Basic HTML version of Foils prepared 7 October 97

Foil 24 Parallel Cache Management (PCM) and Distributed Lock Manager(DLM)

From Second Part of RDBMS Discussion ARL Database Tutorial -- February 98. by Gang Cheng Marek Podgorny Chao-Wei Ou
1 Concurrency control of buffer caches among instances
2 Maintain independent, yet consistent cache of blocks for all instances
3 Allow multiple nodes to simultaneous read/write to database (disks)
4 Maximizes concurrent data access and minimizes (disk) I/O activity
5 Block-level PCM locks, row-level locking preserved
6 A fault-tolerant DLM: if a node goes down, the DLMs will automatically recover and reconstruct the lock database on behalf of the failed DLM
in Table To:

© on Mon Oct 13 1997