Cluster characteristics
When many OracleAS instances are grouped together according to identical configuration and application membership information, they form a cluster. OracleAS is run on such clusters in order to benefit from the following cluster characteristics:


Clustering enhances availability of the entire system. Applications that run in different instances process user requests amongst themselves. Because the load is shared between the redundant instances, it is possible to make effective use of all the resources, such as memory and processor speed. When a component fails, the cluster automatically restarts the failed component while other cluster components take over the failed component’s task. The high availability infrastructure also minimizes unplanned downtime due to system failure or human error.

Clustering enables easy management of the system. A farm is made up of multiple instances or a cluster of instances. Members of a farm cannot act as a single unit or have failover capabilities unless they are part of a cluster. Each farm instance shares the same repository. When an instance of the farm is reconfigured, the repository updates the other instances in the cluster. Uniformity is thus
maintained without having to reconfigure each instance manually. Because each instance may work on a different hardware configuration, not all instances share the same performance. It is important that the same operating system and application deployments are used when clustering instances across different machines. Some configurations may remain particular to an individual instance, for example host name, port number, and number of processes.

Clustering enables scalability of the number of instances. A cluster can span multiple hosts and distribute application operations across a scalable number of CPUs. A cluster can function as a single instance and each instance in a cluster can serve client requests. You can easily add or delete instances from a cluster. Scalability thus enables the system to share the processing load amongst its instances. It also affects performance by facilitating load balancing, which distributes requests amongst multiple instances.