Amazon AWS, Amazon RDS, Azure SQL Database, Editorials, Open Source

Multi-Platform Development – a Pro or Con?

In working with different companies to try to define how they’ll best move forward with their data requirements, it’s increasingly obvious that multi-platform, where you have multiple database systems on tap to provide support for requirements, is an ideal solution.  By embracing the platform-agnostic approach, and looking at the requirements and using those to determine what options to consider, you can have a best of breed environment.  This, coupled with the fact that many systems requirements are now growing and changing at the speed of light, and it’s not a feat of getting less complex, means you’ll want to have all the resources you can muster at your disposal.

But, of course, it’s not that easy.  Just saying you want to use various platforms for functional or cost reasons isn’t enough.  The resources and expertise that comes into play are substantial.  In some cases, the platforms are different in very significant ways.  From NoSQL to MySQL to Aurora and Azure SQL DB.  From all of those – you also need to consider the platform.  On-premises, cloud, mixed environment.

None of this may be news to you and your efforts, but once you get outside your “orbit” of influencers, and quite possibly to the decision makers in terms of information use, things get pretty slippery pretty quickly.   There have been many times of late where a decision maker gets it in their mind just how things should work, how they must work.  This is all without regard to the technology pros and cons and without any knowledge of internal resources and expertise to support the choice.  In some cases, the conversations have become fairly panicked, as the IT folks try to pull back the technology stack to one that is manageable with on-hand resources.

We’ve recently been fighting this battle on several fronts.  From consulting to our own development, I fully admit that I suffer from “look, a rabbit!” syndrome with technology.  I get a good number of eyerolls when I start a meeting with something along the lines of “so… I have this idea…”  I can cement the scene with outright frustration sighs if I add “should be really simple and only take a day or two with the new tools!”  But this is reality.  The reporting and intelligence requirements of the users of data are, more and more, determining the tool selection, the platforms, the options.  This comes from their misunderstanding the how (or simply not caring “how”) and being more focused on the “what” – they want to know what they can extract from all of this information at their fingertips.

I’m positive this is where the now-famous “pointy-haired-boss” comics started way back when.

The thing is, the pull from these end-users of the information that we provide is getting very strong.  The expectations of fast solutions are driving the adoption of BI tools to work around having to wait.  Those same tools often insulate the users from the source of information and how it’s created and massaged.  This is a whole ‘nother issue with data ownership and pedigree.

I’m not sure we’ll win that control battle over tools, platforms and such.  I think that’s not necessarily a bad thing though.  It keeps the skills moving forward, it keeps things (certainly) very interesting and it keeps your job as a data person, whether it’s a DBA or analyst or scientist or however you’re going about it, important and a key to your company and clients.

So, my suggestion would be to embrace these requests.  To diving into the new options.  Do the webinars.  Learn.  Try new platforms.  Set up that database on a new database platform and see what the differences are.  Learn about NoSQL, Hadoop – the things you’re hearing about as rumors right now in your organization and from clients.  You don’t have to (can’t) learn it all at once.  But you can sure start learning a few of the differences and wrapping your head around the tools and options.

Have fun with it!