Facebook Creates An Open Source Group, Introduces A New Routing Technology
Facebook announced on Monday at @Scale 2014 conference the creation of an open source group called TODO, which is short for “talk openly, develop openly.” With this group, Facebook aims to address the challenges that open source community faces. They will guide and advise the companies that want to consume open source software and run open source programs.
For those who do not know what open source is, open source is a development model in which the code and design of software is made public and anyone can see it and improve it. This development model is a great way for researchers to develop products quickly. It also eliminates the possibility of reinventing the wheel and helps researchers reuse the already developed software for their own product following the license terms at the same time.
Facebook’s TODO organization aims to use better open source programs in their companies, and help everyone who want to create open source programs or consume them. Many companies including Box, Dropbox, GitHub, Google, Khan Academy, Stripe, Square, Twitter, and Walmart Labs have collaborated with TODO and Facebook hopes that more will join soon. More details will be shared by Facebook in coming weeks.
Facebook also announced on the same day about their new routing technology called “mcrouter” (read: mik-router) which is used to handle the data from Facebook’s cache servers. Cache servers store the frequently or recently used data so that it can be fetched quickly than is possible if it is stored on databases, ultimately increasing the response time for Facebook users.
mcrouter is a memcached protocol that is capable of servicing as much as 5 billion requests per second because of its distributed design. Facebook said on the blog post announcing mcrouter that the technology was used as a standalone binary in an Amazon Web Services setup used by Instagram last year before it was fully transitioned to Facebook’s infrastructure.
Now, the social media giant wants to provide their knowledge, gained by their experience with big data, to the companies that aim to scale. Moreover, these initiatives will also help in innovating new ways to handle big data.
Facebook already keeps their system designs and techs open and anyone can learn about how they have managed to service billions of users. They have also released the details about their file system called Hadoop Distributed Filesystem (HDFS) responsible for forming the basis of many large-scale storage systems at Facebook.