Pooling Connection Object

[Ryan P. Kennedy]

Hello:

How does one pool a connection object? I have the same application running
on 4 different machines, all connecting to the same server/SQL Server 2000
instance for DB activity.

Some posts have mentioned pooling the connection objects to reduce overhead,
but how do I do that for the 4 separate computers.

Appreciate any response.

Regards,
Ryan Kennedy

[Joe Weinstein]

Ryan P. Kennedy wrote:

Hello:

How does one pool a connection object? I have the same application running
on 4 different machines, all connecting to the same server/SQL Server 2000
instance for DB activity.

Some posts have mentioned pooling the connection objects to reduce overhead,
but how do I do that for the 4 separate computers.

Appreciate any response.

You would want to connect your four applications to a single middle tier
process, which remained running, and which maintained one or more open
connections to the DBMS. The middle tier process would act as a proxy for
your client applications, receiving requests, and sending them to the
DBMS via a perpetually-kept connection, and then relaying the returns from
the DBMS to the clients.
A pooled connection can only live as long as the client that made it.
There are two main savings from pooled connections: They save the typically
longer time it takes to make a new connection, and they can be used to
limit the number of simultaneously open DBMS connections, in cases where
licensing or performance dictate that you cannot make a separate connection
for each client.
Joe Weinstein at BEA Regards,
Ryan Kennedy