Wanna be Facebook? It just open-sourced some of its web server code. Now to find 1bn users...

Facebook has rolled out another chunk of open-source code, this time a C++ HTTP stack called Proxygen, which includes a web server. Techies at the social network hope other developers will use the BSD-licensed software as the basis of their own web apps, and have included a bunch of “sensible defaults” to get people on their …

[Christian Berger]

That reminds me of the AOL web server

That also eventually became a piece of open source software.

http://en.wikipedia.org/wiki/AOLserver

I wonder if that's still in use.

[btrower]

Very nice

Beyond the fact that it is an honorable thing to release the code as open source, from a quick review the code quality is better than the average open source stuff. In fact, though I quibble with multiple points of exit and maybe a bit too much verbosity in comments separating some lines of code, it is significantly cleaner and more mannerly than the vast majority of code I see out there.

Two very big thumbs up both for facebook fostering and freeing the code and the developers and contributors for producing an exemplary work product.

[Bill Redmond]

But..

Will it run on a raspberry pi??

[btrower]

Re: But..

It is written in C++ and I did not see anything that looked particularly heavyweight. Unless there is something unusual happening it should run fine on a raspberry pi. I think a big motivation behind someone doing a stand-alone like this is precisely to get rid of mountains of dependencies they don't need. Small footprint means better performance and arguably small amounts of increased percentage performance on facebook's massive infrastructure is more important than it is on a small device like the raspberry pi. Small savings are multiplied many times over on a huge system like that.

My attention is directed elsewhere but I may at some point pull that stuff down, compile and give it a whirl.

I am still a little bit hopeful that facebook will turn into a good player. Their monstrous privacy invasion and erection of a walled garden was basically a necessary self-defense against encroachment by another player. They are large now and the dynamics of group forming networks makes them enormously powerful and qualitatively different from a smaller version of this type of network. However, I still see some vulnerabilities and given the value of maintaining their position (I write elsewhere about a facebook financial 'singularity') you can't fault them for playing things close.

To the extent that facebook has had the latitude to do good things without self-immolation they appear to have done so. Time will tell. My instinct is that these gigantic organisms are intrinsically evil, but this recent open handed release of code is a hopeful sign.

[Tom Samplonius]

Re: But..

"... but this recent open handed release of code is a hopeful sign."

Because Facebook hasn't be making open source releases continuously over the last 10 years. And their unique hardware designs aren't open source either. And Presto (http://prestodb.io/), the tool used for the querying data from their "monstrous privacy invasion" isn't open source either. So this is a complete turn around for Facebook.

[wikkity]

Incorrect link

The link for Folly is incorrect, you are missing a :/ after https. Firefox sends me to https.com that redirects me to HomeImprovement.com. I know many people look at open source as DIY software so maybve thats right.

[Anonymous Coward]

Maybe it's me

But I didn't see a way to integrate this into an event loop, which is quite essential if you want to actually serve something useful (like the result of database queries, etc...)