NoSQL as a governance arbitrage
24Sep11
I got into a conversation earlier in the week with a techie friend about the merits of SSDs, which we both use these days for our main machines. It look a odd left turn when he said:
Funny part for me is that I truly believe the SSD revolution will result in a swing back to more traditional storage tech like RDBs.If you could get 100k+ IOPS for good old postgres, would you use something else? ;)
I replied that this ignored some of the social/cultural aspects that are driving NoSQL adoption. To a certain extent I see the entanglement of DBAs into the enterprise as a barrier of adoption to NoSQL, but there’s a flip side to this coin – so I replied:
NoSQL in the enterprise is sometimes about performance envelope, but in most cases I think it’s actually about escaping from the ‘cult of the DBA’ (and their friends the ‘priesthood of storage’). Most enterprise have a Dev:Ops barrier with the OR mapper on one side and the RDB on the other, so if people want to get with the DevOps groove they have to go some flavour of NoSQL – basic governance arbitrage.
He replied:
Possibly, but I think people underestimate the tech behind the top class RDB engines, and making them go faster is sometimes a better solution then switching.. What the SSDs are doing is lowering, dramatically so, the price to increase the performance of said RDBs…
He is of course right on both counts. RDBMS has dominated for so long, and found its way into use cases way beyond what ‘relational’ was intended for precisely because the underlying engines are so fast and flexible, and SSDs will just help to sustain that status quo (at least if people can come up with file systems that provide less of an impedance mismatch between the RDB workload and underlying storage substrate – but that’s another story). I do however think that NoSQL will continue to thrive and disrupt in situations where organisational structure and the governance processes that go with it are seen as barriers to progress.
Filed under: technology | 2 Comments
Tags: dba, enterprise, governance, nosql, rdb, rdbms, ssd
Subscribe
Recent Comments
Tim Coote on Ross Anderson RIP iain on Skiing in Espace Killy (Val… Chris Swan on Getting more from a British Ga… Murray Cowell on Getting more from a British Ga… JL on AI MacGuffin 
Pinboard.in bookmarks- Commission proposes youth mobility between EU and UK
- DEA fuels moral panic over ADHD meds to justify its failed drug war
- Meta.ai Oh My!
- Looking for AI use-cases — Benedict Evans
- the biggest threat facing your team
- Why has Hugh Grant settled his phone hacking claim against the Sun?
- Open Source Security (OpenSSF) and OpenJS Foundations Issue Alert for Social Engineering Takeovers of Open Source Projects
- AI isn't useless. But is it worth it?
- Pluralistic: The true post-cyberpunk hero is a noir forensic accountant
- Subaru battery drain due to missing 3G network
Chris Swan's Weblog 
Hacker News comment thread
Hi,
On the one hand I hope your friend is right: RDBMS isn’t suddenly a useless technology.
It does, however, suffer from an image problem. Both in terms of the ‘cult of the DBA’ (and their friends the ‘priesthood of storage’) as well as more emotional reactions to limitations of certain widespread and freely available engines. It seems that RDBMS as a whole has been tarred with that brush.
From where I’m standing, most of the NoSQL advocates have had a bad experience and want something else. That doesn’t mean that the big, professional, commercial engines haven’t been pulling their weight for a long time on enormous data sets. It’s just that perhaps NoSQL users have never had the opportunity to either see that, or afford it. I find it hard to believe that the average NoSQL user today has significantly more data than the biggest RDBMS users of yesteryear.
Of course, there are other factors in play as well. RDBMS has always had a reputation of being difficult to use and difficult or costly to architect for. Data Architecture as a discipline seems to be well and truly out of fashion and I’m guessing that the majority of NoSQL deployments are still small enough that it’s possible to brute force the way through, especially with hardware being what it is today. It remains to be seen whether this becomes a problem, or is admitted as a problem, in the long term.
There’s also the issue of talent. Using NoSQL allows you to attract the kind of developer that likes to use modern tools and keep themselves current. We have a quote on file from a CTO of a large web company that attests to the fact that his decision to “Go NoSQL” was motivated more by his desire to attract a generation of new talant than the feature lists of the engines.
Finally, whilst it may be the case that NoSQL users currently have relatively small datasets, the fact that it’s now possible to juggle large amounts of data inexpensively will almost certainly open up a whole host of opportunities that would previously have gone unexplored.
Regards,
@ndy