----------------------------------------------------------------------- WORKLOG TASK -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- TASK...........: Dynamique versionning of query plan for performance metric and downgrade . CREATION DATE..: Tue, 30 Jun 2009, 21:37 SUPERVISOR.....: Bothorsen IMPLEMENTOR....: COPIES TO......: CATEGORY.......: Server-RawIdeaBin TASK ID........: 33 (http://askmonty.org/worklog/?tid=33) VERSION........: WorkLog-3.4 STATUS.........: Un-Assigned PRIORITY.......: 30 WORKED HOURS...: 0 ESTIMATE.......: 0 (hours remain) ORIG. ESTIMATE.: 0 PROGRESS NOTES: DESCRIPTION: Just for comparing apple and oranges , A lot of internal SUN/ORACLE benchmarks are reporting performance improvements. But they are only tested on specific workload and predefined scenario like DBT2. MariaDb should provide dynamique QP and provide a ratio of efficiency in regard with the number of handler operations, each user would so on, be able to found out, if an improvement or a bug in the data acess path match is wokload . With such feature 5.0 to 5.1 would have found an inconsistant ratio of 2 to 1/1000 with a serious fluctuation on time depending on closing , reopening , reclosing and reopening bugs like http://bugs.mysql.com/bug.php?id=28404 ESTIMATED WORK TIME ESTIMATED COMPLETION DATE ----------------------------------------------------------------------- WorkLog (v3.5.9)