top posting:
- it is a generally accepted notion that some sort of identifier(s)
needs to be added to the existing architecture (I'm writing up a short
notes on my own understanding of why now, and why *not* day one)
- since god did not write us a bible on how to design networks, we
have to learn as we go. And one of the things I have learned from
last 30 years: never be too sure about the future's need. A winning
strategy seems including the following ingredients:
+ start simple, and
+ stay flexible, keep the mind open.
Along that line of thought: we need to understand the interplay
between address (used for routing) and this forthcoming identifier
thing, but I would keep the focus on address/routing, not nailing down
the exactly numbers and definitions of identifiers.
Lixia
On Jun 26, 2009, at 10:24 AM, Christian Vogt wrote:
Dear all -
I am catching up on folks' previous discussion on identifier
properties.
This discussion is really useful; I wish I had participated earlier.
One thought that has been brought up is the inevitable coexistence of
multiple types of identifiers. This makes sense given the variety of
things to be identified. In fact, the multitudinousness of
identifiers
is not new: Many applications have their own identifiers already
today.
But are all these identifier types essential elements of an Internet
architecture? I would argue most of them are not -- they are useful
within the scope of a particular Internet application, but they are
not
essential for the Internet per se to function. In fact, I see only
two
purposes for which the Internet architecture must have identifiers:
(1) service identification, identifying a piece of communication
software that responds to incoming contact establishment attempts
(2) session identification, identifying the protocol state
corresponding
to a particular session after contact establishment
Individual Internet architecture solutions may use combinations of
more
than one identifier for either of these two purposes. For example,
service identification in the existing Internet is achieved by
combining
a host identifier (DNS name, or IP address in its role as a host
identifier) and a host-local service identifier (well-known port
number). So an Internet architecture may have more than two types of
identifiers. But again, it seems that all of those identifiers would
serve only the above two purposes, after all.
Given the distinct identifier purposes, the properties of identifiers
may naturally be specific to a particular purpose. For example, the
property of enabling session referrals clearly applies to only session
identification, but not to service identification. On the other hand,
stability is a property that is more important for service
identification than for session identification, since two hosts
engaged
in a session may be able to handle changes of session identifiers.
Thoughts?
- Christian
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
_______________________________________________
rrg mailing list
[email protected]
http://www.irtf.org/mailman/listinfo/rrg
