A properly implemented end-to-end high availability solution enables you to reduce the possibility that the entire system becomes vulnerable due to a single point of failure. OracleAS is constructed on a grid to ensure high availability.
End-to-end high availability solution
The end-to-end high availability solution begins when the Web Cache or the Oracle HTTP Server (OHS) receives a request from the client. At the Web Cache tier, multiple Web Cache instances are clustered in order to share cached information, support the failover feature, and conduct load balancing to multiple middle tier Application Server instances.
At the middle tier, multiple Application Server instances are clustered in order to implement OHS and multiple Oracle Containers for Java (OC4J) that run multiple copies of the application.
The Infrastructure tier contains the Infrastructure instance and the repository database.
The Application Servers are followed by a continuous availability database that uses Real Application Cluster (RAC), Data Guard, or replication. RAC creates multiple database instances of the same database. Data Guard implements a standby database that is constantly updated from the active database. Replication implements multiple active and synchronized databases.
Grid computing makes use of available processing power amongst all network resources. The grid groups together large numbers of servers and storage capacity. In this way, it supports clustering and offers an end-to-end high availability solution.
OracleAS can be efficiently deployed on the grid using the following features:
- Distributed Configuration Management (DCM), which prepares the Application Server instances for use on the grid
- Oracle Enterprise Manager Grid Control (Grid Control), which is the interface with which you manage the Oracle Application Server grid environment
- Process Management, which adds Oracle Application Servers according to scalability and availability requirements
- Workload Management and Load Balancing, which detect and route requests among a cluster to share the load and circumvent failed cluster components
The major challenge when implementing high availability is to design a grid infrastructure that reduces planned downtime and keeps unplanned downtime to a minimum. Planned downtime includes the times when you conduct system and data maintenance. You reduce planned downtime with the help of rolling upgrades. The rolling upgrade enables you to upgrade and test components of the Application Server one by one while redundant components support the application. Unplanned downtime includes instances that are caused by system failure, data failure, and human error. You reduce such downtime with the aid of the following OracleAS features:
Backup and recovery. The backup and recovery tool enables you to recover – rather than repair – failed components by creating a checkpoint to which you recover. It also replicates data from the production site onto the standby site.
Disaster recovery. The OracleAS disaster recovery tool provides a standby site with a replication of the active production site’s Application Server. This becomes a necessary disaster recovery strategy for when the active production site is destroyed.
Distributed Configuration Management (DCM). DCM prepares the Application Server instances for use on the grid. DCM enables you to prevent human errors with its configuration management tool.
Identity Management (IM) Services Replication. IM supplies Oracle Internet Directory (OID) and Single Sign-On (SSO) to the user authentication and authorization features. The IM middle tier is installed on multiple machines. If one fails, the IM middle tier is automatically routed to an operational machine.
Infrastructure Active Failover Cluster. You can configure the OracleAS Infrastructure within an Active Failover Cluster. OracleAS Infrastructure runs an active-active setup. When one node fails, the others are ready to take over, because all nodes are active.
Infrastructure Cold Failover Cluster. You can configure the Infrastructure tier within a Cold Failover Cluster. OracleAS Infrastructure runs an active-passive setup whereby the passive node – situated at the standby site – takes on failover duty only when an active node at the production site has failed.
Java Naming and Directory Interface (JNDI). JNDI is an Application Programming Interface (API) that assists with the interfacing of multiple naming and directory services.
Load balancing. Load balancing facilities detect and route requests among a cluster of OracleAS instances to ensure all capacities are used effectively.
Oracle Process Monitor and Notification Server (OPMN). OPMN offers the Application Server high availability options because it manages all Infrastructure processes in OracleAS – with the exception of the database and its listener. OPMN monitors and deals internally with any failure of OPMN-managed processes.
High availability works on the principle that each component of Application Server can create redundancy. You need to ensure that you create a high availability Infrastructure by using high availability features that are built into the Application Server 10g and the RAC Database 10g.
As a summary of high availability in OracleAS. High availability works on the principle that each component of Application Server can create redundancy. You need to ensure that you create a high availability Infrastructure by using high availability features that are built into the Application Server 10g and the RAC Database 10g.
