nodes. In addition, there is a disk subsystem that is directly linked to multiple computers
within the cluster. In this figure, the common disk subsystem is a RAID system.
The use of RAID or some similar redundant disk technology is common in
clusters so that the high availability achieved by the presence of multiple computers
is not compromised by a shared disk that is a single point of failure.
A clearer picture of the range of cluster options can be gained by looking at
functional alternatives. Table 17.2 provides a useful classification along functional
lines, which we now discuss.
A common, older method, known as passive standby, is simply to have one
computer handle all of the processing load while the other computer remains inactive,
standing by to take over in the event of a failure of the primary.To coordinate
the machines, the active, or primary, system periodically sends a “heartbeat” message
to the standby machine. Should these messages stop arriving, the standby
assumes that the primary server has failed and puts itself into operation. This approach
increases availability but does not improve performance. Further, if the only
information that is exchanged between the two systems is a heartbeat message, and