Storage Clusters - Hitachi Universal Storage Platform V Reference Manual

Storage Clusters

Each controller consists of two redundant controller halves called storage
clusters. Each storage cluster contains all physical and logical elements (for
example, power supplies, front-end directors, back-end directors, cache,
control storage) needed to sustain processing within the storage system. Both
storage clusters should be connected to each host using an alternate path
scheme, so that if one storage cluster fails, the other storage cluster can
continue processing for the entire storage system.
The front-end and back-end directors are split between clusters to provide full
backup. Each storage cluster also contains a separate, duplicate copy of cache
and shared memory contents. In addition to the high-level redundancy that
this type of storage clustering provides, many of the individual components
within each storage cluster contain redundant circuits, paths, and/or
processors to allow the storage cluster to remain operational even with
multiple component failures. Each storage cluster is powered by its own set of
power supplies, which can provide power for the entire storage system in the
event of power supply failure. Because of this redundancy, the USP V/VM can
sustain the loss of multiple power supplies and still continue operation.
The redundancy and backup features of the USP V/VM eliminate all active
single points of failure, no matter how unlikely, to provide an additional level
of reliability and data availability.
2-4
Hitachi Universal Storage Platform V/VM User and Reference Guide
Architecture and Components