Although OPS was good for its time, it had several limitations, especially with scalability. In addition, it required careful setup, configuration, and a heavy amount of administration. It also required raw devices for storage, which were not easy for inexperienced system administrators to set up and administer.
Dynamic lock remastering allows frequently used database objects to be managed in the local instance, which greatly reduces the interconnect traffic and increases performance.
Oracle RAC also comes with an integrated clusterware and storage management framework, removing the dependency on the vendor clusterware and providing ultimate scalability and availability for database applications.
Oracle RAC systems can be configured to have no single point of failure, even when running on low-cost, commodity hardware and storage. If a monolithic server fails for any reason, the applications and databases running on the server are completely unavailable. With Oracle RAC, if any of the database servers fail, applications simply keep running and the services are continuously available through the remaining servers. The underlying clusterware makes sure the services are relocated to other running servers, and failover is frequently transparent to applications and occurs in seconds.
Oracle RAC has proven time and again in all “near death” situations that it can save the day with its controlled reliability. The paradigm has shifted from reliability to maximum reliability. Analyzing and identifying the subsystem and system failure rates in accordance with the appropriate standard ensures reliability, which is a key component in the Oracle RAC technology.
Oracle RAC allows multiple servers in a cluster to manage a single database transparently and without any interruptions to database availability. Additional nodes can be added when the database is up and running. With dynamic node addition and effective workload management, Oracle RAC allows database systems to scale out in both directions, with an ability to run an increased workload by increasing the number of nodes as well as the ability to reduce the amount of time to complete a workload. This means that the previous ceilings on scalability have been removed. Collections of servers can work together transparently to manage a single database with linear scalability.
Oracle RAC requires no changes to existing database applications. A clustered Oracle RAC database appears to applications as a traditional single-instance database environment. As a result, customers can easily migrate from single-instance configurations to Oracle RAC without needing to change their applications. Oracle RAC also requires no changes to existing database schemas.
Oracle RAC allows organizations to use collections of low-cost computers to manage large databases rather than needing to purchase a single large, expensive computer. Clusters of small, commodity-class servers can now satisfy any database workload. For example, a customer needing to manage a large Oracle database might choose to purchase a cluster of eight industry-standard servers with four CPUs each, rather than buying a single server with 32 CPUs.
Oracle RAC allows clusters of low-cost, industry-standard servers running Linux to scale to meet workloads that previously demanded the use of a single, larger, more expensive computer. In the scale-up model of computing, sophisticated hardware and operating systems are needed to deliver scalability and availability to business applications. But with Oracle RAC, scalability and availability are achieved through functionality delivered by Oracle above and beyond the operating system.
Clustering is a great solution for scaling up and speeding up the database workload processing with extreme availability. Because the storage is shared among the nodes, all the members of a cluster can concurrently access the storage, which helps the system be available during node failures. Oracle databases implement the shared everything architecture to achieve on-demand scalability and ultimate availability.
The right clustering technology can achieve infinite scalability on demand and uninterrupted availability. Oracle databases have a good track record of clusters with OPS and its successor, Oracle RAC. The limitations of OPS are addressed in RAC, with the introduction of the Cache Fusion framework. Later versions further improved Cache Fusion with the introduction of the read-only and read-mostly locking frameworks.
With the introduction of the Grid Infrastructure, Oracle integrated Automatic Storage Management, a true volume manager and file system for the Oracle database infrastructure, with its clustered database solution. This eliminated the requirement of having a third-party file system manager for the Oracle infrastructure. Oracle RAC 12c provides impressive scalability, availability, and flexibility at a low cost. It makes the consolidation of databases affordable and reliable by leveraging scale across the architecture.