id != address. An address is used in routing, and therefore must include the locator.
IMHO, this definition doesn't define anything. On Jun 11, 2010, at 6:00 PM, Dae Young KIM wrote: > On Fri, Jun 11, 2010 at 5:48 PM, Scott Brim <scott.b...@gmail.com> wrote: >> Fred Baker allegedly wrote on 06/11/2010 01:02 GMT+02:00: >>> This one gives me some pause in the context of a virtual machine in a >>> cloud computing environment. I rather like the idea of associating >>> identifiers with an application or set of applications, and only by >>> extension of that referring to physical or virtual machines. For >>> example, in cloud environments, if a process or set of processes >>> moves from one machine (physical or virtual) to another, it would be >>> nice to be able to move all of its/their sessions with it/them. >> >> Fred, you're right, I hadn't noticed that. Identifiers are used in many >> ways, particularly for different session instantiations. The definition >> here has a particular scope, and the scope should be explicitly stated. > > Hi, Scott and Fred, > > How about looking at the thing in this way? > > o There are various identifiers in various places, vertically or > hierarchically: > > - APP ID > - Node ID > - Subnet ID > - Site ID > - ISP ID > > o They're floating over each other: > > - APP floats over Nodes. (process migration over machines) > - Node floats over subnets. (node mobility) > - Subnet floats in a site. (subnet mobility) > - Site floats over ISPs. (ISP migration) > > And, as long as these ID's are used in moving around > (migration/floating), they are equivalently called 'address'. > > -- > DY http://www.ipinc.net/IPv4.GIF _______________________________________________ rrg mailing list rrg@irtf.org http://www.irtf.org/mailman/listinfo/rrg