Not quite Star Trek

Capsenta is, unfortunately, not called Star Trek. Otherwise I could make bad puns about the Federation and the next generation. Because Capsenta’s Ultrawrap is what you might call a next-generation data federation engine.

I have written enough about data federation and data virtualisation in the past that I should not need to describe what it does or extol its benefits. But Ultrawrap does some things that I have not seen from any other vendor in this space.

To go back a step, Capsenta is a spin-off from the University of Texas at Austin (which is a shareholder). Specifically, it is based on research around the use of semantics and RDF (Resource Description Framework) in conjunction with relational databases (as well as other traditional sources such as Excel).

The company has done several things that are seriously interesting. First, it automatically semantically analyses your data (in situ), which may be contained in multiple sources, and creates a graph of all the relationships that exist across the data that it then maps to standard ontologies using probabilistic matching techniques (this is subject to user verification). The company has focused particularly on the healthcare sector because there are many established ontologies in this sector but, in principle, Ultrawrap would be suitable in any industry wherever an ontology can be established (i.e. Finance Industry Business Ontology for finance).

Second, Ultrawrap provides semantic search capabilities and, in particular, what is known as faceted search. In case you don’t know what this is, it means that the software will semantically parse search terms to get the best results. For example, you might enter the terms “child kidney cancer” but the software will process “paediatric renal oncology”.

The third really interesting thing that Capsenta has done is that it uses SPARQL for query processing. That’s no big deal. What is a big deal is that it has developed a patented method that dismisses the notion of performance degradation of graph based queries. In a nutshell, Capsenta has figured out how to run SPARQL queries as fast as SQL queries.

That is seriously impressive: apart from anything else it means that Capsenta does not have to write its own database optimiser. Given that companies like IBM and Oracle have been developing optimisers for thirty years or more, it is reasonable to assume that – in the words of the song – nobody does it better. So, it’s better to leverage their expertise rather than reinvent the wheel. With Capsenta’s patented approach, query performance in a virtual federated setting is optimal.

Capsenta’s current implementation is just the beginning: Ultrawrap supports what the company calls simple federation, where you are aggregating data from multiple data sources. In subsequent releases Ultrawrap will have the ability to do things like joins and matching records across databases.

If “cool” was in any way a contemporary word for expressing admiration I would describe Ultrawrap as cool. But it isn’t so I will content myself by saying that I am impressed.