Autonomic Index Management

This work presents the core ideas for an approach that enables index tuning with no human intervention at all. SQL statements submitted to the database are monitored and indices are created or dropped according to the benefits or overheads regarding the actual workload. We have implemented this strategy using a software agent architecture embedded in PostgreSQL. We discuss here some of the architecture core ideas and present initial experimental results.

• Type: Conference
• Year: 2005
• Where: ICAC
• Authors: Sérgio Lifschitz, Marcos Antonio Vaz Salles
• Keywords: Architecture Core Ideas, Autonomic Computing, Core Ideas, ICAC 2005, Software Agent Architecture