What is High availability concepts?
OracleAS 10g provides high availability on all Oracle application tiers by integrating redundancy into the Application Server network infrastructure. A correctly designed redundant network infrastructure eliminates all single points of failure that could bring down an entire system.
Availability refers to the amount of time in which an application is available for use according to defined company and network performance indicators. System availability is a flexible concept, and not all applications need to have the same level of availability because not all systems in a network are required to run at full capacity throughout an entire day. The decision of when to run a particular application at a particular level of availability depends on the network installation and the company’s business goals.
Redundant components act as a backup system when the primary system fails. You create high availability by applying redundancy to every component from the network to the back-end database. When implementing redundancy, you should consider that the redundant components increase the availability of the system only if they start to function as soon as the primary system fails. A gap of time may exist between when the problem is detected and when it is responded to, causing downtime. Implementing high availability is more complex in the database tier than in the middle tier.
Clustering is the cornerstone of high availability and refers to the grouping together of components in such a way that they act as a unit. A cluster of applications is identified by its constituent components’ identical functionality and redundancy. This characteristic enables the components of a cluster to take over the functions of a failed component. Also, you can scale a cluster by adding components to it.
High availability functions according to the principles of hardware clusters and failovers. Hardware clusters refer to a collection of computers – also called nodes – that become a cluster because they share the same configurations. Each node functions independently and runs its own processes, but all nodes in the cluster will respond to a request when a cluster component fails. A Load Balancer distributes requests only to nodes that are functioning and are not overloaded.
Failover refers to the automatic transfer of the function from a failed node to a redundant component. Clients may need to reconnect to the application after the failover. However, the client will not know which server is actually providing them with the application.
A hardware cluster receives either an active-passive or active-active set-up. In the former instance, the cluster contains an active server that runs its own applications, and a passive server that runs none of its own applications. The passive server acts as a redundant component whose sole duty it is to become active when the Load Balancer assigns the redundant server failover duty. In the active-active instance, all servers in one cluster perform their own applications, but reserve space for failover duty.
