How PayPal Was Able to Support a Billion Transactions per Day With Only 8 Virtual Machines

This post discusses how PayPal scaled its operations to handle a billion daily transactions with just eight virtual machines by utilizing the actor model supported by the Akka framework. Initially, PayPal's early growth involved scaling horizontally with over 1000 virtual machines, which led to high latency, infrastructure complexity, and resource wastage. Transitioning to the actor model, which is known for its concurrency and lightweight structure, allowed efficient message handling and reduced resource consumption. The actor model enables enhanced performance by processing messages asynchronously and in a sequential order, minimizing the dependency on external databases due to local state caching. Fault tolerance is managed by supervising actors that can restart if failures occur, offering scalable and error-resilient performance. Although it simplifies concurrency management, the actor model introduces a learning curve for developers and requires careful management of race conditions and deadlocks. This approach allowed PayPal to significantly scale operations with reduced infrastructure, showcasing an innovative solution to scalability challenges.