Oracle Senior DBA Group

Oracle Senior DBA Group

Have something to say? Join LinkedIn for free to participate in the conversation. When you join, you can comment and post your own discussions.


Why would anyone want to migrate from Oracle to SQL Server?

Product Marketing Manager - Database Delivery at Red Gate Software

  • Comment (34)
  • April 14, 2011
  • Close viewer


  • Wilson B.


    Wilson B.


    I'm with Derek on the bigger picture, including ROI. Part of the issue is that people "like what they know and know what they like", so it is difficult to get an objective opinion.

    However, my experience of the UK market in particular, leads me to think that SQL Server has come of age and is a serious contender to Oracle in the enterprise database (or application platform) space. I see many companies and software houses migrating from Oracle (& Sybase) to MS SQL almost every day.

    I'll put my cards in the table and tell you that my business is based on migrating Oracle (& Sybase) to MS SQL and to me, the current environment looks very much like the early 90's when I was engaged in migrating clients from proprietary & mainframe to UNIX/Oracle.

    If you are old enough or live long enough, you will see that technology is cyclic.

    and by the way......look at an old map of Silicon Valley and see how many of the mighty have wrong turn and all that!

  • Venkata Sudheer Y.

    Venkata Sudheer

    Venkata Sudheer Y.

    Oracle DBA at IBM

    This is a very nice discussion in deed, my points what I have observerd is the ability to handle the large concurrent transaction is lower in SQL-Server than Oracle, also Oracle is platform independent, not sure why this point is not comming in the discussion, before the Microsoft cam up with SQL there was already bussiness running on Oracle and SQl is not path breaking or versetile.
    I can see only two options
    1) Graphical Interface than oracle
    2) Cost
    3) All at one (Microsoft can provide all needed to start the bussiness at one location from email management to backup solutions)

  • Ian

    Ian W.

    Ideas Hamster. 35 yrs dev experience. All sectors. All enivornments

    I've implemented a couple of conversions, and my view is that out of the box SQL Server functionality now implements virtually all Oracle modes of operation and most functionality, and the same is not true in reverse. The major concern of Oracle dbas, non blocked reads, is now one of the standard SQL concurrency modes, for instance. (This doesn't mean I'd write new app software for SQL using Oracle's native mode of operation, though.)

    The one thing I can think of sittin here now, that isn't native, i.e. Sequences, is coming in Denali (MS aren't stupid, they know what the target markets are, I believe) - though SQL Server does implement this functionality through a different mechanism which allows some very close hardware integration, particularly if you're a Kimball oriented chap.

    On top of that, virtually everything is free, Integration Services, Reporting Services, Analysis Services etc, once you've bought the licence.

    I don't consider SQL Server to be a relational database anymore, merely a software solution, one component of which implements a relational DB.

    This means that there's nothing between Oracle and SQL on functionality, (if anything SQL provides more,) and it's all down to cost. Sure, Oracle runs on everything, but this means hardware abstraction. SQL is never going to be non-windows based, but this conversely means lots of close coupled optimisation.

    From the dba perspective too, in my experience at least, it requires less to get good performance out of a box, because of the nature of the evolution of the product. It's not uncommon to see teams of 20 year dbas with a few lesser experienced chaps, whereas in SQL, there are many dbas with much less experience still managing to make it work. This is again, cost.

    Overall the bottom line is cost. For new projects, this is a simple calculation, and for migrations isn't isn't much more difficult, bearing in mind the nature of migration tools. One looks at a conversion, as I'm sure Wilson Bigg will agree, and after an automated conversion there is then still much more scope for further improvement, your ongoing development will improve performance, not destroy it, from the transfer point. Code written for Oracle ported over, is unlikely to be as fast as it could be in SQL Server, due to the nature of who wrote it. Once you've got the code over, retargetting is always going to provide more for your money, because of the close OS coupling.

    The reality is that Moore's law affects RDMBS's as well. At one time your local Bainbridges needed a mainframe, now you could run (though Engineering wise, I wouldn't advise it,) Centrica's entire billing on a Acer 8943G laptop. As performance increases, more and more programmes become a mere cost problem.

  • terry G.


    terry G.

    at Oracle

    Ian, where exactly is HCC in the evil empires failed attempt at rewriting sybase?
    The fact is they're are many features in even Oracle standard edition that are not in
    this failed attempt at an RDBMS. But spread the fud the evil empire is losing
    OS mkt share by the day. Of course running on stable and secure OS is key to
    any DBMS working properly. Given Linux is free or darn near it and the evil
    empire has the nerve to charge $2K for it's least failed attempt at a server OS
    is reason enough to pass it by for mysql, and other open source rdbms if cost
    is the #1 concern.

    Given the low cost of Oracle per user licenses I'm seeing a biz opening
    to move db off of the evil empire's failed attempt @ db and consolidate those
    one schema trick ponies on a low cost linux server running Oracle 11g.

  • Earl D S.

    Earl D

    Earl D S.

    Senior Director - Database Architecture at Scivantage


  • Yevgen S.


  • Alessandro

    Alessandro D.

    Oracle DBA & Developer at Tiscali

    "Google, Facebook use the techology that allows them to scale-up their applications better ", but at what cost? Generally at the cost of transaction integrity and you would trust your financial transactions to a technology that does not guarantee transactional integrity?!

    There are other (most) appliations than Facebook that need transaction integrity, finance applications for example, bank apllications, etc. ;

    if a manager choose to invest in a different technology when the staff is skilled on another technology, well I would change the manager... This would cost far less


Have something to say? Join LinkedIn for free to participate in the conversation. When you join, you can comment and post your own discussions.