The big challenge for engineers is to keep the site up alive and running smoothly despite handling nearly half a billion active users. This article looks at some of the software and the techniques they use to achieve this.
The scale of the challenge Facebook
Before we get into this in more detail, here are some facts as an illustration of the scale of the challenges that must be taken care Facebook:
1. Facebook serve 570 billion page views per month (according to Google Ad Planner).
2. Photos on Facebook more than all other photo sites.
3. More than 3 billion photos are uploaded each month.
4. Facebook serves 1.2 million pictures per second. This does not include images served by the CDN Facebook.
5. More than 25 billion content (status updates, comments, etc.) are shared each month.
6. Facebook has more than 30,000 servers (this number is the data from last year
7. Software Behind Facebook
In some ways Facebook is still a LAMP-based sites (like), but modified and expanded its operations by incorporating many elements and other services, and modified its approach to be like that exist today.
For example:
* Facebook is still using PHP, but have made compilernya so it can be converted to native code on its web servers, thus improving performance.
* Free use Linux, but has been optimized for its own purposes (especially in terms of tissue).
* Free use MySQL, but mainly as a key-value persistent storage, moving joins and logic to the web server because the optimization is more easily done over there (in the "other side" of memcached).
Then there are custom-written systems, such as Haystack, a highly scalable object store is used to serve the photos on Facebook a lot, or Scribe, logging system that can operate on a scale up.
Now let us discuss (some of) the software (software) that are used up in providing the largest social networking site in the world.
Memcached
Memcached is now one of the most popular software on the internet. It is a distributed memory caching system that is used up (and many other sites) as a caching layer between the web server and MySQL server (because of relatively slow database access).
Over the years, Facebook has repeatedly perform optimization on the memcached and the surrounding software (such as optimizing the network buildup).
Facebook running memcached server with tens of thousands of terabytes of data cache at one point in time. This possibility is the world's largest Memcached installation.
HipHop for PHP
PHP, a scripting language, is relatively slow when compared with the code that runs natively on the server. Hiphop convert PHP into C code that can then be compiled to produce better performance.
This has allowed up to optimize its web server because it relies heavily on PHP for serving content.
A small team of engineers (initially three people) at Facebook spent 18 months developing hip-hop, and now has in production.
Haystack
Haystack is a storage system / high-performance search photos on Facebook (Truly Haystack is a storage object, so it does not save the image). He was doing heavy work, there are more than 20 billion photos uploaded in Facebook, and each is stored in four different resolutions, resulting in more than 80 billion photos.
And it is not just about the ability to handle billions of photographs, performance is also very important. As previously mentioned, Facebook serves approximately 1.2 million pictures per second, the numbers do not include pictures / photos are served by the CDN Facebook.
BigPipe
BigPipe is a dynamic web page service system developed by Facebook. Facebook used to serve every web page in multiple sections (called "pagelets" to produce optimal performance.
For example, the chat window is taken separately, the news feed is taken separately, and so on. Pagelets this can be taken in parallel, so kenerja stay awake, and also allows the user to access the site properly though some parts of disabled or damaged.
Cassandra
Cassandra is an open source system that functions as a distributed database management. Facebook then develop it and use it to feature NoSQL Inbox Search. In addition to facebook, a number of other services also use, for example Digg.
Scribe
Scribe is a flexible logging system is used internally by Facebook for many purposes. It was developed to handle logging in Facebook, and automatically handles the emerging new logging category.
Hadoop and Hive
Hadoop is an implementation map-Reduced open source that allows for the calculation of large amounts of data. Facebook use it for data analysis (and as we know, Facebook has large amounts of data).
Hive developed by Facebook, and with it possible to use SQL queries against Hadoop, making it easier for non-programmers to use it.
Both Hadoop and Hive is open source and is used by a large number of services, such as Yahoo and Twitter.
Thrift
Facebook uses several different languages ??for some services. PHP is used for front-end, Erlang is used for Chat, Java and C are also used in some places (TC ESSWA
OIV AW
