New kinds of Infrastructure
Three new services have launched recently which provide infrastructure for our open mesh of the future. These services all share in common the notion of ‘digital lifestyle aggregation’ in that they incorporate features which connect together, route or provide a backbone for communication between apps and services.
Most importantly these three new services establish models which can be used by OTHER infrastructure providers, including (theoretically) the BigCos (we all know and love so well) Google, Microsoft, Facebook, MySpace, Bebo, Yahoo, etc. But they all also fall short in the idealistic checkbox list of features I always apply to infrastructure plays.
Identi.ca is an open source implementation of a Twitter-like communication platform. The open source aspect tries to solve one of the greatest hangups we’ve all experienced first hand when we put our collective balls in a single-vendor vice. So Identi.ca tries and solve sthis by making it’s source code available. Coolio dudes - thanks. But as Mark Hopkins points out without being able to connect with other instantiations and create a DNS-like backbone of federated IDs, then Identi.ca falls short of the idealistic ‘decentralized Twitter‘ we’ve all been talking about.
Ping.fm is a routing service, enabling me to send my posts (whether they be text blog posts, media uploads or messages) to any of my tools, apllications or services aI use. We call these services “destinations’. As cool as ping.fm is, I’d like to request one more coolioicious-ness. How ’bout making that list of services (destinations) a standard file that is shared between other routing services? I as user don’t wanna have to maintain mutiple copies of these lists and have to add to each one, when I add an additional destination.
So the list of services that FriendFeed utilizes should be the same list that ping.fm utilizes. We call that service ‘OutputThis‘ and it ALSO has to be open source and not controlled by any one vendor.
Finally we have Gnip - a new infrastructure for pushing updates to data consuming sites - the aggregators we all have grown to love so well. Give credit to Facebook for popularizing their ‘News Feed’ service and FriendFeed for turning ’subscribing to people’ into a new artform. Gnip helps solve the scalability problem, so that data consuming services don’t have to be constantly pinging the data sources for updates. Give credit to Dave Winer for inventing ping services and Matt Mullenweg for taking ping to a new level. Now we have Gnip which has turned this problem into yet another piece of our infrastructure future.
But where’s the ‘Identi.ca of Gnip?” We can’t lock our collective selves into yet ANOTHER single vendor solution! What if Gnip goes down? So as perfect as Gnip is, it needs to take a clue from Identi.ca and provide an open source version of it.
And Identi.ca needs to get a clue from the DNS servers that drive the web and provide a way for ALL our social graphs, IDs, personas, etc. to share and federate together - to avoid single vendor lock-in. Maybe we can use OpenDNS?
Between Google Friend Connect, MySpace Data Availability, Microsoft’s Live Contacts and Facebook’s dynamic privacy - we’ll soon have plenty of technology and methodology to move our social graphs and relationships between systems.
So the ‘lock-in’ that we’re all experiencing with Twitter will go away.
These new kinds of infrastructures are important as they establish important precedents. But we can’t fall into the same trap that we have with Twitter. We MUST have redundant solutions, even if they’re not identical. Or Identi.ca or Identi.it or Identi.sex.
But it’s becoming clearer and clearer what our infrastructure will look like moving forward. But all these infrastructure innovators have to realize that as coolio as they are, they’re not the only game in town - and won’t be - moving forward.

July 13th, 2008 at 2:24 pm
Yo Marc, give me a call. Your phone goes straight to voice mail.
July 13th, 2008 at 2:37 pm
But Identi.ca DOES support a federated system of messages.
July 13th, 2008 at 3:03 pm
[...] « New kinds of Infrastructure [...]
July 13th, 2008 at 5:33 pm
I have some ideas Marc…
July 13th, 2008 at 9:01 pm
Hey Marc, I’ve followed your open mesh thoughts from some time now. I am not sure how to enter the dialogue, but I did want to encourage the project. I will jump in as I can.
I set up the Global Social Media Discussion room on FriendFeed to see if there was any “platform” there. Not sure yet, since I am the only poster. Good number of respected friends have joined, so we will see. I picked up a funny idea of the FFoFF platform. FriendFeed on FireFox. And in a recent discussion on that topic I began sketching out a visualization of the UBER Social App requirements. Now perhaps it is really FireFox 6 and some very slim and simple plugins, ala 37 signals philosophy.
Here is a link to V1 of the graphic for any input you might have. http://bit.ly/3p703j
Open Mesh here I come.
July 14th, 2008 at 3:47 am
Nice Marc. A simple analysis of the where we are and where we need to be going with infrastructure 2.0.
July 14th, 2008 at 7:27 am
[...] David Ulevitch (of OpenDNS and EveryDNS) has some ideas on how to help enable the open mesh. [...]