Research in Brief: Neural Egg Separation: Training ML systems to extract audio and video from noisy, cluttered environments on, Facebook's Engineering blog


The Facebook open switching system (FBOSS) is switch software created specifically to address the needs of a large, rapidly growing network. At ACM SIGCOMM 2018, we shared new details about the open source version of FBOSS and our experiences and challenges using FBOSS in the Facebook network.


To make FBOSS easier to implement, maintain, and update, we developed it as if we were building a large-scale software service. The code is leaner than standard network switch software because we stripped away everything not needed for its particular application here at Facebook and leveraged a “deploy early and iterate” approach. Also, thanks to its modular abstractions, FBOSS is not tied to a specific hardware or feature set.

This approach has helped us overcome the complexity and scaling challenges involved in implementation. As a result, FBOSS deployments in our data centers have increased by 30x over two years and we’ve been able to nimbly respond to new developments. FBOSS was an important part of our implementation of IPv6 for our specific operational needs and more recently, our inter-building Fabric Aggregator.


Our family of applications demands immersive, real-time interactions and imposes technical challenges for our large-scale data centers. FBOSS addresses many of those needs specific to Facebook. By sharing our design and experiences, we hope to encourage further research and development in this area. As our network grows, the importance of open and streamlined infrastructure design such as FBOSS will only increase.


FBOSS: Building Switch Software at Scale

To help personalize content, tailor and measure ads and provide a safer experience, we use cookies. By clicking or navigating the site, you agree to allow our collection of information on and off Facebook through cookies. Learn more, including about available controls: Cookie Policy