Managing access to shared resources is important in concurrent programming to ensure data consistency. Traditional synchronized keyword lacks flexibility for complex scenarios like fairness, acquire lock immediately, wait for a specific time to acquire lock, etc. Lock, ReadWriteLock, and StampedLock APIs are introduced in Java 1.5 to addresses these limitation and provides great control to developers to handle shared resources access. They are part of java.util.concurrent
package.
Lock API
Lock API is an interface and provides below methods to handle thread synchronization.
-
void lock()
It is used to acquire the lock. If the lock is not available, the thread will be blocked until the lock is acquired. -
void lockInterruptibly()
It is same aslock()
, but this thread can be interrupted. -
boolean tryLock()
It returns true if the lock is acquired; otherwise, false. The thread will not be blocked if the lock is not granted. -
boolean tryLock(long timeout, TimeUnit unit)
It is same astryLock()
but waits for a specified amount of time only. -
void unlock()
Releases the lock. -
Condition newCondition()
Provides the ability for a thread to wait for some condition to occur while executing the critical part.
ReentrantLock
- ReentrantLock class implements the Lock interface.
- A reentrant lock allows a thread to acquire a lock multiple times and release it multiple times.
- If you want
fairness
policy, use parameterized constructor of ReentrantLock. - If
fairness
is enabled, It makes sure that the longest waiting thread is given access to the lock. - Use finally block to release the lock to avoid deadlock situation.
ReentrantLock lock = new ReentrantLock(true); //fair lock
public void methodExecution() {
lock.lock();
try {
// Critical section here
} finally {
lock.unlock();
}
}
ReadWriteLock API
ReadWriteLock API is an interface and maintains a pair of locks to maintain read and write scenarios. The read lock can be held simultaneously by multiple threads if there are no writers. The write lock is exclusive.
-
Lock readLock()
- Returns the lock used for reading. -
Lock writeLock()
- Returns the lock used for writing.
Key Rules
- If you have a read lock, you can't acquire a write lock.
- If you have a write lock, you can't acquire a read lock in any other thread.
- If you have a write lock, you can acquire another write lock in the same thread.
- A write lock is allowed when no other read or write lock is active.
- While waiting for a write lock, no new read locks from other threads are allowed.
ReentrantReadWriteLock
- ReentrantReadWriteLock class implements the ReadWriteLock interface.
- Useful if we have more read compare to write.
Example: In micro-service communication, assume Service B expects a JWT from Service A. JWT is generated by Service A and can be cacheable for couple of minutes. In this scenarios ReentrantReadWriteLock will be more useful. We can renew the token if it is expired or about to expire. I am not acquiring read lock here to avoid read starvation.
String jwt = JwtUtil.generateJwt();
ReadWriteLock lock = new ReentrantReadWriteLock();
Lock writeLock = lock.writeLock();
Lock readLock = lock.readLock();
public String getJwt(String key, String value) {
if (StringUtils.hasLength(jwt)) {
if (JwtUtil.isJwtEligibleToRenew(jwt)) {
generateJwt();
}
} else {
generateJwt();
}
return this.jwt;
}
public void generateJwt() {
this.writeLock.lock(); //write lock
try {
if (JwtUtil.isJwtEligibleToRenew(jwt)) {
this.jwt = JwtUtil.generateJwt();
}
} finally {
this.writeLock.unlock(); //release write lock
}
}
StampedLock API
StampedLock was introduced in Java 8 and it is a class. It supports three modes of locking and returns a stamp that is used to release the lock. It allows lock upgradation.
-
Read Lock
Allows multiple threads to read concurrently. -
Write Lock
Returns a lock used for writing. -
Optimistic Read Lock
Allows a thread to read without acquiring a traditional read lock, which improves performance as it avoids contention.
-
long writeLock()
Acquires the exclusive lock, blocking if necessary until available. -
void unlockWrite(long stamp)
Releases the write lock. -
long readLock()
Acquires the read lock, blocks if a write lock is in progress. -
void unlockRead(long stamp)
Releases the read lock. -
long tryOptimisticRead()
Returns a stamp that can later be validated, or zero if exclusively locked. -
boolean validate(long stamp)
Returnstrue
if the lock has not been exclusively acquired (write lock) since the issuance of the given stamp; otherwise,false
. -
long tryConvertToWriteLock(long stamp)
Upgrades a read or optimistic lock to a write lock. Blocks until it is available.
Key Points:
- It is not reentrant, meaning a thread holding a lock cannot acquire it again.
- It doesn’t support conditions.
- It doesn’t support fairness.
public void readAndWrite() {
long l = this.stampedLock.readLock();
try {
//critical section
} finally {
this.stampedLock.unlock(l);
}
long w = this.stampedLock.writeLock();
try {
//critical section
} finally {
this.stampedLock.unlock(w);
}
}
Optimistic read example. This allows a thread to read without acquiring a traditional read lock, which improves performance as it avoids lock contention. If any write lock acquired after acquiring optimistic read lock time, validate() will return false
otherwise it will be true
.
public class StampedLockExample {
private StampedLock stampedLock = new StampedLock();
private int cachedValue = 0;
private long optimisticLock = stampedLock.tryOptimisticRead();
public int readWithOptimisticLock() {
if (!stampedLock.validate(optimisticLock)) {
System.out.println(" Optimistic lock is invalid");
optimisticLock = stampedLock.readLock();
try {
} finally {
stampedLock.unlock(optimisticLock);
}
this.optimisticLock = stampedLock.tryOptimisticRead();
} else {
System.out.println("Optimistic lock is still valid");
}
return cachedValue;
}
public static void main(String[] args) {
StampedLockExample threadsExample = new StampedLockExample();
threadsExample.readWithOptimisticLock();
long l = threadsExample.stampedLock.writeLock();
threadsExample.stampedLock.unlock(l);
threadsExample.readWithOptimisticLock();
}
}
Happy Coding and Learning !!!
Please drop a comment if you have any question.
Top comments (0)