No Slide Zone – Tom Kyte

Today was the first time to attend a session in the ‘No Slide Zone’ which is one of the special programs presented at Oracle Open World. During these sessions no slides can be used, in other words this is no typical conference track.

The keywords for a ‘No Slide Zone’-session: ‘discussion, interaction, whiteboarding and more’.

The session I had chosen, ‘Tom Kyte presents 10 new features for Oracle Database 11g’ … and it was fun ;o)

Using different paper notes he mentioned and discussed the 11 different features which are now made available in 11g and explained which problems they solved.

Every time he mentioned a feature and explained the usage and gain of it, he threw away the card which was a really fun way of presenting.

After he had mentioned the 11 new features, the people could ask some questions … and to make sure questions were asked … books regarding 11g were given to the first 12 people asking a question. Everyone in the room had a question … this is a reminder for myself for future presentations and sessions.

So let’s stop talking and give you those features wright ;o):

  • Virtual Columns: ability to use functions instead of creating a new column for calculations => the outcome is beter cardinality which means a beter execution plan
  • Compress Data & MetaData
  • Finer Grained Dependencies => if the signature of a procedure doesn’t change (the input/output parameters), the procedure won’t be invalidated. When columns are added or dropped from tables, no invalidation occurs, if specs are added to packages, no invalidation occurs
  • Compress data for transactional data not only for read-only data => compressioning will work for conventional paths as well (CRUD-methods) which will save space and audit retrieval will be faster
  • Flash Data Archive: enew background process is used, flashback-process which will use the flashback query for data retrieval
  • Partioning is extended and now enables: partionining by virtual columpns; reference-partioning which will give the ability to partition on master-data from a child-table. The child-table doesn’t need to be denormalized anymore, so the attribute of the master-table is available in the detail-table. The last partioning-feature added is interval-partioning which will give you the ability to performed sparsed partioning
  • Encrypted tablespace => store data on block, in the index in clear data and afterwards when the data is written in clear text, the tablespace block will be encrypted
  • Cache more stuff => statement results can be cached; server-result cache which means just-in-time materialized views are used (use the buffer cache or shared pool); pl/Sql function cache which will mark the function as result cache
  • Stand-by database => for logical standby’s datatype support has been extended and for physical standby’s, active standy is now fully supported
  • Real Application Testing => catch the workload of your db 10g release 1 db and run the same workload on your db release 2 database using Real Application Testing

It was an amazing session and these are amazing features which will improve the DB’s performance and maybe, somewhere in the very, very far future DBA’s aren’t needed anymore ;o)