Data access (JDBC / ADO / OLEDB) performance comparison for Sybase ASE 12.5.3 and Microsoft SQL Server 2005

As I mentioned in an earlier post we are currently in the process of migrating from Sybase ASE 12.5.3 to Microsoft SQL Server 2005.

One of my big concerns from the outset was the performance of Microsoft SQL Server 2005 or more to the point the performance of the Microsoft SQL Server 2005 JDBC driver.

We have been doing some performance analysis which is producing some interesting results.

We have developed a Java console application which uses our applications database abstraction layer to call a stored procedure several million times. The stored procedure simply inserts a record into a database table.

We ran this program against Sybase ASE 12.5.3 on Redhat Enterprise Linux and Microsoft SQL Server 2005 x64 SP1 / Windows 2003 Server x64 SP1. Both RDBMS’s were installed on servers of the same hardware specification.

  • Syabse ASE 12.5.3 using the Sybase JDBC driver managed 780 calls of the stored procedure a second.
  • Microsoft SQL Server 2005 using the Microsoft JDBC driver managed 463 calls of the stored procedure a second.
  • At first glance it appears that Sybase ASE 12.5.3 is much faster than Microsoft SQL Server 2005.

    However we developed an equivalent program in Visual Basic 6 using ADO and the respective OLEDB drivers to call the same stored procedures. The results were

  • Syabse ASE 12.5.3 using the Sybase OLEDB driver only managed 120 calls of the stored procedure a second.
  • Microsoft SQL Server 2005 using the Microsoft OLEDB driver managed 1,106 calls of the stored procedure a second.
  • The Sybase ASE 12.5.3 results were so poor that we downloaded the new OLEDB drivers that Sybase have recently released from their web-site. The performance was significantly better, although still well short of the Microsoft SQL Server OLEDB driver, but it appears the new Syabse OLEDB drivers have a serious memory leak and we were unable to complete the test.

    What does this prove? It goes a long way to proving that the performance of your application depends a lot on the technology you use to connect your application to your database and that you should thoroughly test the performance of the drivers.

    I am now off to post an article on one of Microsoft’s MSDN Managed Newsgroups and email somebody I know at Microsoft.

    Over the next few days I will post on the raw performance of Sybase ASE 12.5.3 and Microsoft SQL Server 2005.

    Advertisements

    Leave a Reply

    Fill in your details below or click an icon to log in:

    WordPress.com Logo

    You are commenting using your WordPress.com account. Log Out / Change )

    Twitter picture

    You are commenting using your Twitter account. Log Out / Change )

    Facebook photo

    You are commenting using your Facebook account. Log Out / Change )

    Google+ photo

    You are commenting using your Google+ account. Log Out / Change )

    Connecting to %s


    %d bloggers like this: