Interview: Eben Moglen - Freedom vs. The Cloud Log (The H)
[Posted March 17, 2010 by corbet]
The H has an
interview with Eben Moglen. "And so, basically, what I am
proposing is that we build a social networking stack based around the
existing free software we have, which is pretty much the same existing free
software the server-side social networking stacks are built on; and we
provide ourselves with an appliance which contains a free distribution
everybody can make as much of as they want, and cheap hardware of a type
which is going to take over the world whether we do it or we don't, because
it's so attractive a form factor and function, at the price."
(Log in to post comments)
Freedom vs. The Cloud Log
Posted Mar 17, 2010 21:21 UTC (Wed) by kragil (subscriber, #34373)
[Link]
That personal social server idea is awesome .. for geeks.
I thought about this a lot and for something like this to really overtake
facebook you need lots of really hardcore development. Some glue code and a
bit of Python won't do.
First security:
I think a Wua.la-like crypttree implementation where access to certain
parts of your profile/files is handled via keys (exchange) is needed.
Currently I know of no FOSS solution for this problem.
Second storage:
Facebook stores tons of pictures and videos. Especially the videos could
kill such a small wall wart.
Possible solutions:
- owner free data (like OFFsystem.sf.net) where donors donate hosting,
storage and bandwidth to keep the system running
- something working like wuala, also with donors instead of one company
Retroshare.sf.net is probably the closest thing currently working. But it
lacks all the important bits.
I think this social wall wart will not kill facebook, but a software that
would allow me to share anything encrypted with my friend (real friends)
with my current computer could have a lot users and that software should be
available for any platform.
Some challenges still remain with these solutions. ISP will not like the
private hosting and governments will not like the encryption.
Freedom vs. The Cloud Log
Posted Mar 17, 2010 23:10 UTC (Wed) by njwhite (subscriber, #51848)
[Link]
I disagree about storage being unreasonable in such a model. I have a uSD card which is
8GB which is about the size of my fingernail. I don't know if the SheevaPlug referenced in
the article has a uSD slot (doesn't look like it), but it certainly is perfectly conceivable to
create such a device, today. 8GB is a lot of storage by 'storing my data in the cloud'
standards, and could be extended very cheaply.
Another advantage of this pushing servers out to the edge of the network model is that
scalability isn't such a problem - if I serve family photos to 20 friends a day that's quite a
lot.
What's 'crypttree' technology? I haven't heard of it before - sounds like I should read
more about it.
Freedom vs. The Cloud Log
Posted Mar 18, 2010 1:28 UTC (Thu) by JohnLenz (subscriber, #42089)
[Link]
I think the grandparent is referring to Attribute Based Encryption, see
this paper (pdf).
Freedom vs. The Cloud Log
Posted Mar 18, 2010 8:15 UTC (Thu) by kragil (subscriber, #34373)
[Link]
And at the moment Facebook saves 2.5 billion pictures every month. Once
people upload HD video from cell phones at a similar rate SD cards won't
work.
A FOSS solution should work for 2012 and beyond, not just today.
Freedom vs. The Cloud Log
Posted Mar 18, 2010 12:53 UTC (Thu) by pboddie (subscriber, #50784)
[Link]
He makes a good point about the physically private hosting, but I think that we're not technically far away from a complete Free Software alternative to proprietary social networking services if you're willing to have content hosting with a provider you trust somewhere. Indeed, the surprising thing is that apart from the integration, which occurs as a consequence of the monolithic nature of such proprietary services with their continuing scalability problems (and much wailing about such matters from users and the services themselves), there really isn't anything that couldn't be done right now with existing open technologies, and the integration is just that "glue code" mentioned in the article (that people have probably already written).
Of course, if you're willing to put together your own social networking experience, there's probably nothing that the proprietary services offer apart from the ability to surf the often promiscuously shared information about people in order to "connect" to them. For genuine collaboration beyond the occasional quip or shared link/image/video, such proprietary services are the things lagging behind.
I've given this area a bit of thought in the past, and maybe I should write it up in a public place, just to contribute to the debate. Sadly, many people think that competing with the big social networks is all about maintaining a big server farm, but that just falls into the trap of competing with them on their own terms: those services have to "own" all their (or rather your) data, yet a Free Software competitor could be decentralised, just as most Internet applications were meant to be.
Freedom vs. The Cloud Log
Posted Mar 18, 2010 22:32 UTC (Thu) by kragil (subscriber, #34373)
[Link]
I disagree. A few glue scripts won't work. The data has to be secure,
encrypted and redundant. Add high availability to the mix and your glue will
get really useless really fast.
There is no point in building a half assed distributed Facebook clone where
you have to do maintainance, backup and monitoring. The system should really
be distributed and fault tolerant and above all robust, secure and easy.
A distributed FOSS Wuala would be a much better starting point than some
Sheevaplugs + lots of glue.
Freedom vs. The Cloud Log
Posted Mar 19, 2010 13:16 UTC (Fri) by pboddie (subscriber, #50784)
[Link]
Well, you can get "secure, encrypted" via various hosting providers already, also "redundant" to an extent, although I concede that setting stuff up from the lower levels (HTTPS, having a "valid" site certificate, and all that) can be awkward unless you go for a provider that tries to make that easy: the distinction between generic hosting and, say, blog hosting. I note the option of using such a provider in my very first sentence.
Really, there's a fair amount of choice in content hosting, which admittedly does make it all less of a "no brainer" for those wanting to just go somewhere, sign in, then hand over their data, but there's no reason why providers couldn't adopt a common standard (and brand) for this kind of activity. Google, WordPress(.com) and other blogging providers could have dealt with the competition from social networks a lot more convincingly by leveraging their expertise in the infrastructure department and providing a platform to build upon, but the temptation for companies like Google is to own as much of the data as they can.
As for "a half assed distributed Facebook clone", you just have to compare what Facebook does and what existing technologies do to see that there's less magic than is generally believed. I'm not advocating a monster aggregator of data - that's playing on Facebook's terms, exactly as I described - nor is a physically local server essential, either - you could delegate hosting to someone you trust who offers the maintenance, backup and monitoring. Instead, you could get a long way with smart clients, just like the various desktop applications offering multi-protocol, multi-provider support: exactly as I described in my final sentence.
Interview: Eben Moglen - Freedom vs. The Cloud Log (The H)
Posted Mar 19, 2010 11:40 UTC (Fri) by sylware (guest, #35259)
[Link]
... and that open source private cloud should be packaged in small boxes that can be configured in a few clicks, maybe using some info provided by the IAP on the domestic LAN. If too complex, the lambda user will flee.
But are going to be able to do it, we are already not able to remove proprietary OSes from the desktop??