Ambry is optimized to serve immutable objects of few KBs to multiple GBs in size with high throughput and low latency. Ambry can play a critical role in this future, and in the future of any company that is interested in diverse kinds of media to a global audience.Īmbry is a distributed immutable object store that is highly available and very easy to scale. Media pipelines will need to be supported by more companies, especially with the advancement of video and virtual reality. Media content has become critical for any website to increase user engagement, virality and monetization. Ambry is optimized to store and serve media. Today, we are announcing that Ambry is now available as an open source project under the Apache 2.0 license. Also, we have given presentations to a few companies about the system and have seen great interest in adopting it for their use case. Since we started sharing data about this internal project back in 2014, Ambry has made significant performance improvements, in both latency and network efficiency. Two years ago, we revisited the technology we were using and started on a journey to fix these issues. The backend storage system acts as the origin server for the content served by the CDNs.Īs LinkedIn traffic grew, the system that we were using to store media content traditionally had increasing scalability, availability and operational issues. These media types get stored on our backend and are predominantly served by our Content Delivery Networks (CDN). Profile photos, email attachments, logos, and influencer posts are a few examples of where photos, videos, PDFs, and other media types get uploaded and displayed to the end user. Media content has become ubiquitous around the web and almost all of Linkedin's new features interact with media in some form or the other.