InterSystems Summit 2022

I’ve just been to the InterSystems Summit in October, in Birmingham, and it was both informative and great fun. A good opportunity to renew my acquaintance with Mike Fuller, Regional Director of Marketing, InterSystems (IS) in person.

For some time, IS has been focusing on IRIS, basically a fusion of its Caché database and Ensemble data integration, with other capabilities, for the Cloud. See our InBrief on IRIS here. So, what is new just now? Well, in GA or Preview we have: embedded Python, kafka adapters; FHIR enhancements; adaptive analytics and SQL operators; columnar storage (this doesn’t really compromise IS’ famous abstraction, it’s in the form of “hints” to the platform); a new rule editor; and, of course, new cloud services and platform updates.

To just look at one of these, what may excite developers new to IRIS and its capabilities is the availability of Python as a first-class language for IRIS, built into the core product. Experienced developers tell me that the language originally provided still has advantages and isn’t going away – but lots more people speak Python.

So, InterSystems is alive and well and now very much into verticals and selling “products”. Even though the IRIS data platform/data fabric (what I tend to call, in my old-fashioned way, a DBMS – database management system, although that old term has become rather debased recently) underneath it is entirely generic. In part, this is because it is punching above its weight – with limited marketing budget. I am still very impressed by the technology. It is one of very few DBMSs that completely abstracts the logical data models (note the plural – the relational model is only one option) from the underlying storage mechanism (sparse bit maps).

My view that Phillip T. (Terry) Ragon, InterSystems CEO and founder is very protective of his product was confirmed. In part, this makes the IS platform as good as it is, and he gives chosen employees (IS is very picky about who it employs – technical excellence is just the start, passion, ethics etc also matter) a lot of autonomy. But any changes to the underlying (basically Caché) database model or its associated language has to get past Terry’s detailed QA, although the Python story suggests that this isn’t stifling innovation in any way.

One of his top people asked him “what happens when you die” (which was my question to the top person), probably expressed in terms of “succession”, and Terry replied “I’m going nowhere”. Which is one of the few issues I have with IS – but the person I was talking to said there was already a strong management team below Terry. There needs to be.

Any product needs an “exit strategy” in my opinion, as do its customers, even if no-one expects to use it. But, that said, much of IS’ welcome customer and quality focus comes from its being a private company and I see no signs of IS facing succession problems yet. I also see good public companies distracted by “shareholder value” instead of “customer value” – and valuable public companies killed by hostile takeovers, by competitors who just want to kill the competition. I think IS is still a private company that is well worth watching.