* Tune physical IO SQL. As I mentioned, there can be no significant LRU chain latch contention without physical IO. So, from an application perspective, look for the SQL statements that perform the majority of block reads; that is, physical IO activity. Do whatever you can to reduce their physical IO consumption. This means classic SQL tuning, including indexing, as well as reducing the execution rate of the top physical IO SQL during the performance-critical times.

* Increase CPU power. Just as with CBC latch contention, or any other latch contention for that matter, if more CPU power is available, memory management will take less time. This means latch hold time and also latching acquisition time (both spinning and sleeping) could be reduced. Increasing CPU power also means looking for creative ways to reduce CPU consumption during peak contention times.

* Increase LRU latches. By default, most Oracle systems have only 8 to 32 LRU chain latches. LRU concurrency can be increased by adding latches, which means increasing the hidden parameter _db_block_lru_latches. This can be especially effective if you have a multigigabyte buffer cache.