High Availability (BS2000)
High availability and scalability for the entire IT infrastructure
A cluster consists of a number of server systems linked together to perform as a single system. This ensures scalability and the high availability of applications or services in the event of failures or planned downtime.
HIPLEX (highly integrated system complex) is the Fujitsu concept for supporting a load and availability-sharing cluster consisting of several BS2000/OSD business servers. (We also call each multi-server configuration having the attributes of one of the aforementioned cluster types a HIPLEX.)
A HIPLEX consists of several uniprocessors or multiprocessors linked via a network. Each one of these processor systems has a BS2000/OSD operating system of its own or runs with VM2000. This HIPLEX setup offers a wider range of scalability and redundancy than traditional uniprocessor or multiprocessor systems with shared memory architecture. (Symmetric Multiprocessor = SMP).
HIPLEX AF and HIPLEX MSCF enable applications to be implemented with the highest possible availability as part of a multi-server concept:
- If one system fails, the applications monitored by HIPLEX AF are automatically switched to a second intact system (hot standby) together with their resources. Downtime is minimized by an automatic failure detection facility as well as the fact that a system restart is not required and thus does not influence application availability. There is also a warm standby configuration option by which the failed system is automatically restarted on a second server.
- In addition to managing system failure, HIPLEX AF can increase availability after an application failure and provide the basis for a load-sharing cluster. Here, specific applications can be relocated to another server with the aid of HIPLEX AF.
- In case of a catastrophe, such as a local crash at a site of both the server or servers and the (mirrored) storage subsystem, HIPLEX AF autonomously switches over to a remote data center (failover) without any data loss. Support for transferring applications back to the recovered data center (failback) is available at the push of a button.
Enables applications to be implemented with highest possible availability as part of a multi-server concept.
This is the cluster concept to support an availability, operational and load network with several BS2000/OSD Business Servers.
Offers the option of mirroring disk drives in the BS2000 via doubled I/0 events