Comments on: Oracle Identity Management Event http://www.alwynvanniekerk.com/2008/oracle-identity-management-event/ Sat, 27 Dec 2008 06:35:22 -0500 http://wordpress.org/?v=2.8.1 hourly 1 By: Alwyn http://www.alwynvanniekerk.com/2008/oracle-identity-management-event/comment-page-1/#comment-41 Alwyn Tue, 29 Apr 2008 06:48:50 +0000 http://www.alwynvanniekerk.com/?p=28#comment-41 There's currently no serious consideration to toss our Novell implementation, but the (Oracle) DBA's do like to joke about it... Our DB is my authoritative source for the majority of our identity attributes, and the data structures used to extract the desired information in one go is pretty heavy, albeit well optimized. It's a balancing act between expected service levels and system impact on overall performance - but that is the price you pay for having all your information originate in external systems. There’s currently no serious consideration to toss our Novell implementation, but the (Oracle) DBA’s do like to joke about it…

Our DB is my authoritative source for the majority of our identity attributes, and the data structures used to extract the desired information in one go is pretty heavy, albeit well optimized.

It’s a balancing act between expected service levels and system impact on overall performance – but that is the price you pay for having all your information originate in external systems.

]]> By: Dominic White http://www.alwynvanniekerk.com/2008/oracle-identity-management-event/comment-page-1/#comment-40 Dominic White Tue, 29 Apr 2008 06:40:36 +0000 http://www.alwynvanniekerk.com/?p=28#comment-40 Hmm, changing IAM products based on DB performance seems strange. Have the DBA's fiddled with the indexing, caching and connection pooling settings first? Even then, there are other options such as upping the DB HW specs or replicating identity info to a separate DB dedicated to IAM. All of the above would be much cheaper than a new IAM implementation. Hmm, changing IAM products based on DB performance seems strange. Have the DBA’s fiddled with the indexing, caching and connection pooling settings first? Even then, there are other options such as upping the DB HW specs or replicating identity info to a separate DB dedicated to IAM. All of the above would be much cheaper than a new IAM implementation.

]]>