In-memory databases are able to store and react to changes in your data an order of magnitude faster than even databases that store data on solid state drives, allowing single thread benchmark speeds of over 140,000 storage operations in under 1 millisecond, and the same number of reads in under 0.1 millisecond on consumer grade hardware. This raw speed advantage enables in-memory databases to address a different set of use cases like intelligent caching, accelerated user experiences, and adaptive traffic shaping.
This white paper first analyzes the business value derived from low latency applications, it then examines some architectural patterns for in-memory databases, and finally looks at the leading in-memory database and our recommendations.