcast, which has already been going very, very well, has a new market that it can enter.

Looking forward to the future, what do you expect Hazelcast to achieve in the next year?

Matsumura: I’m happy to weigh in on this! Obviously, we’ve got big aspirations and goals…In a year, at this time, we’ll have rounded out our dream team, we’ll have booked a significant multiple of 2013 revenue, plus a significant multiple of 2012 revenue, and so we’ll continue to hug the exponential curve of revenue growth. From a company and corporate perspective, it’s very much on track to be  a high growth startup.

On the Java side of the world, we’ll certainly have our implementation of JCache complete, and we’ll be hopefully at a standard open source way for Java to talk to RAM in large clusters.

Caching will no longer be an after thought. In our view, in a year’s time, particularly with JCache, there’s going to be a tremendous feeling of, “Ah, memory first!”.

From our perspective, we feel that people are going to completely change about the way they think about caching.

I think right now, people think caching means something you do later. Greg has this thing he calls operational data store that we feel is going to be ubiquitous in the Java world, you know, with the adoption of JCache, which is that people are going to start thinking about in memory as the standard way. And they’re going to start doing it straight out of the gate.

Inline Feedbacks
View all comments