Earlier, Toni Stoev wrote:
> Issue #2, technical: Identity/location separation with locally unique 
> identifiers
> 
> On Sunday 13 June 2010 at 00:30:00 Tony Li sent:
> > > What happens when a node with a locally
> > > unique identifier in a subnetwork moves to another subnetwork
> > > where there is another node with exactly the same (locally
> > > unique) identifier? How are the two nodes distinguished, and how
> > > are their ongoing sessions preserved?
> > 
> > You are correct, locally unique identifiers will have mobility issues and
> > hosts should not use their locally unique identifier outside if its scope.
> 
> There is no identity/location separation with locally unique identifiers.

I think we disagree.  

Even with local-scope Identifier values, the Identifier
does not name a location && the Locator does not identify
a node.  So the concepts of identity and location are 
crisply separated, with clear semantics for each.

Another way of examining this might be to consider that 
with ILNP, routing only uses Locators, while upper-layer 
protocols (e.g. TCP/UDP pseudo-header checksum; IPsec 
Security Association) bind only to the Identifiers, 
never to the Locators.  Again, this shows clear separation 
of location from identity, and this is also true with
local-scope Identifier values.

Again, this is a major architectural change from IPv4/IPv6, 
where the upper-layer protocols (e.g. TCP/UDP pseudo-header 
checksum, IPsec Security Association) bind to the entire 
IP address and misuse it as a node identifier.  The issues
the deployed world has with things like NAT or various 
Mobile IP approaches or even (indirectly) site multi-homing
are because location was incorrectly bound up in the 
upper-layer protocols when those protocols only really
cared about node identity.

Yours,

Ran Atkinson


_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg

Reply via email to