Systems Engineering and RDBMS

Scalable Shared Databases enhancements in SQL Server 2008

Posted by decipherinfosys on February 13, 2009

Scalable Shared Databases was a new feature in SQL Server 2005.  If you have not looked into this feature yet, you can read this KB article from MSFT to understand the concepts as well as see the steps involved.  This scheme provides an alternative approach to scaling out for read only databases.  The word “read only” is the key here.  Using this scheme, one can mount the same physical drives on even less powerful machines and thus allow multiple instances of SQL Server to use the same set of data files.  Reminds you of Oracle RAC where multiple instances work off the same database, however over there it is for both read and write and is much more powerful and of course, the architecture is very different.

So, this scheme does not require data duplication since all those instances work off the same data files but does allow to use the additional processing power of those machines – so, you can take advantage of more CPU, more memory, tempdb.  But this does restrict the IO bandwidth since the same set of physical files are being shared across the instances.  These are mostly used for offloading reporting (or adhoc read only queries – by Operations folks) load.

So, what’s new in SQL Server 2008 in this regard?   When this feature came out in SQL Server 2005, one of the immediate questions raised was whether this is supported for Analysis Services as well?  It is supported now in SQL Server 2008.  So, a single read only copy of an analysis services database can be now shared across different Analysis Services instances through a virtual IP address.  Advantages are obvious ones – such a load balancing architecture allows us to better utilize the disk resources since there is no redundancy of the data, cube processing is not impacted by queries thus allowing us to process a larger number of concurrent queris with faster response times.

This feature is available only in the Enterprise Edition.

3 Responses to “Scalable Shared Databases enhancements in SQL Server 2008”

  1. alee said

    I work for xkoto ( and our GRIDSCALE solution is like Oracle RAC for SQL Server, but with some significant architectural differences. GRIDSCALE manages an active-active pool of SQL Server instances (physical or virtual, local or remote) so you can do reads or writes. The reads are load balanced to provide a big scalability boost on performance. The writes are sent to all databases, and the application only waits for the 1st commit to return; however, GRIDSCALE ensures data consistency.

    Our CTO is giving a tech webinar/demo coming up soon – if you visit you can get all the details.

    • Hello,
      I am a product manager in Beijing SamerUnit Info.& Tech. Ltd.,Co.,I am interested in the gridscale for sqlsever,Could you give any Documents to me ? thank you!

      • Hi Yang,

        Thanks for your note. Gridscale was a product of xkoto who was recently acquired by TeraData. If you google “xkoto” and “gridscale” and “TeraData”, you can find several posts about it. You can find some older whitepapers and documents on it but since the acquisition, we haven’t seen much about it other than people commenting in discussion boards that gridscale has now been taken off the market.

Leave a Reply

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

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

Google+ photo

You are commenting using your Google+ 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 )


Connecting to %s

%d bloggers like this: