Member-only story
Distributed Transactions & Two-phase Commit
Anatomy of Two-phase commit
Introduction
In today’s world, data is growing at an enormous rate. Enterprises have built innovative solutions to handle a humongous amount of data. It’s not uncommon to see data distributed across many machines or databases. This technique is known as ‘Sharding’, helps in building scalable & reliable systems.
Tech companies also are adopting a microservices architecture. In this type of architecture, every microservice manages its own database. To add or modify data in a different database, it calls the responsible microservice.
Distributing data across many machines comes up with its own set of challenges. Data management for a monolithic non-sharded system is straightforward. Relational Databases such as Postgres, MySQL, etc. offer A.C.I.D properties out of the box.
The same isn’t applicable for a sharded database or data distributed across microservices. In this article, we will see understand atomicity while handling distributed transactions. We will take a look at a protocol called 2-Phase Commit, that helps us achieve the same. So, let’s get started.
