High availability
High availability refers to a minimal loss of service to critical applications, such that planned and unplanned outages have minimal impact on client request processing. In a highly available environment, your enterprise application continues to process client requests efficiently in the event that part of the system fails. To ensure high availability, create a multiple machine configuration in which two or more physical machines host multiple application server processes. It is recommended that you also use iSeries logical partitioning (LPAR) to enhance availability.
The number of machines required to ensure high availability depends on the size of the workload, the platform of operation, and the size of the machines' processors and associated memory. In most cases two iSeries servers are adequate to handle a workload in a highly available and scalable manner. You can add processors or application servers as needed to handle increases in workload.
LPAR on iSeries enhances availability by providing additional process and hardware isolation. Process and hardware isolation improves security, failover support, and manageability (server consolidation). However, the use of LPAR on a single machine does not guarantee high availability, because the machine is still a single point of failure in the configuration. In the iSeries Information Center, see the Logical partitions topic in Systems management:
These topics provide more information on high availability:
Benefits of high availability configurations
High availability configurations provide several benefits,Manage high availability configurations
To adminster high availablilty configurations, use the WebSphere Application Server Network Deployment deployment manager. This topic describes the resources managed by the Network Deployment systems administration tools.High availability topologies
Several configuration options are available, each with its own advantages. You can choose a single topology or combine topologies to optimize client request processing. See this topic for samples of the available topologies.Workload management
Use workload management (WLM) to distribute client requests among multiple application servers according to their capabilities. This topic describes WLM and how it can help you improve the availability and performance of critical applications.Clusters and cluster members
In WebSphere Application Server, a cluster is a group of application servers that work together under workload management. The application servers in a cluster host identical copies of an application and process client requests for that application. This topic provides information about clustering in WebSphere Application Server.