🚀 go-pugleaf

I am attempting to understand the proper use of the selector fields 
in the OSI stack: NSEL, TSEL, SEL and PSEL. In the discussion of the
ISO network layer addresses in "Interconnections", page 157, 
Radia Perlman says
"The purpose of the SEL octet is to differentiate among multiple
network layer users". On page 161 of the same book, it also states
"..an NET should take the same form as an NSAP and that the selector
value of 0 would be reserved to mean the network layer itself." 

RFC 1139 uses NSAP selectors in a similar way: "The choice of 
echo-request and echo-reply NSAP selectors is a local matter.
However, to insure interoperability, and as an interim measure
until use of the directory service becomes widespread, this memo
will recommend the following default values ...."

So, it seems, selectors allow multiple n+1 layers to make use
of layer n's services. Somehow I need to know which selector
values to use. There are several possibilites:

1. Assign values permanently, as seems to be recommended in the
quotations above. This could become yet another registration problem.

2. Agree locally before hand which values to use. This could become
an  interoperations nightmare.

3. Depend on a directory service to provide all information needed.
How far along are we on actually getting such a directory implemented?
If we do depend on the directory service, what values of the
selector will we use for the directory service?

4. And all the other possibilities....

In summary I have several questions:

1. Are there standards or defacto standards on the use of selectors?
Any detailed formation on these would be appreciated.

2. If there are no standards, how should these fields be used?

Wendy Teller
Tellabs Operations, Inc
wendy@tellabs.com
708-512-7956