CODEWITHSUNDEEP
javaperformancescheduledexecutorservice
fatih
fatih

Reputation: 105

ScheduledExecutorService performance

I have a list of objects whose maximum size is 5000. When an object has not been updated for a type specific amount of time like 5, 10 or 100 seconds, it will be deleted from the list.

What is the best or preferred option to handle such a situation?

What do you recommend?

Upvotes: 5

Views: 854

Answers (2)

fps
fps

Reputation: 34450

If you can use Google's Guava, you should give the Cache class a try.

For example, you could have one cache for each type of object:

LoadingCache<String, ObjectOfType1> type1Cache = CacheBuilder.newBuilder()
   .maximumSize(5000)
   .expireAfterWrite(5, TimeUnit.SECONDS)
   .removalListener(MY_LISTENER)
   .build(
       new CacheLoader<String, ObjectOfType1>() {
         public Graph load(String key) throws AnyException {
           return createExpensiveGraph(key);
         }
       });

And for Type2:

LoadingCache<String, ObjectOfType2> type2Cache = CacheBuilder.newBuilder()
   .maximumSize(5000)
   .expireAfterWrite(10, TimeUnit.SECONDS)
   .removalListener(MY_LISTENER)
   .build(
       new CacheLoader<String, ObjectOfType2>() {
         public Graph load(String key) throws AnyException {
           return createExpensiveGraph(key);
         }
       });

Then, you could just use a cache as if it were a Map:

ObjectOfType1 o1 = type1Cache.get("1");
ObjectOfType2 o2 = type2Cache.get("2");

Upvotes: 1

user520458
user520458

Reputation:

I'm actually using an implementation which uses a Map<Object, Long> for storing the expiration times of each element, and a java.util.Timer to run every n seconds, removing every expired element.

I can't really say it's the best implementation, since I only use it for a few hundred simple elements (i.e. not complex objects).

Upvotes: 0

Related Questions