Our database architect told a client of ours that SQL Server 2014 is a poor choice over 2012 because it lacks the performance and stability advantages that 2012 has. Everything I've read contradicts this. Aaron Bertrand says the exact opposite and all the white papers I've read from MS agree too - 2014 is enhanced and resolves some of the AG issues people had with 2012.

Is there any major disadvantage to choosing 2014 over 2012 that I'm missing?

  • 11
    I think that your database architect is drunk. Feb 15, 2016 at 13:11
  • 1
    Extraordinary claims require extraordinary evidence. You have your arguments and references (and they are good ones). Approach your architect and ask them for details to support their position. Feb 15, 2016 at 13:17
  • 2
    Actually you can not choose SQL Server 2012 over SQL Server 2014 for new installations today because you can't buy licenses for SQL Server 2012 any more (at least not directly from Microsoft)
    – user1822
    Feb 15, 2016 at 13:20
  • You should make choice according to your requirement and budget. I would say both are good and you should not compare. Though 2014 has some very good features but if you have legacy application and you are trying to migrate 2014 the new CE is going to give you more pain than 2012. Apart from this 2014 is very good choice.
    – Shanky
    Feb 15, 2016 at 13:29
  • @Shanky regressions with the new CE are rare compared to the number of queries that benefit overall. Regressions can be dealt with individually or, in the worst case, the old CE can be implemented by force, making it a poor argument to avoid the version itself IMHO. Feb 15, 2016 at 13:38

2 Answers 2


You obviously don't need my confirmation but I can't think of a single possible reason why I would suggest to a client they go with 2012 over 2014. Particularly for support / end of life reasons, but also because of additional features and enhancements to existing capabilities.

Some reasons you might hear:

  • Performance degradation. Sure, this can happen, but it can also happen when going to 2012 or making any other major change (some of the reasons can also be attributed to migration without upgrade, failover, etc). Regressions are an expected part of change, and it's easy to blame an upgrade (just ask the VSTS team).

  • Higher cost. I've seen a few claims that 2014 costs more than 2012 for the same configuration, but haven't observed any licensing changes that would cause that to happen except that standby secondaries now require software assurance to be licensed. The other aspect is if someone without SA planned poorly and bought 2012 and now wants 2014 - yeah, that's going to cost them.

Generally worse performance? Less stability? No, both seem laughable to me. Ask your drunk friend for some real evidence, and maybe keep him distanced from clients, too. :-)


Well SQL Server 2014 comes with some serious changes to the Cardinality Estimator. I can only guess he was afraid of these changes and the effort of testing everything so the performance will not degrade.

Or he simply upgraded SQL Server without properly analyzing the changes occurring due to the new Cardinality Estimator. And worst comes to worse, had some queries in his application, which were negatively effected by the Estimator changes!

Be aware, the new Cardinal Estimator might handle your queries differently and some are bound to be effected negatively and will cause bad performance and an unstable system; well at least they might! But (!!) that is - if at all - only true for certain queries, not for the product in general! And these queries should be fixable.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.