Is a Perfect Storm Forming For Distributed Social Networking?
Maybe it's better to host your own. That's the thinking coming from a growing number of early technology adopters as service after service goes down, sells out or otherwise frustrates the users who have published their content online only to see the tools they use become broken or less desirable.
The prospect of a distributed, interoperable, self-hosted network of publishing, reading and discussion tools is nothing new - but the idea is gaining a lot more support as more people react to recent news like FriendFeed's sale to Facebook, Tr.im's up and down and Twitter's denial of service attacks. The tide may not be turning, but there's sure to be some new waves of innovation that come out of this period of frustration.
Isn't This What Blogging Does For Us Already?
One of the analogies people are drawing is that we need a WordPress.org-type version of Twitter to put on our own servers as an alternative to the Twitter-hosted version that exists now like WordPress hosts blogs on WordPress.com.
Why do we need self-hosted lifestreaming, microblogging or social networks though when we've already got the ability to host our own blogs, own our own data there and set our own rules? Simply because these technologies fill different needs. Blogs are good for longer-form, author-centric communication. Quick, very social conversations around objects like links or media items can best be had in other settings. Thus the interest many people have in both writing a blog and sharing and discussing items on sites like Facebook (social networks), Twitter (microblogging) or FriendFeed (activity streams).
Twitter's Down Time
Twitter went down again today, possibly for the second time in two weeks because of a Distributed Denial of Service attack. A swarm of zombified computers, distributed all around the world, is hitting Twitter's centralized infrastructure over and over again until it can't stay up.
If we all had a little piece of our microblogging network on our own servers and they spoke to each other, that couldn't happen.
We'd also own our own data, our archives, our interface design and more. It would be like publishing little messages... like grown ups.
The two systems could co-exist, a hosted service has its advantages and many people wouldn't use anything else. Realistically, no one is going to build something too much like Twitter if they could build a distributed version of something like FriendFeed or Facebook.
Facebook Eats FriendFeed
Social activity stream discussion network FriendFeed announced that it was selling itself to Facebook yesterday and many of its users were very upset. The acquisition is likely to change Facebook in interesting ways (FriendFeed's creators were the inventors of GMail and Google Maps) but FriendFeed itself was important to its users.
The feeling of betrayal that comes from a transaction like this makes it hard to trust a hosted social networking company again.
Fortunately, there's a long and growing list of ways to put all of your activity around the web in one place on your own website. When will those tools begin to include subscription to other peoples' activity feeds and posting comments from your social lifestream viewing page that will appear back out on everyone else's?
That's a big part of the vision articulated by Anil Dash in his recent essay about what he calls The Push Button Web. It's related as well to RSS pioneer Dave Winer's recent promotion of a part of RSS called RSS Cloud. Developers are actively building on RSS Cloud and a similar protocol with the humorous name PubSubHubbub.
That's also part of the vision of the Distributed Social Networking Project (DiSo). We haven't heard much lately from this project, probably because its founders are busy building the technical standards that will allow the information to flow from one social network to another.
Tr.im Your Expectations
This weekend link shortening service Tr.im announced that it was shutting its doors. It was too expensive and hopeless to run the service without the funding, hype and official blessing from Twitter that competitor Bit.ly had won.
Big deal, right? It turns out that people freaked out. Tr.im's biggest users were developers who were hip to the opportunities to do interesting things with the service. They had built on it and they felt a lot of frustration when they heard the news.
A dead URL shortener means dead links, broken content, lost data.
There are a number of different solutions being explored in response to this part of the problem. Developer Brian Hendrickson has already begun working on a service called rp.ly, a "community-owned URL shortener" based on cloning the Tr.im API.
There will, no doubt, be any number of other efforts that rise from the ashes of the trust that's been burnt over the last week or more.
Are all of these circumstances and conversations going to push the social web over the edge, toward a more distributed and less centralized model? Probably not in a big way, immediately, but we're pretty sure that some interesting innovation is going to come out of this. Dissatisfied engineers, working on a problem that a lot of people are interested in, can produce some fun and important work.
Some will hold out for Google Wave, the forthcoming open-source hyper communication head shift. We're hearing that Wave may be too complicated, though, and we suspect that the most important innovations will come from coders building the kind of software that many, many people can hack on and help evolve.
In the future many of us may be microblogging, lifestreaming and social networking over technology that we control and can customize ourselves, instead of inside the owned networks of major companies like Facebook or Google. Those companies are seeking to branch out as well, trying to colonize the web (in the words of Forrester's Jeremiah Owyang) with tools like Facebook Connect.
But many of us may decide not to trust them anymore, and to use the tools that are becoming available to build and host our own systems of communication. People who control their own systems of communication can innovate on them outside the boundaries of the financial interests of big communication companies and we can all benefit from those innovations.
This summer is an important period in answering those questions.
Share
Facebook
Comments
Subscribe to comments for this post OR Subscribe to comments for all ReadWriteWeb posts
I wrote a sort of "Open Proposal" wiki for a Distributed microblog protocol about a year ago or so (a week before Laconi.ca came out) due to frustrations working with Twitter on my site CommuterFeed.com
Its here if anybody's interested: http://open.zenjiweb.com/Open_Microblog_Protocol
I really think the now OpenSource Jaiku needs to be repurposed as the backend for a distributed Twitter..
Maybe the Opera 10 approach is a better than you thought! It needs a killer ap of a "twitter" reader and a "twitter" push publisher.
Liked the Twitter section a lot. Does anyone think Web companies with access to cheap hardware and programming talent, while helpful for getting off the ground, can lead to shortcomings on upfront development, IT discipline/comm gaps and not being able to scale once they really start taking off?
We had a long discussion on this two years back. Check out http://blog.krishworld.com/2007/06/17/a-decentralized-social-networking-platform/
Being someone who is actively thinking about a solution and someone who is somewhat disappointed in ff sale if it ends ff.
The issue with self hosting is aggregation and discussions.
Yes it would be pretty easy to import all data but where will the discussions, likes, discovery and sharing take place?
Blogs are great but unless you have decent traffic no one will see this stuff and discuss it.
Maybe the answer is a site that acts an aweful like technorati but also enables discussion.
I am working on something that involves microformats and ping...
You've hit the nail on the head. This is going to be a growing problem. I said today on my blog that this sort of opportunities present new opportunities for startups. This has to evolve similar to how the web evolved and how blogs evolved. If you recall many years ago people would go to Geocities.com to get a website on a shared hosting service. At that time domain names and individual hosting packages were costly. Now days more and more people are able to afford their own website because it is not only cheap but more reliable. This is a similar example to how you've explained wordpress.com blog vs individual blogs. I think we can be certain that distributed social networks is not far from reality.
-Guna
Short answer: Yes.
Posted by: Gina
|
August 11, 2009 4:12 PM
Spot on re: tr.im, that's why Cleartext is building it's XMPP SaaS offering out with URL filtering and shortening.
We're delivering a new category of SaaS which encompasses enterprise messaging like IM and micro-blogging. We call this Enterprise Social Messaging (ESM).
Larger companies will start looking at ways to pull micro-blogging into their business but they need availability, security and compliance addressed first.
As Dennis said, Opera seems to be working on a browser-based implementation of distributed social networking. I think Mozilla Snowl is also a step in this direction. I imagine Google Gears might be a plausible means to store data in the browser as well.
I'm pretty sure that identi.ca has an open source self-hosted option and so does Reddit. These are steps in the right direction.
However, in order to really take off, someone will have to devise a simple, user-friendly system. Probably a modular one. It should be manageable through one central dashboard. Even FriendFeed, Twitter and Facebook features are a bit too complicated for the average person to wrap their head around. It needs to be something that can be shown and explained in everyday terms.
The future of social networking is git-based. I'm just not sure how, yet.
Posted by: Panayotis Vryonis
|
August 11, 2009 4:36 PM
There's a lot of truth in the idea that individuals and groups are wanting to take back ownership of their data, but the real problem is that of a platform. How far out is Google Wave? :p there's quite a few developers waiting to pop that cherry.
Nice post Marshall. I sure hope this gets a movement going but I'm not holding my breath. There's some pretty big hurdles to overcome. P.s. I like your use of Thomas' Lifestream screenshot for the post. I'm guessing you saw that in my latest custom Lifestream gallery post yes? :)
Posted by: Mark Krynsky
|
August 11, 2009 4:44 PM
or emacs. META-X-POST-KITTEN-PICTURE
Posted by: mikepk
|
August 11, 2009 4:44 PM
i have outlined how distributed options could work in old threads here - after all, irc, laconi.ca/identi.ca, newsgroups and jabber all work perfectly well in a distributed manner - transparent to the users but sharing/splitting the load across dozens or hundreds of machines. Protocols like that give people the option to host their own, use something provided by their ISP or local providers, and yet interact with everyone.
Posted by: Joelle Nebbe (iphigenie)
|
August 11, 2009 4:47 PM
A "WordPress.org-type version of Twitter" already exists in the form of laconica (http://laconi.ca/trac/), which powers identi.ca
it does hash tags, groups, conversation threading (context) and can be federated with other laconica servers.
Posted by: imabug.myopenid.com
|
August 11, 2009 6:05 PM
I think Facebook has bitten off more than it can chew.
It's not only a perfect storm for distributed systems but also for owning your namespace. We had the discussion two months ago with Facebook's username landrush already. Now Friendfeed gets bought by Facebook. Will you have the same username with both services? And what about the broken links from tr.im? Or maybe you'd like to switch back from Facebook to Myspace?
It would be much better to have:
- domain.com/profile for your social network profile
- domain.com/discuss for a Friendfeed-like service
- domain.com/activity to display your lifestream
- domain.com/microblog for microblogging
- domain.com/a56e for your own URL shortener
Distributed systems are the right thing but a domain-centric identity is equally important.
Good that we are working on exactly that stuff now :)
io4.biz as commercial version and io4.me (input & output for me) as private version.
Have been in stealth mode yet, but stay tuned for our first release soon.
If interested, follow me on twitter @ungerik or send me an email to mail@ungerik.net
Gina's got it: Answer is "YES!".
This is a similar example to how you've explained wordpress.com blog vs individual blogs.
In a comment i made over on dave winer's scripting.com, i referrer to a rrw post from a while back regarding standardizing url structure:
http://www.scripting.com/stories/2009/08/07/tradingOneCentralizedNetFo.html#comment-14552834
http://www.readwriteweb.com/archives/standard_urls_proposal.php
I was always fond of that concept and it seems apropos here and now.
Many ideas will be discussed in the coming months as this topic is reignited with new vigor. And the conversation is not for everyone. My wife uses twitter and facebook and tumblr in conjunction with her b&m retail store. she is not concerned with the discourse of federating the social web. she is not going to seek out software to install and run etc (thats what i am for ;) but despite the lack of concern for many people... we technologists who are interested in the social web and content creation/publishing (this all also relates to publishing video, audio, photos, words, works.. not just morsels of scattered thought output) need to be interested and need to discuss and innovate and federate and balance the ecosystem so that companies like twitter and facebook can co-exist with open (really open) distributed (really distributed) platforms.
the next 5 years will lead us to streamlining our online presence while hopefully not letting it NEED to be under the control of any one bigco. we may even end up with a government issued digital id that resolves to a domain name that branches out to our various forms of contact and content (a service paid by our taxes). companies (looking to profit) and governments and freedom of choice for digital services are part of this discussion.
sull
What sites are going to start with the Diso type technologies first? Will we actually see traction? Does laconica figure into this stuff at all?
I'd sure love to see more of Laconi.ca. I should have mentioned them here but I don't know that they are growing much. Identi.ca has been around for more than a year now and is still 1/20th the size of ReadWriteWeb. I hope they spread far and wide, I really do.
I'm surprised nobody has mentioned BuddyPress. They don't natively employ OpenID, but the fledgling community is working on it, and it will likely get a good jump start soon since it's already got a built-in community on the wings of WordPress. Andy Peatling does regular releases, and it's already come a long way since release 1.
BuddyPress is another one that could become increasingly important. I've not spent much time with it myself but I love the idea. That said, it too has been around for awhile and I'm not sure how much it's setting the world on fire. Maybe I'm wrong though. That would be great.
Google Chrome/Sync/Wave, Opera Unite, Live Framework, The Mine! Project. Go back a bit in time: Mylifebits, Memex
I wrote a book on this very subject on how 'dashboards' can be a common notion we can all use for a distributed web. It will take time, but eventually we'll have distributed friending, distributed access controls and all sort of shared servers - that no one owns, everyone can use and anyone can improve upon.
I was asked to teach a class on this at CWRU - so I wrote a second book - as a textbook for the class (EECS 396) - and I'm hard at work trying to create a pilot distributed environment of this sort of stuff in Northern Ohio.
pdf's are available of the two books:
"How to build the Open Mesh"
large pdf = http://www.box.net/shared/u65izvyipq
small pdf = http://www.box.net/shared/sdic22rbzu
"How to build a Digital City"
large pdf = http://www.box.net/shared/y40efkliqk
small pdf = http://www.box.net/shared/y7hbssnng3
How about YouTwitter? http://softwareindustrialization.com/YouTwitterNextGenTwitter.aspx
Or www.peerscape.org ? (peer-to-peer social networking)
I like Jisko. http://jisko.org/
That Twitter like features for WordPress already exists:
http://en.blog.wordpress.com/2008/01/28/introducing-prologue/
http://www.freshpressthemes.com/twitter-wordpress-theme/
Flagged - I think that's a little more like a Twitter-like UI than Twitter-like features. One install of Prologue can't comunicate with another, as far as I know.
YES
Posted by: Dobromir Hadzhiev
|
August 12, 2009 12:32 AM
Although I agree that it can be very nerve recking to to feel exposed to the big social networks like facebook and their actions and attacks on them, for the average member of these mass networks there is simply no other solution at hand.
When your knowledgeable about what there is out there to replace these networks functionalities and lets you do everything on your own space under your control, you also need to be able to use it and adjust it to your requirements.
Might be good for some people but in our experience at relenet the mass simply wants to use the servive and doesn't have the skills or the time to do it themselves.
So I guess we will have to stick to what there is right now.
There's a catch though. facebook and twitter is where all the people are. Any de-centralised system also needs search to find the others and ways of following the others. Yes, IRC and Jabber (and smtp and nntp) demonstrate how these kinds of systems can be built but did you ever try to find somebody's jabber address or the IRC room that hosted the discussion you were interested in?
Posted by: http://www.voidstar.com/yadis.xrds
|
August 12, 2009 1:56 AM
I'm glad Marc Canter left a comment.
I'm not sure how you could write an article like this without mentioning Marc Canter's work.
He was talking about open standards for this stuff since before the centralized version were even popular.
There is an OS social network system, that you've reviewed here in the past: http://elgg.org and a "WordPress.com" type of service that uses elgg at http://eduspaces.net/
I wrote recently wrote about how, short URLs could be federated:
http://blog.richardcunningham.co.uk/2009/08/federated-short-urls.html
I think this is where google's Wave might help, the massive strength of google's support, with the fact it's pretty much open sourced the protocol and theoretically everyone can have their own wave 'host', it doesn't have to be run by google. It can handle posting of different media, at the moment it seems to be more 'topic' oriented than the random stream that twitter is, but I don't see why there can't be a wave robot that aggregates twitter-style postings from people you're interested in, especially if they're using the wave protocol.
I'm not saying it's the answer to all the problems, but I think a lot of people have kind of dismissed it as complicated and a solution looking for a problem, but maybe something like this is what it will be best at, rather than trying to replace email.
But then again, it might just be a big flop...
Hi Marshall,
I think we're going to see two ends of the spectrum jostle and hopefully improve the whole lot.
At one extreme will be companies like Google/ Microsoft who will aggregate and grow larger - living out the rule of 3
At the other end of the spectrum will be decentralised systems challenging the big 3 - and pushing for simplification.
Cheers
@Anita_Lobo
*cough* Opera *cough*
But Seriously, web standards are already there we just need to learn and/or follow them and rethink their implementations from time-to-time.
Software still matters. Programming matters.
However you approach it, it's nice to have power or luxury to control your data, code, or at least have some degree of input and community-type control.
Increasingly too, we have the excess capacity to be 'smart nodes' A browser-centric vision of the future is fine, that doesn't mean the browser/client isn't or shouldn't also be robust.
I hate to Microsoft's phraseology here, but it really is a "mesh" approach.
It's not just the social web, but the application web, too. The current, central, single-point-of-failure model is simply untenable, and the web is evolving towards making a distributed web of active tools as well as content possible.
The social media tools out there are still going through their initial phases. They are getting funded but over time, they will be either consumed, dead, or on top of the hill. As soon as this phase is over with, then the better all of us will be. I am starting to see the horror stories now and rest be assured there will be more in the future. Every product goes through this, pick any and you will see the same thing over and over in history!
One huge advantage of user owned profiles is that they would make a lot of propietary networks obsolete.
Imagine if there truly were a distributed social network where users owned their own profiles and hosted them with whomever they wished. Now imagine Google created a Profile Search Engine that allowed anyone to query all these profiles at once for any information that each user chose to make public.
Instead of creating an account with the New York Times website and voting on their poll of your favorite breed of dog, for instance, you could simply post a properly formatted statement on your own profile saying something like "My favorite breed of dog is Labrador." You could then query the profile search engine and see the results from everyone. This could happen with more serious issues as well, like finding opinions on proposed laws in Congress.
Or imagine if your phone kept track of your GPS data when you were driving, and then posted it to your profile. This would allow developers to create applications that accessed true real-time traffic data from across the country, and would prevent the chicken-and-egg problem of getting everyone to use one specific iPhone/Android app for this purpose. Your profile host could also anonymize the data for you before the profile search engine could access it.
Dave Winer is right: Everything has to be able to talk to everything else, everything has to be distributed in a way that nothing is a central point that can be taken down, and everything has to be open so it can be customized in ways people don't consider in the beginning.
Dead links alone make the current Internet a frustrating place to navigate. I've had to give up keeping links on my blogs up to date because sometimes there's just no way to route around a 404. The Internet has a big fat Memory Hole.
I definitely think that at least in terms of Twitter-like activity there needs to be a distribution across multiple networks, not unlike email. In some cases the behavior is far too valuable to leave up to one company.
If for example an emergency response system wanted to use Twitter to coordinate with the public and Twitter goes down that is a big problem.
I feel like we're still in the "AOL" phase of microblogging. Eventually a protocol will surface that will open it up. What form it will take I don't know, but to be totally honest I don't think it will be identi.ca / laconi.ca - there are a number of flaws in proposals and implementation.
I feel the future will come from coordinated development with backing from some of the larger players in the field. I wouldn't be surprised if Google is putting resources into this area already, particularly with the interesting choice to opensource Jaiku
GReat article, relly gives you something to think about! The points discussed here definitely make sense after the weakness that was shown by twitter after the latest attack on that particular site!
Very informative!
Distributed makes more sense when you can sign in to the various distributed instances using a common identity. OpenID, Facebook Connect and Twitter OAuth enables this and all are fairly easy to integrate, avoiding the need for users to have a separate username and password on each instance. This is critical to this concept working.
But then I guess if you have a stack of different instances you can also tweak each to make it more relevant to the audience. For example, a service in one country may be local language, or may use a common sign in to a "verticalised" version of the service.
I can see it happening and, in fact, made this one of my predictions under the guise of "niche networks", each providing a better service for people with a special interest or specific need.
The mainstream guys with their centralised sites won't disappear, as it's likely we'll need them for idenity management and some level of activity aggregation.
Ian Hendry
CEO, WeCanDo.BIZ
http://www.wecando.biz
This is a similar example to how you've explained wordpress.com blog vs individual blogs
With the frequent death of API's, acquisitions and downtime, I think the storm is already here. Users are left in the cold, their data either lost or now "dormant" because no one can do anything with it. It happens all the time and will only happen more in the future as people launched bootstrapped services and try to get by without funding.
I'm not sure what the solution is, but like others have said - the answer is definitely "yes".
I've been pondering solutions to problems like this for years, and I think it lies somewhere in standards-land, but I've been disappointed with pretty much every standards exercise I've ever participated in. Maybe we're ultimately waiting on an innovative solution from some brilliant set of minds to fix this for us... or we're working on it ourselves.
1 2 Next