Pool Allocation [Noble 2001] describes similar concepts in terms of memory allocation.
Connection pools and statement caches are pervasive data access optimizations. Many database drivers and most application servers provide some form of connection pool support. The JDBC 2.0 standard provides many support classes similar to those that this pattern describes.
You can pool logical resources, such as those that specialize Data Accessor (9). This can improve resource utilization and also helps to manage multiple resources within a central component.
PooledResource specializes Resource Decorator (103), since it decorates any resource with customized close behavior.
Resource Timer (137) describes a strategy for cleaning up inactive resources. In one capacity, you can use it to counter defective clients that neglect to return inactive resources to the pool. In another, you can automatically trim the size of the pool when a large number of its resources go unused.
A common problem in connection pools is that some databases implicitly close connections after given inactivity conditions are met. When a connection pool attempts to recycle a closed connection, unpredictable results often occur. Retryer (171) describes a strategy that can be used to transparently detect and discard stale connections and then replace them with fresh connections.
Cache Statistics (361) describes a consistent mechanism for publishing resource pool statistics such as pool size and number of hits and misses. A hit happens when the resource pool recycles an existing pooled resource. A miss indicates that the pool cannot find a resource to match a request and needs to create a brand new resource.
