I am trying to figure out what you two are talking about now that doesn't
make any sense about the speed of something running through SQL Server QA,
which is in direct contact with SQL Server, as opposed to something that's
running from a program that using OleDb
One of the reasons the SQL .NET Data Provider has gotten so much hype
(which
is why all the samples use it), is how it has been optimized. The SQL
Managed Provider talks directly to SQL Server without using OLEDB (the
benefits of using products from the same company). Microsoft claims that
the
speed of moving data between SQL Server and your ASP.NET application can
increase as much as 300% or more using the SQL Managed Provider because of
this direct communication.
I posted the results of a little test here back in June:
SQL Client (2.0): INSERT 20000 rows: 9,109375 seconds
SQL Client (2.0): 50 SELECT 20000 rows: 1,75 seconds
OLE DB (2000): INSERT 20000 rows: 20 seconds
OLE DB (2000): 50 SELECT 20000 rows: 62,140625 seconds
ODBC (2000): INSERT 20000 rows: 16,65625 seconds
ODBC (2000): 50 SELECT 20000 rows: 47,8125 seconds
OLE DB (2005): INSERT 20000 rows: 17,28125 seconds
OLE DB (2005): 50 SELECT 20000 rows: 61,484375 seconds
ODBC (2005): INSERT 20000 rows: 13,96875 seconds
ODBC (2005): 50 SELECT 20000 rows: 47,734375 seconds
<end>
Anything coming from OLE DB or ODBC is going to be slower because those
solutions are not in direct contact with SQL Server, as opposed to SQL
Client or something running within SQL Server QA with those solutions being
in direct contact with SQL Server.