Facebook open sources McRouter tool

By
Follow google news

Handles close to five billion requests at Facebook.

Social network Facebook today said it would open source the McRouter (memcached protocol router) software it uses to speed up web server access to database back-ends.

Facebook open sources McRouter tool

Facebook Engineering said it originally built McRouter to make Memcached more flexible and scale to its MySQL database and web server clusters. 

The scalability comes from the ability to allow working data datasets grow beyond individual Memcached servers and be served between several different servers through an algorithm that segments the cached information and splits it into different pools.

McRouter was developed in C and C++ and is licensed under the liberal BSD license. It supports any caching program that follows the Memcached protocols.

Memcached itself is an open source distributed in-memory object caching system aimed at storing small amounts of arbitrary data such as application programming interfaces and database calls.

Although McRouter was written for Facebook's internal use, it has been adopted by Instagram, as well internet forum Reddit.

At Facebook and Instagram, McRouter is claimed to manage close to five billion requests during peak usage times.

The source code for McRouter can be found on Github.

Read iTnews's Labs guide to five zero-cost ways to improve the performance of MySQL.

Add iTnews as your trusted source

Got a news tip for our journalists? Share it with us anonymously here.
Copyright © iTnews.com.au . All rights reserved.
Tags:

Most Read Articles

Westpac brings automation and AIOps to life, chasing CPU and memory alerts

Westpac brings automation and AIOps to life, chasing CPU and memory alerts

Suncorp to have AI agents in insurance claims process as soon as this month

Suncorp to have AI agents in insurance claims process as soon as this month

Services Australia describes fraud, debt-related machine learning use cases

Services Australia describes fraud, debt-related machine learning use cases

Qld gov backs technology projects with at least $340m

Qld gov backs technology projects with at least $340m

Log In

  |  Forgot your password?