Hey now,
OK, I looked and was correct.
The MDAC path to a SQL Server is:
ADO
|
OLEDB---------|
| |-SQL SERVER
ODBC-----------|
The code path is
ADO or other
OLEDB Consumer
OLE DB
OLE DB
Core Services
Core Services
|
|
Native OLE DB Provider OLE
DB Provider
|
for ODBC Drivers
Database
|
ODBC
Core Services
|
ODBC Driver
|
Database
You can see that the code path for ODBC on the right has two more layers to
go through than the OLD DB directly does on the left.
You do not want to store connections in the session object, but use OLE DB
Session Pooling.
Your DSN-less connection will look like:
SConn = "Provider = MSDASQL;Password=??;UID=??;Driver = SQL
Server;server=???"
For a Native Provider, the shortest route in, use:
SConn = "provider=SQLOLEDB;server=MyServer;UID=??, PWD=??"
HTH,
Doo
--
Doo
Senior Data Architect / DBA
PlanetJam Media Group
> I'm no OS guru...but at the SQL2TheMax conference I attended a session on
> this. What stuck is that OLE-DB has less levels to go through to get to
the
> database than ODBC, therefore it's faster.
> I stand corrected if I got that wrong as I don't have the session notes
> handy...Friday is a telecommute day and they're at the office.
> --
> Doo
> Senior Data Architect / DBA
> PlanetJam Media Group
> > ODBC Connection Pooling Vs OLE DB Resource Pooling
> > Hi
> > Everyone
> > I am a Sql server DBA at a B2B Internet Startup
> > I have to determine which type of pooling Should we use to increase
> > overall performance
> > we use Clustering /Sqlserver /ADO/VB/ASP
> > There is no document which tells me which Pooling is Beneficial for
> > Performance
> > If some one can help me with some Practical Advice
> > Thank you
> > John