HIPLEX is the concept which allows several hosts to be grouped together in a computer network. This means that users can address their requests to a computer network and not have to send them to one of the hosts. The advantage of this concept is that continuous request processing is guaranteed even if one or more hosts of the network are unavailable. Files that have been backed up or archived on one host, can be restored on another host in the network if that host crashes. When the crashed host is up and running again, the situation can be returned to the pre-crash status while the continuity of HSMS activities is guaranteed.
To be able to move HSMS activities from one host in the network to another without problems, the environment has to be adjusted. The data and metadata available on the first host must be made available on the substitute host. General availability within the network is achieved by means of SM pubsets. If the data and metadata is located on an SM pubset, the HSMS activities can be continued after a host crashes by importing the SM pubset on a substitute host (see also section "HSMS in an SM pubset environment").
When the HSMS administrator issues the HSMS statement RECOVER-REQUESTS, all requests that were issued on the crashed host are integrated in the local requests. The SHOW-REQUESTS statement can be used to output the status of the requests at the moment the host crashed. Requests that had not been started on the crashed host can be processed on the substitute host. Requests that were running just as the host crashed and had not been interrupted can be restarted on the substitute host.
For the user, the host changeover is transparent: the user starts off working with an SM pubset and a number of archives on that SM pubset. After a host crash, the user’s requests are processed by another, the host itself not needing to perform any actions.