Building highly reliable and elastically scalable backend services is critical for serving web workloads like Facebook’s. At the heart of such services is a distributed coordination service. Zookeeper was evaluated at Facebook for this and we found its performance and scalability inadequate. Project Zeus was started to solve this problem, with the Zookeeper code base as a starting point. Zeus significantly improved upon Zookeeper and added several features and performance enhancements that were required for our workload.
Join Vishal Khaturia and learn more about the Facebook workload, how Zookeeper fared on that workload and a dive in to the architectural elements and features in Zeus that enabled it to handle this workload well.
A little about Vishal, he is an engineer at Facebook and he works on making services at Facebook fault tolerant, elastically scalable and reliable in the presence of dynamically changing workload. He led the design and development several systems at Facebook, including Zeus: Facebook wide coordination service, Facebook’s service discovery and shard management systems and a fault tolerant distributed key/value store. Prior to Facebook, Vishal was a lead engineer at Microsoft on a skunks works project to build a large scale distributed database system.
Want an Uber driver to get you here? Get $20 off your first ride Text SVCLOUD to 827-222.You can also sign up at uber.com/go/SVCLOUD.