[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: GSE IDs [Re: IETF multihoming powder: just add IPv6 and stir]

Christian,


The focus on the site misses an important point.

I thought this working group was working on site multi-homing.

On Monday, May 5, 2003, at 03:11  PM, Christian Huitema wrote:

Computers are often
multi-homed to several sites, e.g. WiFi and GPRS.
Terminology issue.

According to the working group draft:

A "site" is an entity autonomously operating a network using IP and,
in particular, determining the addressing plan and routing policy for
that network. This definition is intended to be equivalent to
"enterprise" as defined in [RFC 1918].

WiFi and GPRS are access methods. A network provider could offer connectivity to a "site" via WiFi or GPRS.

However, with that said, presumably the device you are thinking of is a dual-mode cell phone/PDA or equivalent. As I have said before, a "site" in such a case could be defined to be the PAN (of perhaps no devices) connected to the Internet via the cell phone/PDA or equivalent, connected via WiFi and/or GPRS.

If you do something as
radical as changing the behavior of TCP, then you want a solution that
handles host multi-homing as well as site multi-homing.
If you are doing something as radical as changing the behavior of TCP, then there are a lot of things that could be fixed, e.g., the layering violations of having the network address be aware to the transport and application layers, creation of a real session layer, etc. etc.

However, I don't think this would be the right working group (or even the right area) to try to come up with (Transport Layer Protocol)++.

You should not force multi-homed computers to attach the same
"global identifier" tag in all of their IP addresses, and you may expect
privacy advocates to forcefully remind you if you ignore that point.
I think this is a bit overblown.

Privacy advocates went non-linear on the idea that _individuals_ could be identified by the traffic they originate or receive due to IPv6's stateless auto-configuration using the EUI-64 associated with hardware interfaces on the user's computer. I do not believe they were concerned that _sites_ could be identified. If they were, I would imagine they'd be pretty unhappy with IPv4.

Rgds,
-drc