The companies behind Internet.org, the organization formed by Facebook, Qualcomm and several others to bring the Internet to areas that still don't have it, released a document (PDF) yesterday detailing some of their plans for the initiative. One section stands out in particular, if only for its title – Facebook for Every Phone.
Other parts of the document involve advice for data center infrastructure (based on Facebook’s Open Compute Project), mobile app efficiency, and contributions from Qualcomm and Ericsson. It all follows Internet.org’s plan to increase the efficiency of delivering data to users by 100 times in the next decade, primarily by reducing both the costs of delivering data and the amount of data that mobile apps use.
The project is clearly part of Facebook’s long-term strategy. Both The New York Times and USA Today have reported that Facebook has made little-to-no money selling ads through Facebook for Every Phone. The app didn’t even have ad space until this July. But in certain emerging markets, which eMarketer identified specifically as India, Indonesia, Mexico, Brazil, and Vietnam, Facebook is considered responsible for connecting people to the Internet.
“In a lot of foreign markets, people think that the Internet is Facebook,” Clark Fredricksen, a vice president at eMarketer, told The New York Times in a July interview.
How it works
A lot of vendors make a lot of feature phones, so Facebook incorporated a “gateway,” which “creates an abstraction of a ‘canonical device’ on top of which the application logic is being built,” according to the document. This ensures Facebook for Every Phone lived up to its name and works with the more than 3,000 models of feature phones that access it today. Facebook also handles most high-impact decisions and processes for the app on the server side, enabling the company to fix bugs remotely without requiring users to upgrade.
From there, Facebook for Every Phone incorporates several tactics to reduce the amount of data sent to feature phones located in low-connectivity areas. The app’s client/server protocol accommodates 2G network bottlenecks by processing and re-sizing large files, such as images, on the server side before sending them to the device.
A process called “screen diffs” reduces the amount of data sent when users try to access screens that they had viewed earlier. For these, Facebook’s servers only send the updated content to the screens, rather than sending the entire screen every time the user accesses it.
Because most screens change only minimally, this process eliminates a lot of unnecessary data involved when users load a full copy of a screen that they had accessed earlier. (Facebook also must have recognized the potential privacy concern with its servers monitoring and storing the pages that users load, so it included this sentence – “When the client flushes a screen from its cache, the server is being notified to also eliminate its own copy.” Proactive denial of privacy invasion is par for the course these days.)
Persistent caching is handled on the server side as well, allowing Facebook to send the files stored on the client cache once, without a client’s request, and enabling the company to change its caching algorithm whenever needed.
Text translation, which is important for a project targeting such a diverse user base, is loaded for each specific user’s language as the screen loads. That means the app doesn’t need to include potential translations for every language, naturally reducing the size of the app when it’s downloaded.
A few other features help fit Facebook onto a feature phone, but the goal is clear – adapt Facebook to the needs of its untapped markets, rather than wait and hope for them to get the technology on their own.
What it means
Ostensibly, including information on how the Facebook for Every Phone app works is no different than the information Facebook provided on Open Compute. The company has had projects in the past that relate to the goals of Internet.org, and now it’s sharing information on how these projects could improve Internet.org.
The difference, though, relates exactly to what Fredricksen said in that New York Times article – some people think the Internet and Facebook are one and the same.
Facebook for Every Phone is, obviously, a Facebook-branded product. Whether or not Facebook makes money off it right now is less relevant than how Facebook establishes itself among a massive market of recently connected users. If Facebook contributes its Open Compute strategy to Internet.org, users won’t really know the difference, aside from those who go out of their way to look it up.
So far, it’s unclear if Internet.org will promote the branded Facebook for Every Phone app for feature phones or simply create a non-Facebook-branded version. If it does use Facebook’s product, though, the tangible benefits of Internet.org become clearer for Facebook. The company doesn’t have to just hope that the new users that its collaborative initiative connects to the Internet will one day use Facebook. With Facebook for Every Phone, it would force them to use it.
Colin Neagle covers emerging technologies and the startup scene for Network World. Follow him on Twitter and keep up with the Microsoft, Cisco and Open Source community blogs. Colin's email address is email@example.com.