Hotsos Symposium Speaker – Riyaj Shamsudeen

Riyaj Shamsudeen is a returning Hotsos Symposium presenter.


Riyaj ShamsudeenRiyaj Shamsudeen has 17+ years of experience in Oracle and 16+ years as an Oracle DBA/Oracle Applications DBA. He is the principal DBA for OraInternals, a consulting company resolving many advanced performance, recovery, RAC, and EBS11i issues. Riyaj specializes in RAC and performance issues and has authored many articles, mostly about performance, database internals, optimizer internals, etc. He is a frequent presenter in major conferences such as Hotsos Symposia, UKOUG, RMOUG etc. He is a proud member of OakTable network. He blogs regularly at

Presentation Title

Session #1: Why Does the Optimizer Hate My SQL?
Session #2: A Close Encounter with Real World (and Odd) Performance Issues


Session #1, Why Does the Optimizer Hate My SQL?: This presentation explores the various SQL performance issues encountered in the real world and discusses root causes for those issues:

  • Review of cardinality, selectivity and histograms. Every performance tuner must completely understand these concepts.
  • What to look for in a 10053 trace. How to effectively use 10053 trace files to identify root cause of a SQL performance issue.
  • Bind peeking issues.
  • Effective use of system statistics to combat optimizer discrepancies.
  • Index efficiency, indices on non-selective columns.
  • Nulls and indices.
  • Issues with a few hints such as ORDERED, INDEX etc. Use of profiles and outlines to resolve these issues.
  • Subquery transformation and a few optimizer features related to sub query transformation.
  • Predicate based first_k_rows mode.

Session #2, A Close Encounter with Real World (and Odd) Performance Issues: This presentation explores various customer performance issues encountered in real world performance situations, including odd performance issues. This presentation uses scientific methods to show that even odd problems can be resolved using scientific techniques:

  • A case of high Kernel mode CPU usage — High kernel mode CPU usage in a database node caused ASM timeout issues leading to downtime issues. Why?
  • Cluster hang situation due to enqueues.
  • Shared pool fragmentation, duration, and more.
  • A case of heavy latch contention and x$ table at the heart of it.
  • A few more if time permits.

Presentation Materials

Presentation materials are available online to attendees only.


Riyaj ShamsudeenThe speaker schedule is as follows: