Here is the video I like to watch.
Ding Hui's profile is here.
Python is simple and clean, and favors pragmatism
1. Scope the problem, AKA, do simple things first.
2.
Scaling python to support user and feature growth
Python efficiency strategy
1. Build extensive tools to profile and understand perf bottleneck
2. Moving stable, critical components to C/C++, e.g., memcached access
3. Cythonization
4. Sync? New python runtime?
Ding Hui's profile is here.
Head of Instagram infrastructure team, overseeing Instagram's core infrastructure including Django/Python platform, social graph and Cassandra key-value storage systems, video infrastructure and realtime infrastructure. Responsible for driving reliability, scalability, efficiency and quality of backend infrastructure and web platform. Built out the team from 2 engineers to 70+ engineers across multiple teams. Proud of building a strong engineering culture within the team, and cultivated broad cross-functional collaborations within Facebook. Proven track record of successful delivery of multiple large scale projects. My team also plays a key role in delivering major products at scale, together with our product engineering teams, e.g., Instagram Direct, Stories, Live, IGTV.
Instagram runs the world's largest python fleet in production and we pride ourselves in the way we drive scale, efficiency and technical innovation with Python.
Instagram is also the world's top 3 user of Apache Cassandra, open source key-value storage system. We have been leading some of the 10x changes to Cassandra performance.
No comments:
Post a Comment