asterisk users - Nov 2007 - ODBC connection to Microsoft SQL Server

Hi,

I wish to integrate a Microsoft SQL server with Asterisk for CDRs and
for dialplan routing based on database values, and have this application
scale to a large number of simultaneous calls:  The Asterisk: The Future
of Telephony 2nd edition book states that:

? The pooling and limit options are quite useful for MS SQL Server and
Sybase databases. These permit you
  to establish multiple connections (up to limit connections) to a
database while ensuring that each connection
  has only one statement executing at once (this is due to a limitation
in the protocol used by these database
  servers).

Does this suggest any kind of performance issue with scaling?  I am
assuming not as all this indicates is that DB queries are pooled from
the ODBC connection on the Asterisk Server side rather than the SQL
Server?  Has anyone done this before in a large implementation?

Any advice appreciated.

Cheers

Robert McNaught

-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.digium.com/pipermail/asterisk-users/attachments/20071112/a2297fa6/attachment.htm

On Monday 12 November 2007 18:14:57 Robert McNaught
wrote:
> I wish to integrate a Microsoft SQL server with Asterisk for CDRs and
> for dialplan routing based on database values, and have this application
> scale to a large number of simultaneous calls:  The Asterisk: The Future
> of Telephony 2nd edition book states that:
>
> ? The pooling and limit options are quite useful for MS SQL Server and
> Sybase databases. These permit you
>   to establish multiple connections (up to limit connections) to a
> database while ensuring that each connection
>   has only one statement executing at once (this is due to a limitation
> in the protocol used by these database
>   servers).
>
> Does this suggest any kind of performance issue with scaling? 
If there is any such issue, it is due to a limitation with the database
server protocol, not with Asterisk.  The description should have made
that clear.
> I am 
> assuming not as all this indicates is that DB queries are pooled from
> the ODBC connection on the Asterisk Server side rather than the SQL
> Server? 
MS SQL Server is simply unable to accept multiple simultaneous queries on the
same connection, so this is the only solution possible to avoid any
collisions.
> Has anyone done this before in a large implementation? 
I've never done it in a large implementation, but I don't see any reason
why
it wouldn't work.  If you're really concerned with the scale of opening
a lot
of connections, might I suggest that you use PostgreSQL?

-- 
Tilghman