JCS, or the Java Caching System, is a relatively new entrant into the growing family of distributed Java caches. Even so, it has been in development since 2001, according to the project's home page, which announced the release of JCS 1.3 this week.

The most important use-case for JCS is to speed up, and scale up, Web applications:

JCS is a front-tier cache that can be configured to maintain consistency across multiple servers by using a centralized remote server or by lateral distribution of cache updates. Other caches, like the Javlin EJB data cache, are basically in-memory databases that sit between your EJB's and your database. Rather than trying to speed up your slow EJB's, you can avoid most of the network traffic and the complexity by implementing JCS front-tier caching. Centralize your EJB access or your JDBC data access into local managers and perform the caching there.

The project documentation also highlights that JCS can "scale into the tens of servers."

JCS is based on what the project describes as a "composite cache," potentially consisting of four different types of cache implementations:

Four types of caches can be plugged into the Composite Cache for any given region: (1) Memory, (2) Disk, (3) Lateral, and (4) Remote. The Composite Cache orchestrates access to the various caches configured for use in a region.

The JCS distribution provides an implementation of each. Its lateral cache, for instance, uses UDP to discover other cache nodes and to distribute data to available nodes. The disk cache implementation, by contrast,

Follows the fastest pattern for disk swapping. Cache elements are written to disk via a continuous queue-based process. The length of the item is stored in the first few bytes of the entry. The offset is stored in memory and can be reference via the key. When items are removed from the disk cache, the location and size are recorded and reused when possible. Every aspect of the disk cache is configurable, and a thread pool can be used to reduce the number of queue worker threads across the system.

Have you used JCS in your projects? If so, what experiences with this cache can you share?