Intelligent Disk Subsystems
Hard disks and tapes are currently the most important media for the storage of data. When storage networks are introduced, the existing small storage devices are replaced by a few large storage systems (storage consolidation). For example, individual hard disks and small disk stacks are replaced by large disk subsystems that can store between a few hundred gigabytes and several ten petabytes of data, depending upon size. Furthermore, they have the advantage that functions such as high availability, high performance, instant copies and remote mirroring are available at a reasonable price even in the field of open systems (Unix, Windows, OS/400, Novell Netware, MacOS). The administration of a few large storage systems is significantly simpler, and thus cheaper, than the administration of many small disk stacks. However, the administrator must plan what he is doing more precisely when working with large disk subsystems. This chapter describes the functions of such modern disk subsystems.
This chapter begins with an overview of the internal structure of a disk subsystem (Section 2.1). We then go on to consider the hard disks used inside the system and the configuration options for the internal I/O channels (Section 2.2). The controller represents the control centre of a disk subsystem. Disk subsystems without controllers are called JBODs (Just a Bunch of Disks); JBODs provide only an enclosure and a common power supply for several hard disks (Section ...