Home Forums Article Discussions Article Discussions by Author Discuss content posted by Deepa Parikh The FILESTREAM Data Type in SQL Server 2008 RE: The FILESTREAM Data Type in SQL Server 2008

  • September 24, 2008 at 12:45 pm

    #876617

    Kevin Farlee (9/24/2008)

    We have done extensive perf testing. Generally for blobs under 256K to 1 MB it's quicker to access via TSQL. For larger blobs it's quicker to access the file directly.

    [snip]

    The files can only be accessed via SQL functions, and under the control of SQL Server. There is a filter driver which verifies that the identity attempting the open has access to that cell in the database. Also, the access happens in the context of the current SQL transaction.

    Hi Kevin,

    Thanks for responding. One more point of clarification if you don't mind.

    When you say "For larger blobs it's quicker to access the file directly" does that mean "For larger blobs it's quicker to access the file directly via SQL functions, and under the control of SQL Server" (presumably using "column.GETPATH()" and "OpenSqlFilestream")?

    Thanks again,

    Scott

    Edit: I think your previous post answers my question.