||Karl Arao is an OCP-DBA, RHCE, and Oracle ACE. He works for Enkitec as a senior technical consultant where he currently spends most of his time exploring on Exadata capacity planning and performance. Prior to that, Karl went to work for SQL*Wizard as a solutions architect and an R&D guy where he had his five years of solid DBA experience including proactive database maintenance, database installation and configuration for high availability and disaster recovery, migration and upgrades, backup and recovery, database hardening, capacity planning, and performance tuning. He shares his experiences, adventures, and discoveries in his blog (karlarao.wordpress.com), he tweets at @karlarao, and owns a Wiki site (karlarao.tiddlyspot.com) where he shares his quick guides and documentations on technologies. He presents at local Oracle Events and DBA round tables, and has also presented at the Oracle OpenWorld Unconference, Oracle Closed World, Virtathon, and Hotsos. He is also a proud member of the Philippine Software Industry Association Technology Council. In his spare time, he loves to mine the ASH and AWR and do adventure sports like underwater hockey, mountaineering, mountain biking, aquathlon, and sprint distance triathlon.
||New CPUs and storage arrays are getting faster, but these resources are finite and come at a cost. Hence, capacity planning plays a very important role to ensure proper resources are available and to handle expected and unexpected workloads. Another critical matter for the DBAs and IT managers is justifying the expense of adding resources on the system. With guesswork, you'll end up getting the most expensive hardware. With proper measurement, proper planning, and management of growth, you'll be able to get just the right hardware for your workload with allowance for a particular growth period. This will result in huge savings for the company and a happier IT shop. AWR is a built-in data store that started in 10gR1 and is very much like a "Statspack on steroids." It has improved significantly in 11gR2, enabling you to have a far better workload information when going through all the AWR snapshots. From the AWR data samples, we could build amazing reports that will let us notice trends and makes it possible to visualize data and use statistical methods for analysis. Even more surprising about the AWR data samples is we are able to define the database server's Capacity, Requirements, and Utilization in terms of CPU, IO, memory, and network, which are very important key metrics for Capacity Planning.