Performance and Scalability - Java Concurrency in Practice - page 272

Listing 11.5. Reducing Lock Duration.
Reducing the scope of the lock in userLocationMatches substantially reduces the num-
ber of instructions that are executed with the lock held. By Amdahl's law, this removes an
impediment to scalability because the amount of serialized code is reduced.
Because AttributeStore has only one state variable, attributes , we can improve
it further by the technique of delegating thread safety ( Section 4.3 ). By replacing at-
tributes with a thread-safe Map (a Hashtable , synchronizedMap , or Concur-
rentHashMap ), AttributeStore can delegate all its thread safety obligations to the
underlying thread-safe collection. This eliminates the need for explicit synchronization in
AttributeStore , reduces the lock scope to the duration of the Map access, and removes
the risk that a future maintainer will undermine thread safety by forgetting to acquire the ap-
propriate lock before accessing attributes .
While shrinking synchronized blocks can improve scalability, a synchronized block
can be too small—operations that need to be atomic (such updating multiple variables that
participate in an invariant) must be contained in a single synchronized block. And be-
cause the cost of synchronization is nonzero, breaking one synchronized block into mul-
tiple synchronized blocks (correctness permitting) at some point becomes counterpro-
ductive in terms of performance. [9] The ideal balance is of course platform-dependent, but in