Oracle Multiple data cache buffers Oracle Tips by Burleson Consulting

While multiple blocksizes are new to Oracle, I?ve been using them since the 1980?s (In IDMS) where the benefit of buffer segregation and I/O reduction are well-understood.  If you are planning to try-out multiple blocksizes, first make sure that your database is using direct I/O.
 

From the research of author Robin Schumacher, we now know that moving Oracle indexes to a larger blocksize can make more efficient use of limited RAM regions (db_cache_size, db_32k_cache_size, etc.) and the intelligent segregation of objects to reduce logical I/O (consistent gets) for multi-block scan reads. 

Because the blocksize affects the number of keys within each index block, it follows that the blocksize will have an effect on the structure of the index tree. All else being equal, large 32k blocksizes will have more keys per block, resulting in a flatter index than the same index created in a 2k tablespace.

Today, most Oracle tuning experts utilize the multiple blocksize feature of Oracle because it provides buffer segregation and the ability to place objects with the most appropriate blocksize to reduce buffer waste. Some of the world record Oracle benchmarks use very large data buffers and multiple blocksizes.

According to an article by Christopher Foot, author of the OCP Instructors Guide for Oracle DBA Certification, larger block sizes can help in certain situations:

"A bigger block size means more space for key storage in the branch nodes of B-tree indexes, which reduces index height and improves the performance of indexed queries."

In any case, there appears to be evidence that block size affects the tree structure, which supports the argument that data blocks affect the structure of the tree.

Today, MOSC notes that the multiple blocksize parameters are among the most important in Oracle tuning, and noted experts such as Robin Schumacher has demonstrated that Oracle indexes will build more-optimal b-tree structures within a large blocksize. We also see end-user reports showing I/O reductions of 20% with this simple ?Oracle Silver Bullet?.

 

 

Here is one real-world comment about index segregation into a larger blocksize:

 

"My favourite recent article was on 32KB indexes - Our client (200GB+) saw a 20% reduction in I/O from this simple change... " 
 
 The server had 8 CPUs 32gb RAM with db_cache_size = 3g and db_32k_cache_size = 500mb. The database was over 200 gigabytes.
 
 Steve Taylor
 Technical Services Manager EMEA
 Eagle Investment Systems Corp.

 

However, widespread acceptance of using multiple blocksizes has been hindered because the I/O reduction cannot be ?proven? using simple SQL*Plus scripts on personal computers and because multiple blocksizes were originally created to support transportable tablespaces:

 

For example, some Research DBA?s attempt to prove the Rule-of-thumb that ?Indexes run better in a large blocksize? using small, artificial single-user experiments.  They falsely suggest that multiple blocksizes are unlikely to help in a real-world database.  For example:

 

Minimal difference in elapsed time. . . . So the number of waits has halved, but performance is not much changed. Queries against both tablespaces are performing pretty much the same as before.

 

Some Oracle researchers say that using multiple blocksizes is ?Voodoo? tuning, for some interesting reasons:

 

1 - Because multiple blocksizes were never intended as a performance tool.
 

2 - Because they cannot replicate the performance benefit on a personal computer.

 

And bear in mind I stress that you should TEST the use of multiple block sizes. They were actually only invented to make transportable tablespaces work properly. So you are pushing them into realms they were never really intended for if you use them as a performance tuning tool. And that's despite one well-known leading author claiming all indexes should be rebuilt into 32K block size tablespace (a lovely example of totally Voodoo Tuning advice)