Home > Workload Solutions > Oracle > White Papers > Oracle TimesTen In-Memory Database on Dell S5000 Series Servers > Business challenge
Businesses require real-time information, meaning they must currently know what is happening with their business operations. For many businesses, real-time applications are not elective but a requirement to stay competitive in their markets.
Companies such as telecom operators, security exchanges, airlines, shipping, and logistics, to name a few, are prime examples of enterprises that require real-time applications to process, analyze, and respond intelligently to key events to make sound decisions.
In a traditional relational database, user applications communicate with a database server process over an IPC connection which adds substantial performance overhead to all SQL operations. Furthermore, traditional disk-based databases have been primarily optimized under the assumption that the data is disk resident. Optimization algorithms, buffer pool management, and index retrieval techniques are designed based on this fundamental assumption, leading to slower response times.
Even when a traditional relational database has been configured to use its main memory to hold all its data, the performance is still not optimized because the database engine has been engineered for disk-based residency. It is not easy to re-engineer due to decades of research and development within traditional database processing logic, indexing schemes, and data access mechanisms.
Furthermore, conventional disk-based databases rely heavily on the disk I/O subsystem to be optimized. Even though disk technology has improved from spinning spindles to solid-state technology, database I/O operations still need to traverse the storage network from media to the CPU adding latency to the applications.