I know this is not pure java but I worked a lot with Weblogic connection pools and I found they are very slow in comparison with the straight JDBC access (about 5-7 times slower). I was accessing both a DB2 and an Oracle database using the Weblogic connection pool (thru the tx data source name) for a brokerage application with tens of thousands of trasactions and part of the performance tunning I had to switch back to a home made connection pool. Maybe Jack or somebody else can explain to me why. Thanks.
0
