Slashdotted (We're not centralized!)

Wow - we just got Slashdotted! (And our servers seem to have withstood a sustained load over six times what their previous peak had been - whew!) Anyway, it seems clear that our messaging around how we work has got some holes in it...

First, it appears many people think that this is a centralized system. Actually, i-brokers, which are based on the open OASIS XRI, XDI and SAML standards, are not centralized. Well, they are now only because there is just one of them, but we have a project underway to package our code for release - under the dual GPL/BSD license - by the end of the year. At that point in time anyone can be an i-broker. As our business model is based on us being trustable (as people can easily move to any other i-broker if ours loses favor) it'll simply have to be so - and provide an excellent set of services that cause people to stay with us.

Also, the fact that there is only one global i-broker may lead to some misunderstandings. While it is true that there are only two (primary) global namespaces - rooted in the '=' (personal) and '@' (organizational) global context symbols - there are two other types of namespaces that open the system wide open. Community i-names, usually rooted with an organizational '@-name', allow each community to create its own local namespace. For example, I'm "broadcatch" in the Slashdot namespace. Every community can create as many local namespaces as it wants to - for free!

The other primary global namespace, which opens things even wider, is the concept of rooting a namespace on an XRI cross reference. This will allow roots based on cryptographic public keys, among other things, and will be ideal in peer-to-peer systems as a vehicle for creating truly decentralized i-names.

Finally, besides the fact that i-name users have i-name portability that enables easy movement between i-brokers, the data that the i-broker manages for them can reside anywhere - including on their personal systems. There is no requirement that the data be stored at the i-broker - in fact, a basic i-broker has no mechanism for such data storage!

More later...

Comments

because of the slashdot post,

because of the slashdot post, I read up more on inames and got one. I like the idea and wanted to support it.. very interesting. now if other organisations would get on board..... :)

=stu

Thanks. The Slashdot article

Thanks. The Slashdot article really helps me to see how our message can be easily misunderstood. It's great to have such a well-educated group as Slashdot tear into you - especially when you have good answers. Maybe next time we'll have them on the site... :-)

hi Fen, I'd like to see

hi Fen,

I'd like to see (e.g., on the 2idi site) examples of community i-names and cross-reference XRI rooting.

I know 2idi is doing something more specific than everything that's possible with XRI, but, even having thought about XRI, i-names, and registries like 2idi for a while, it's hard to picture the contexts that are hoped to emerge around i-names.

In other words, where am I going to be able to use my i-name, and how/why might I have more than one i-name, how would they then relate, etc?

Also, (at least personally) I'd like to know more sooner than later about how we'll be able to build our own i-name host / i-broker. Once I have the software on my server, how does it work that the server is recognized as a i-name host or i-broker?

=jayf

Hi Jay - We'll be expanding

Hi Jay -

We'll be expanding the description of how community i-names work as we roll out the ability for communities to register them - which should be availability before the end of the year (or early next year). As multiple i-brokers and XDI data sharing are not yet available, cross references are not useful yet, either.

The reason it seems like we're "doing something more specific" is because we're strapped for cash and don't have enough developers - but we'll get it all happening by the end of the first quarter 2005.

And we're packaging the i-broker code now for release - and we have two projects underway with volunteer developers to build the XRI 2.0 resolver that we need to make them work. We will absolutely keep you informed - as soon as we know! In the mean time, you might want to look at our help wanted page at http://wiki.idcommons.net/moin.cgi/HelpWanted