Protecting shared sensitive shared resources from race conditions is hard enough when running in a single process. Let's make it harder: how do you do it in a distributed system? Scale it up and run it in a production system that serves millions of transactions. See where we’re going here?
In this talk, I will take you along our journey of implementing distributed locks in Node.js, how they broke, and how we evolved our algorithm over time to support thousands of requests per second in a mission-critical financial system. We will discuss different ways to implement distributed locking, using different tools and approaches. Learn how to improve performance to support more than 10,000 transactions per second which are all trying to access the same resource while running workflows on multiple data centers and how to verify it doesn’t break.
Or started coding at the age of 12 and never stopped since. At Remitly Israel, he is in charge of innovation and leads several engineering teams. With a BSc in Computer Science and Electrical Engineering cum laude from Tel Aviv University, Or officially joined the Node.js community in 2012. He has since fostered a career in tech that spans over the last 15 years and originates in his love of creating great products that scale well. Before joining Remitly Israel, Or gained professional experience from leading tech companies such as Microsoft and Oracle.