TIBCO Now

I was recently at TIBCO’s London user conference. Apart from the hotel (the Hilton Metropole) leaving something to be desired the conference was an excellent one.

I have been wondering for some time what the company’s overall strategy was: the various acquisitions the company has made seemed disparate and the big picture has not always (ever?) been well painted. Clarity, however, is starting to emerge out of the mist. In particular, TIBCO has three major product groupings labelled under the headings Connect, Unify and Augment. I can’t say that I am thrilled with this terminology, but the underlying principle is sound: connect stuff, bring it together and then leverage (analyse) that collected data.

The company also has another triplet: “cloud native”, “open platforms” and “AI foundation” which, while hardly unique, represents a sound set of principles upon which to develop and deliver products.

As far as those products are concerned, there were multiple announcements at the conference, some available now, some in beta. Of these I’ll mention two – TIBCO Cloud Metadata (effectively an enterprise catalogue) and Process Mining – and discuss one: ComputeDB.

Briefly, the Enterprise Catalogue, due to go into beta next month, is being built on top of EBX (the Orchestra Networks acquisition). It’s too early to go into detail about this but the big question for me is whether it will support the ODPi Egeria project. This, apparently, is not yet determined. As for Process Mining this is a bit outside by bailiwick but, as I understand it, it will allow you to identify bottlenecks in your processes as well as areas where you might apply robotic process automation.

And so to ComputeDB. This is based on the SnappyData acquisition and it is essentially a combination of an in-memory data grid with persistent storage. At the front-end it is Spark-based so it supports Spark SQL, Spark ML and so on. However, it does not use HDFS under the covers and is probably best described as a NewSQL database in that it has a distributed architecture but, in effect, stores data in rows and columns and supports CRUD capabilities rather than just creating copies of things.

Perhaps most importantly, ComputeDB integrates with TIBCO Spotfire (though you don’t have to use this) and TIBCO Data Virtualisation (TDV) where it may be used as a cache and a source of data. In addition, both ComputeDB and Spotfire integrate with TIBCO Streaming. Adding in TDV means that you can query historical stores of streaming data in conjunction with locally persisted data and remote data stored in other systems.

In other words you have an integrated environment that makes so-called Lambda and Kappa architectures out-of-date. Moreover, I think this is going to be a trend. Software AG has announced Cumulocity Data Hub which, partnering with Dremio, is taking a very similar approach and I expect other vendors to follow suit. That in turn implies more (streaming) analytics vendors acquiring database and data virtualisation products and/or vice versa. It’s been clear for a long time that there are too many database vendors and too many analytics vendors in this market and the convergence of streaming and batch analytics could just be the nudge the market requires for some long overdue consolidation. As for TIBCO: well, they got there first.