OK - so as I’ve been going through and spending the time to write up these different areas of “How to build the mesh” and what I was hoping for - paid off. [1], [2], [3], [4], [5], [6], [7], [8]

I actually started to grok how to build the open mesh.

Not Microsoft’s mesh which is controlled by a single vendor. But an open mesh which is made up of vendors, standards and glue code - that connects a wide range of services, applications and platforms together.

To pull this all off - we all need to focus on open standards and how the magic of people working together can make great things happen.

Now I’m not saying that there aren’t great people working at Microsoft, Google or Yahoo - what I’m trying to say is “we can’t rely upon those BigCos - or even Evan Williams to build our open mesh infrastructure.” Individual platforms are going to be always looking out for themselves.

So our efforts at establishing an open mesh must be predicated in the spirit of open cooperation - where we think about the good of the whole FIRST then on how our individual companies or efforts can prosper.

We must architect for redundancy.

This is how the DNS works - and it’ll work for the open mesh as well.

So that means that there won’t be one way to do IDs and privacy.

There won’t be ONE reputation system or feed format. Or one way to keep track of lists of people. Or control access to your images.

We won’t have one way to subscribe to people, import our social graphs and synchronize our profiles.

We certainly won’t have even five standard ways of ‘defining what a group is‘ - but lord help me - we WILL be able to aggregate them all together! (That’s a bone to you Steve.)

What’s complicated about all this is that for the open mesh to evolve, it has to AT THE SAME TIME be implemented and built with all sorts of agreed upon standards, protocols and formats (both proprietary and open), cooperative big players and the commodities of scale in place (so we can all afford it) WHILE we resist the urge to overly rely upon a single source for anything and embrace our competitors as if they were our own.

There are some existing standards that we can take for granted, but lots of trade offs, compromises, ego battles, etc. will have to happen before all of the ideas (I list below) will come to fruition.

In the mean time the open mesh will grow organically under ‘natural evolutionary‘ laws - where proprietary technology, lock-in business models and closed minded thinking (the status quo today) continues to operate side by side with more open, shared technologies, attitudes and practices.

This evolutionary hybrid state of the open mesh will continue until BigCo players start to see the benefit of directly supporting and contributing to open mesh principles.

Once enough big players are making money comfortably residing directly in the open mesh (while maintaining their family jewels and assets) other BigCos will join us - as well.

But I sure hope folks out there don’t think that Microsoft is gonna roll over and just hand over their mesh family jewels - in open source form? Our mesh will have to grow independently of Microsoft’s. And Google’s. And Yahoo’s. And just because Ray Ozzie and company choose the MESH word - doesn’t mean they own the idea - either.

There are plenty of independent, state-of-the-art companies out there to build the mesh - independently of the BigCos.

IM clients, Twitter and FriendFeed services and blog tools can all agree upon common ‘backbones’ to federate IDs (OpenID), share authentication procedures (oAuth) and make it easy for users to keep track of all their ‘destinations’ (OutputThis.)

Other standards will emerge for….. (well just check out the list - below. You’ll get an idea of what’s needed.)

We have to make sure that there is no one single point in the mesh where if it breaks down - we’re all fucked. Or put better “no one vendor gets to dictate - anything.”

We’ll just route around them.

Twitter is a great example of relying upon a single infrastructure - and what happens. Great compelling new services (like Twitter, Friendfeed, what’s next?) change people’s lives - and when the infrastructure that so many people have grown to rely upon - fails - then…… what?

So that means that all great open mesh services HAVE to have multiple vendors. Cooperating together.

HAVE to share their membership databases of people in a DNS-like open environment.

HAVE to rely upon services which rely upon services. eg. a recursive set of dependencies which forms a redundant layer of 24/7/365 uptime.

We have to have redundancy throughout the mesh - so if somebody’s father-in-law gets pissed off and shuts down your company - we don’t lose all our coffee house reviews or access to the coolio t-shirt web service.

If you look at Infrastructure and large companies - this is where they think they can lock us in, milk us as fat cows and continue to exploit us as ‘consumers’. But the mesh offers an opportunity to level the playing field, to make it possible for small software developers to make a living off of running software services which provide compelling experiences and functionality to people.

The hooks INTO the mesh - are the crucial touchpoint - as this is where the small guy gets access to the big league infrastructure.

So the ultimate tradeoff we have to make - at the moment - is to learn to live with the BigCos and exploit them for our own ends. We have to warp, manipulate, infest, conspire and work for/against the BigCos ambition to control the mesh. Once we’ve reached that compromise state - we can then break off and do our own thing - but I predict that we’ll always have ‘BigCos’ in our lives.

YouTube turned into a BigCo - so did Facebook and soon - Twitter. Once a SmallCo turns into a BigCo everything changes - so we have to design our mesh to CONSTANTLY be morphing itself, reinventing itself, scaling up and out - while embracing the BigCos and acting as a petri dish for SmallCos to turn into BigCos.

But there still will be 100,000s of others who never get beyond TinceyTinyCos.

Infrastructure is EVERYTHING. The open mesh needs to have an infrastructure that is protected and separate from any BigCo control. This sort of large scale infrastructure will become the nervous system, blood system, skeleton, muscles and bones of the open mesh.

So we can’t (by definition) have Infrastructure that’s JUST provided by BigCos. This is where Evan Williams and Twitter are on OUR side (until they get bought - of course.)

Infrastructure has to be broken up into as many tiny pieces as possible.

Infrastructure has to take the form of feeds, shared servers and simple ID protocols and formats. It has to be so elusive and slippery - it can’t be controlled, tracked or stopped.

Infrastructure is tagging. Tags will be utilized to aggregate in real-time and parse everything and store it away into the proper categories, demographics, contexts and regions.

Needless to say we need a more malleable term and approach to Infrastructure than something that stands for plumbing, walls and highways. But I won’t go off on a terminology tangent right now.

Here are a bunch of Infrastructure ideas that could help define and build the open mesh.

- RSS - Activity streams - Friendfeeds (we call them ‘Rivers of People’)

- Real-time talk, VoIP, IM, casting, twitting, presence channels (we think of them as real-time blood)

- Dataportability - the ability to move your stuff, wherever, whenever with access privileges under your control

- Interoperability - open two-ways APIs which facilitate two disparate systems sharing commands, verbs, actions between each other

- Media - not just streaming media or downloaded media - but media that’s available all the time and becomes infrastructure. Structured media that knows about what it is, can describe itself and list out who’s on the video, what the lyrics are and it’s length and copyright holder. My media that I’ve bought and media that is stored as well as pointed at. Wrap all that up - and make it a pervasive construct - and we’ll start to get close to how important media is.

- Routing - things that I create and where I want to put them - there are a lot of places that I want to send my creative output. There are also a number of destinations that one collaborates in, with others. And official places - like everything from the tax man and local city art contests to official promotions from Coca-Cola or Verizon. There are a lot of places and destinations where people will want to put stuff - and that’s a whole thing unto itself.

- Addressability - so as I’ve been writing up this series of blog posts Blaine Cook contacts me to clue me into some players in the Live Web space (updated #4 accordingly.) But he also dings me to the importance of ‘addressability‘. By that he means: “finding a standard way to identify and/or address a specific person. ” So absolutely once we solve this - that becomes another major aspect of the open mesh!

- Opt In controls - I really wanna respect Facebook’s obsession on privacy, though I suspect they’re really hiding behind the shroud of privacy for their own good. One solution I proposed when the Robert Scoble scandal exposed issue of ‘moving your data out of Facebook’ broke - was for Opt In Controls. Lets give every user the ability to control who sees their data. This means that all data has to have access controls on it. How’s THAT for a vision of infrastructure? Can you imagine how complicated that could get?

- Who I am - where I am - what I’m doing - targeted ads - my current status - not sure what this is called - but it’s the pot at the end of the Rainbow. Whoever controls this - controls the purse strings.

NOTE: OK folks - that’s 7 - now back to working on my mural.

This entry was posted on Thursday, May 1st, 2008 at 3:24 pm and is filed under . You can follow any responses to this entry through the RSS 2.0 feed. Both comments and pings are currently closed.