VPTN-3 has been submitted to the emerging IETF working group MMOX for
consideration as MMOX data protocol:
http://tools.ietf.org/html/draft-wolf-vp-identity-00
The protocol is proposed as input to the MMOX BoF and WG which is currently
being chartered. It shows, that the main point of interoperability, namely the
instantiation of avatars can be achieved with a very simple yet powerful
protocol. It shall serve as a working showcase of light weight interoperation
and as a reference point for the discussion.
The protocol is Web/HTTP based, actually the transport is HTTP, the container
format is XML and avatar data can use any MIME type. The architecture is
inherently distributed by using basically XML to organize asset URLs.
The protocol has been used for 2 years with 3 Mio. users and 25.000 concurrent
clients. The primary purpose of the protocol is to enable "simulators" to
instantiate ("rez") avatars of users they meet in a virtual world. The protocol
has been designed for interoperability, specifically to be able to use avatars
from closed virtual worlds on the Web. But it is equally suited to bridge from
one world to another. Actually, the Web is regarded as just one virtual world
with many regions, which are commonly called Web sites.
The protocol also offers a solution for the "foreign updates" problem, where the
avatar changes in one world and the change should be displayed instantly in
another world. Say your WoW avatar walks in SL and an effect times out. That
should be visible in SL with minimum delay but also without maintaining long
lists of subscriptions. The protocol has a lean way to communicate changes.
The protocol addresses the "the low hanging fruit" of simulator
interoperability. But it is easily extensible, e.g. to virtual goods ("dragon
heads"). It supports variants of avatar models if multiple formats are required.
It provides trust using XML Signature (not specified here, but in use). OAuth is
the canonical choice for access authentication and selective disclosure of
identity data.
I hope that the task of the MMOX WG will be restricted to interoperability
between simulators and asset services, and that it will not include the
communication between display and simulator. There are so many different models
how worlds structure their communication between display and simulator. In some
systems the world simulation runs in the client. In many 3D worlds the simulator
is regarded as a "server" which forwards scene changes to the display. Games
like WoW do much more simulation in the client than Second Life. In the near
future there will be simulation including rendering in the cloud with only thin
displays. For this reason the MMOX WG should only work on simulator
interoperability and keep the display- simulator communication out of the scope.
|