In this presentation, Hal takes a deep dive technical approach to modern day, efficient, physical DB2 Index design. He will review the differences in index types: clustered and non-clustered, partitioned and non-partitioned, UNIQUE and non-UNIQUE.
He will explain the costs and benefits of important index features such as Index Compression, the effect of PADDED vs. NOT PADDED, NPSI vs. DPSI, PIECESIZE, and others, with a focus on best practices. Hal will also offer some modern ideas about Index Page size and index buffer pools.
In particular, Hal will look at partitioned indexes contrasting classic vs. UTS/PBR and make several recommendations, sharing benchmark results from his own experience. The session provides several opportunities for Q & A plus general comments from the audience.