Difference between Optimistic Concurrency and Pessimistic Concurrency

In this article, we discuss about the differences between Optimistic Concurrency and Pessimistic Concurrency.

Concurrency control is another property of a database system that provides isolation for various transactions to run simultaneously without interference. Concurrency control can be divided into two categories Pessimistic and Optimistic. Any one of these approaches would be better than the other depending on the way in which conflicts among transactions are detected or resolved. This article explains such methods and analyse the benefits, drawbacks, and applicable scenarios for each case.

What is the Optimistic Concurrency?

Optimistic concurrency allows transactions to proceed without locking. It permits several transactions to continue without the lock on the resources being placed initially. Some of the conflicts may occur but they can be resolved after the transactions are completed. Each transaction operates independently and assumes that other transactions would not affect it much.

This approach detects conflicts after the transaction tries to commit. When the transaction is completed, it tests whether the data it worked on has been modified by some other transaction. If there is no conflict, the transaction commits; if it fails the check because of a conflict, the transaction is aborted, and it may be retried.

Mechanism of Optimistic Concurrency:

• No Locking: The data are not locked during the transaction, which allows for access by other transactions. These other transactions may access and modify data being changed by this transaction.
• Versioning: A Versioning system is often used to detect conflicts (e.g., timestamps or version numbers on data rows).
• Conflict Resolution: At commit time, the system checks whether another transaction has made a change to it. In case of a conflict, it rolls back one transaction from the two competing choices and may attempt a retry.

Examples:

• Generalized Databases: The restriction of locking is a common occurrence in distributed systems because these locks tend to become the bottleneck at times and slow down the performance.
• Eventual Consistency: Systems like NoSQL databases (e.g.Cassandra, MongoDB) often rely on optimistic concurrency mechanisms for consistency over a period of time.

Advantages of Optimistic Concurrency:

Several advantages of Optimistic Concurrency are as follows:

• In systems with low write contention, there will be higher throughput.
• There is higher concurrency, as no locks are blocking other transactions.
• It provides better performance in a distributed system environment or ones in which most transactions are read-only.

Disadvantages of Optimistic Concurrency:

• Frequent rollbacks under high contention may be harmful to performance.
• It requires complex conflict resolution mechanisms to deal with data that has changed while the transaction was in progress.

What is the Pessimistic Concurrency?

Pessimistic concurrency control assumes that conflicts are most likely to occur. It is the mechanism to Lock resources to avoid conflicts. When a transaction wants to access a data item, it first puts a lock on it; thus, no other transactions can access, read, or modify the same data item until the lock is released (which usually happens once the transaction completes).

The mechanism ensures that one transaction acts on a specific data item at a time, controlling conflicts at the cost of concurrency. Any transaction wishing to access any portion of a locked resource must wait until the lock is freed.

Mechanism of Pessimistic Concurrency:

• Locks: The resource(s) (rows, tables in a database) lock at the beginning of the transaction. The locked resource fulfilling the transaction is not accessible to other transactions until the lock is released.
• Deadlocks: Deadlocks may occur with two or more transactions waiting for locks released whenever seeking access to one another's objects.
• Blocking: Any transaction that requests a locked resource is either blocked or must wait for the lock to be released, giving rise to the serialization of transaction execution.

Examples:

• Banking Systems: Financial activities involving banking systems often make use of pessimistic concurrency control. Its use is especially intended in systems, where data integrity has to be taken care, where conflicts naturally have to be avoided (e.g., money transfers between accounts).
• Inventory Management: Systems managing stock levels depend on locks to ensure that no two transactions procure resources of the same inventory at the same time.

Advantages of Pessimistic Concurrency:

• It ensures integrity of data by preventing conflicting updates.
• It decreases the number of rollbacks because conflicts get avoided upfront.

Disadvantages of Pessimistic Concurrency:

• Performance degradation because of the locking overhead, especially in systems with high concurrency.
• Deadlocks occur when two or more transactions are waiting for each other's locks.
• Lowering of concurrency due to other transactions queuing up for the lock release.

Key differences between Optimistic Concurrency and Pessimistic Concurrency:

There are several key differences between Optimistic Concurrency and Pessimistic Concurrency. Some main differences are as follows:

Conclusion:

In summary, both Optimistic and pessimistic concurrency models have their own merits and are applicable in different systems. Optimistic concurrency is suitable in low-contention areas with more reads, providing improved throughput since no blocking with rollback in case of conflict. An example of such a system would be a distributed or NoSQL database where eventual consistency is tolerable. On the other hand, pessimistic concurrency is ideal in a high contention environment with a strong need for integrity, such as a banking or inventory system, where it blocks the resource early to avoid conflict. However, it may lead to performance penalties because of the overhead with blocking and the possibility of deadlocks. Depending on the system's performance, consistency, and how it handles the conflict, the decision should be made according to the nature of the system.