mctolentino / generic-object-pool Goto Github PK

An Object Pool is used to reduce the overhead of creating and instantiating a new object. Usually it's used to boost performance for objects that are heavy on resources like database, streams and sockets connections.

Concurrency is usually a major concern with these shared resources.

• com.cashwerkz.challenge.ObjectPool
  • The main class that implements the generic object pool is ObjectPool.
  • A queue that is thread-safe (ConcurrentLinkedQueue) was used to save the generic pooled objects.
  • This is an abstract class, and it delegates the createGenericObject method to the implementing class.
  • Initial implementation of the createGenericObject method was a bit hacky (using newInstance()). Making the method abstract feels a bit cleaner. (You may check logs for previous implementation)

• com.cashwerkz.challenge.test.ObjectPoolTest

• This is the main test class. Test cases:

• public void testCreatePool()

• public void testAcquireObjectFromPool()

• public void testReturnObjectToPool()

• public void testIncreasePoolSizeByGrowthRateIfThresholdIsReached()

• public void testThrowExceptionIfCeilingParameterIsExceeded()

• public void testUsePoolByDifferentThreads() throws InterruptedException

• com.cashwerkz.challenge.test.SamplePoolObject

• Sample object with an overriden toString() method, for easier and clearer debugging.

• com.cashwerkz.challenge.test.SampleTask

• A runnable task, created to mimic separate processes that requests from the object pool.

• Always include test cases for concurrency. My initial implementations that are running in a single thread, broke when faced with multiple concurrent processes. Synchronized methods saved the day.