Dynamic storage blocks tiering

摘要

Storage management for database access, in which selected data blocks are associated with particular business operations, and in which those data blocks are disposed in response to an assessment of which of those business operations are more important, more urgent, or have a greater need to be reliable. This has the effect that overall transaction performance is improved without significantly increasing infrastructure cost. Noting the relative importance of those particular business transactions and the relative probability that those data blocks will be needed by those particular business transactions, and assuring that data blocks associated with relatively important business transactions are moved to relatively superior storage in advance of requests for access.

主权项

1. A method of dynamic storage blocks tiering including steps of:
at a database server capable of individually accessing particular data blocks, the data blocks being associated with database structures accessible by the database server, the server coupled to at least a first storage device and a second storage device, the storage devices having different performance parameters; associating the particular data blocks with one or more business transactions, wherein each of the one or more business transactions is of a different type and distinct from another business transaction; maintaining a set of information for each of the particular data blocks, including at least:
an identifier for identifying the data block;a rank that indicates a significance of a particular business transaction with which the data block is associated;a measure of usage time likely to be expended operating with the data block;a type of access indicating whether access to the particular block would be random-access or sequential-access;a probability measured in time units indicating how soon the data block will be accessed; determining an anticipated order of the one or more business transactions based on the step of maintaining; and repositioning at least some of the data blocks between the at least first and second storage devices in response to the step of determining, wherein the repositioning includes moving one or more of the data blocks, less than an entire of the database structure, from a first particular storage device having a relatively inferior performance parameter to a second particular the storage device having a relatively superior performance parameter.