A read cache typically retains copies of the data in fast I/O memory storage such as dynamic random access memory (DRAM) or flash. The main distinguishing characteristic of a read cache is the mechanism by which the data populates the cache. Data enters the cache after it has been retrieved, or read, at least once from its permanent storage hard disk drive (HDD) or solid-state drive (SSD). An algorithm tracks the data reads and determines which data will be deposited in the read cache.
As the cache receives more data, or warms up, more data requests can be accelerated. In a typical scenario, the algorithm orchestrates the placement of copies of the most frequently accessed data into the read cache. A data request that is fulfilled from the read cache is faster than a request fulfilled from the original storage location.
Read caches may operate differently based on the algorithm variant they use. A read cache can contain multiple layers of storage technology, such as DRAM and SSD paired to accelerate data requests to an HDD-based array. Another variable is the implementation location, such as server-side or storage-side.