This paper considers the performance of cached RAID5 using simulations that are driven by database I/O traces collected at customer sites. This is in contrast to previous performance studies using analytical modelling or random-number simulations. We studied issues of cache size, disk buffering, cache replacement policies, cache allocation policies, destage policies and striping. Our results indicate that: read caching has considerable value; a small amount of cache should be used for writes fast write logic can reduce disk utilization for writes by an order of magnitude; priority queueing should be supported at the disks; disk buffering prefetch should be used; for large caches, it pays to cache sequentially accessed blocks; RAID5 with cylinder striping is superior to parity striping.>

A method for managing a remote backup database to provide protection from disasters that destroy the primary database is presented. The method is general enough to accommodate the ARIES-type recovery and concurrency control methods as well as the methods used by other systems such as DB2, DL/I and IMS Fast Path. It provides high performance by exploiting parallelism and by reducing inputs and outputs using different means, like log analysis and choosing a different buffer management policy from the primary one. Techniques are proposed for checkpointing the state of the backup system so that recovery can be performed quickly in case the backup system fails, and for allowing new transaction activity to begin even as the backup is taking over a primary failure. Some performance measurements taken from a prototype are also presented.>