kdb+ versus MongoDB versus McObject versus OneTick versus...

For a project I am assisting a client in a beauty parade to be followed by a bake-off if required (in plain English, search and selection of a software product followed by a limited implementation to allow hands-on evaluation of the tooling).

A partial list of contenders is 

• kdb+

• OneTick

• McObject

• MongoDB (Man AHL 'Arctic' flavour)

[other names removed from list for commercial/contractual/confidentiality reasons]

I have some experience with MongoDB and kdb+, OneTick and McObject are new to me.

Having spent several years in the old world of Codd and the five normal forms along with the big iron required for "enterprise" grade RDBMS the world of NoSQL, NewSQL and columnar/tick databases is exciting but with a steep learning curve...

The lecture series from First Derivatives (owners of kx systems and the kdb+ product) is entitled "q for Gods" and it's fair to say that someone with a relational database mindset will pick up the concepts, it's not so easy.  In part this is due to the lack of really powerful IDE tooling which is something that is often the case with more niche toolsets.

One of the critical parts of the evaluation is the set of tooling that is supported - does the solution support all the tools required by the client?

• kdb+ - R, JAVA, C#, C, Python, Scala, Excel RTD,

• OneTick - C++, JAVA, Perl, C#/.NET, Python, SAS, MATLAB®, S-PLUS, R, TibRV, ODBC, 

• McObject - C/C++, Java, C# (.NET), Python, ODBC

• Arctic - Python

This of course needs to be reconciled against the client requirements today and on a strategic basis, no point on insisting on a Fortran interface if that is not a strategic requirement and can be allowed to wither on the vine as a legacy platform.

More to follow...