The ActiveSpaces® Transactions High Availability feature provides an easy way to ensure system availability using replicated objects. Two or more nodes (generally on different hosts to reduce risk) are linked together with the high availability feature. The application objects are contained in a high availability partition, which has a prioritized list of nodes that host the partition.

	Multiple partitions can be specified; the node lists can differ for each partition.

Normally the highest priority node in a partition's node list is the active node, and it processes work. Figure 2.11, “High availability with node X active” shows a client directing traffic for processing to node X, which is the active node for the partition. As the requests are processed, objects are transactionally replicated onto node Y.

Figure 2.11. High availability with node X active

If a node goes out of service for some reason, the next highest priority node in the node list for the partition becomes active and all processing can continue uninterrupted on that node. Figure 2.12, “High availability with Y active” shows this taking place. When node X is later brought back up and restored, it will again be the active node and the data and processing will move back to it from node Y.

Figure 2.12. High availability with Y active

It is possible to have all partitions active on one node and all the replica nodes for the partitions on the other; an alternative approach provides load balancing by allocating the active nodes for partitions across several nodes. It is also possible, using ActiveSpaces® Transactions management tools, to migrate a partition from one node to another. This can be used to scale up a system by adding nodes, or to manage processing load by moving hot partitions to dedicated nodes.