A high performance thread-safe map that expires entries
Java Shell

README.md

ExpiringMap

Build Status Maven Central License JavaDoc

A high performance, low-overhead, zero dependency, thread-safe ConcurrentMap implementation that expires entries. Features include:

Supports Java 6+ though the documentation uses lambdas for simplicity.

Usage

ExpiringMap allows you to create a map that expires entries after a certain time period:

Map<String, Connection> map = ExpiringMap.builder()
  .expiration(30, TimeUnit.SECONDS)
  .build();

// Expires after 30 seconds
map.put("connection", connection);

Expiration Policies

Expiration can occur based on an entry's creation time or last access time:

Map<String, Connection> map = ExpiringMap.builder()
  .expirationPolicy(ExpirationPolicy.ACCESSED)
  .build(); 

We can also specify an expiration policy for individual entries:

map.put("connection", connection, ExpirationPolicy.CREATED);

And we can change policies on the fly:

map.setExpirationPolicy("connection", ExpirationPolicy.ACCESSED);

Variable Expiration

Entries can have individually variable expiration times and policies:

ExpiringMap<String, Connection> map = ExpiringMap.builder()
  .variableExpiration()
  .build();

map.put("connection", connection, ExpirationPolicy.ACCESSED, 5, TimeUnit.MINUTES);

Expiration times and policies can also be changed on the fly:

map.setExpiration(connection, 5, TimeUnit.MINUTES);
map.setExpirationPolicy(connection, ExpirationPolicy.ACCESSED);

Expiration Listeners

Expiration listeners can be notified when an entry expires:

Map<String, Connection> map = ExpiringMap.builder()
  .expirationListener((key, connection) -> connection.close())
  .build();

Expiration listeners are called synchronously as entries expire, and write operations to the map are blocked until the listeners complete. Asynchronous expiration listeners can also be configured and are called in a separate thread pool without blocking map operations:

Map<String, Connection> map = ExpiringMap.builder()
  .asyncExpirationListener((key, connection) -> connection.close())
  .build();

Expiration listeners can also be added and removed on the fly:

ExpirationListener<String, Connection> connectionCloser = (key, connection) -> connection.close();
map.addExpirationListener(connectionCloser);
map.removeExpirationListener(connectionCloser);

Lazy Entry Loading

Entries can be lazily loaded via an EntryLoader when ExpiringMap.get is called:

Map<String, Connection> connections = ExpiringMap.builder()
  .expiration(10, TimeUnit.MINUTES)
  .entryLoader(address -> new Connection(address))
  .build();

// Loads a new connection into the map via the EntryLoader
connections.get("jodah.net");

Lazily loaded entries can also be made to expire at varying times:

Map<String, Connection> connections = ExpiringMap.builder()
  .expiringEntry(address -> new ExpiringValue(new Connection(address), 5, TimeUnit.MINUTES))
  .build();

Expiration Introspection

ExpiringMap allows you to learn when an entry is expected to expire:

long expiration = map.getExpectedExpiration("jodah.net");

We can also reset the internal expiration timer for an entry:

map.resetExpiration("jodah.net");

And we can learn the configured expiration for individual entries:

map.getExpiration("jodah.net");

Additional Notes

On Variable Expiration

When variable expiration is disabled (default), put and remove operations have a constant O(1) time complexity. When variable expiration is enabled, put and remove operations have a time complexity of O(log n).

On Google App Engine Integration

Google App Engine users must specify a ThreadFactory prior to constructing an ExpiringMap instance in order to avoid runtime permission errors:

ExpiringMap.setThreadFactory(com.google.appengine.api.ThreadManager.currentRequestThreadFactory());
ExpiringMap.create();

See the GAE docs on threads for more info.

Additional Resources

License

Copyright 2009-2016 Jonathan Halterman - Released under the Apache 2.0 license.