On Apr 2, 2009, Tony Li wrote:

locator    A locator is a name that has topological sensitivity at a
           given layer and changes if the point of attachment at that
           layer changes.  [...]

identifier An identifier is the name of an object at a given layer;
           identifiers have no topological sensitivity, and do not
           have to change, even if the object changes its point(s) of
           attachment within the network topology.  [...]


Hi Folks -

Since "topological sensitivity" is key to distinguishing identifiers and
locators in these definition, I believe we should be more specific with
respect to what we mean by it.  Much of the previous discussion was,
IMHO, due to people having different understandings, as exemplified by
the discussion on whether MAC addresses are identifiers or locators.

In particular, "topological sensitivity" could be defined as either or
both of the following:

(1) as the property of being used in path selection.

(2) as the property of encoding a path, strictly or loosely.

If we use definition (1), the categorization of a name as identifier or
locator depends on the use of the name.  Overloading, e.g., can turn an
identifier into a locator.  On the other hand, if we use definition (2),
the categorization of a name is independent of the use of the name.

The earlier discussion between Tony and Dino on whether MAC addresses
are identifiers or locators exemplifies the need for more specificity of
what we mean by topologial sensitivity.  Dino argued that a MAC address
has topological sensitivity, and hence is a locator, because MAC
addresses are used for path selection in switches. He is absolutely
correct if we define topolocial sensitivity as in (1) and consider
switched Ethernet.  Tony argued that MAC addresses should be considered
identifiers because they were originally not used for path selection. He
is absolutely correct if we use definition (2), or if we use definition
(1) and consider traditional, non-switched Ethernet.

To avoid misunderstandings, I would hence suggest defining "topological
sensitivity" -- either as a separate definition, or by making existing
definitions more specific.  Personally, I am in favor of definition (2)
because it provides terminology that remains stable independent of usage
and overloading.  Obviously, other folks' mileage may vary.

- Christian



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

Reply via email to