Thread Synchronization Problem and Its Solution

Thread synchronization ensures correct execution of concurrent threads by managing their access to shared resources. Key issues include race conditions, deadlocks, and resource contention.

Agenda

• Intro to thread synchronization problems.
• When does the thread synchronization problem happen?
• The solution to thread synchronization problems.
• Properties of a good solution to synchronization problems.

Synchronization problems

1. Suppose we have a shared count variable, initially set to 0.
2. We want to perform 2 actions on this variable at the same time.
   • Add 1 to 100 to this variable using Thread 1.
   • Subtract 1 to 100 to this variable using Thread 2.
3. Finally, Print the count of a variable. Ideally, it should print 0. but it won't Reason is when more than one thread works on the same variable at the same time, it can lead to unexpected results.

When does the synchronization problem happen?

1. Critical Section
   • When more than one threads try to access the same code segment that segment is known as the critical section.
   • So, when more than one thread is there in the critical section at the same time, it can lead to unexpected results and synchronization problems.
2. Race Condition
   • If more than one thread tries to enter inside the critical section at the same time then it might lead to the synchronization problem.
3. Preemption
   • Preemption is the ability of the operating system to preempt(that is, stop or pause) a currently scheduled task in favor of a higher-priority task.
   • A program that is inside the critical section and CPU preempts then can lead to the synchronization problem.

Solutions to the synchronization problem

In C#, there are several ways to synchronize access to shared resources to ensure thread safety and prevent race conditions.

Using lock keyword: The lock keyword provides a convenient way to create a synchronized block of code. It internally uses the Monitor class to achieve synchronization. The lock keyword ensures that only one thread can execute the locked code block at a time.

Example

private static object syncObject = new object();
 private void Increment() 
 {
      lock(syncObject) 
      {
          // Critical section: Access shared resource
      }
 }

Using Monitor Class

Instead of using the lock keyword, we can directly use methods of the Monitor class for synchronization.

Example

private static object syncObject = new object();
 private void Increment() 
 {
      Monitor.Enter(syncObject) ;
      try
      {
          // Critical section: Access shared resource
      }
      finally 
      {
          Monitor.Exit(syncObject);
      }
 } 

Using Mutex

A mutex is a synchronization primitive that allows only one thread to acquire it at a time. It's typically used for inter-process synchronization to synchronize threads within the same process.

Example

private static Mutex mutex = new Mutex();
 private void Increment() 
 {
      mutex.WaitOne();
      try
      {
          // Critical section: Access shared resource
      }
      finally 
     {
         mutex.ReleaseMutex();
     }
 }

Using Semaphore

A semaphore is a synchronization primitive that allows a specified number of threads to enter a critical section simultaneously. It's useful when we want to limit the number of threads accessing a resource.

Example

private static Semaphore semaphore = new Semaphore(1,1); //Limits access to one thread
 private void Increment() 
 {
     semaphore .WaitOne();
     try
     {
         // Critical section: Access shared resource
     }
     finally 
     {
         semaphore.Release();
     }
 }

Using Interlocked Class

The Interlocked class provides atomic operations for variables that are shared between threads. It's useful for performing simple operations like incrementing a counter without the need for locking.

Example

private int counter = 0;
 public void Increment() 
 {
      Interlocked.Increment(ref counter);
 }

Properties of a good solution to the synchronization problem

• Mutual Exclusion: Only one thread should be allowed inside the critical section at any point in time.
• Progress: Overall system should keep on making progress. There shouldn't be a deadlock condition.
• Bounded waiting: No thread should wait outside the critical section infinitely. There should be some bound on the waiting time.
• No Busy Waiting: If a thread has to continuously check if they can enter inside the critical section or not is Busy Waiting.

while(!allowed to enter critical section)

{
    checking();  // This is the busy waiting.
}

there shouldn't be Busy waiting as it can have several consequences, like

• Inefficient use of CPU resources and wasted energy.
• Reduced performance.
• Increased power consumption.
• Potential deadlocks etc.
• The ideal solution should have some kind of Notification system. Once the thread is done with their task in the critical section, they will notify all the threads waiting outside.