A cache on Integration Server occupies a portion of the heap in the Java Virtual Machine (JVM). This portion of a cache is referred to as on-heap cache. You can optionally extend a cache beyond the heap to the following locations, commonly referred to as tiers:

The way in which a cache is divided between on-heap memory and the other tiers is determined by configuration properties that you specify when you create the cache.

The tiered approach enables you to extend a cache beyond the size constraints imposed by the heap. Additionally, when you extend a cache to the Terracotta Server Array, the cache can be simultaneously shared by multiple Integration Servers.

Developers who develop services that use Terracotta Ehcache do not need to know whether the cache they are using resides only on the heap or whether it extends to one of the other tiers. They simply need to code their service to use a particular cache. Whether that cache extends to disk, BigMemory, or a Terracotta Server Array is determined by configuration properties that are assigned to the cache itself. Consequently, developers do not need to change their code when the caching requirements for their applications change.