Roy Fielding: What is the state of this within the IETF?  Is there any intention to move that forward as an informational draft?

Sam Sun: Fix the design before the draft.  We are pretty much done with our design framework.  Timeframe is at first to submit as a RFC.

Q: You claim the architecture of the Handle system is better than DNS. Both seem to be addressing scalable independent names that wont conflict.  You deploy faster but wind up with the scaling issue.  It's misleading to say it's better.

Sam Sun: It's a trade-off. The client does some processing instead of relying on the server. Not the same model as thin client.
Handle provides a service that provides more information than just a server.  Trade off is more work.  But it allows more servers to service the namespace.

Q: Handle is only available as a binary release.  Is there any intent to make source available?

Sam Sun: Binaries for local service is suppose to be source. And for the client.  People had trouble building them so we distributed the binaries instead.  There is an issue on how much of the total code we distribute.

Sam Sun: The software itself relies on a single root service, single root service can be replicated and distributed within a site. Primary service.  Multiple primary sites.  Registration can take place at multiple sites

Daniel LaLiberte: Single admin root? How can the whole namespace be divided up administratively?  For example, UCI can service everything in global sense but why should people trust US servers to do the right thing?

Sam Sun: Lots to do with the policies issues.  Issues of ownership? When two departments merge, what to do?  We don’t have a solution.