Peter Nixon said:
Yep. The problem I have with this change though is that what are the other vendor disconnect keys? If you read the RFC it doesn't specify a specific key (Alan has been holding ongoing discussions about this on the ietf- radext list), just that there should be enough info to identify the correct session. As Cisco also uses it (and Juniper from AFAIK) XAscendSessionSvrKey has become a de-facto standard along with MS-Primary-DNS-Server and its ilk. Given that it is possible to send and have work a Disconnect-Request containing only a User-Name and there is no-where specified that there should be an individual session key I am not sure its worth changing unless someone can come up with a concrete example of a vendor who uses a session key thats different.... Is there?
I don't know of any, but as more gear starts to support PoD / CoA / Disconnect-Message / whatever, I'm sure they'll crop up. Mark my words, the Atlanta Scientific Tanning Salons and Wireless Switches latest model TanNSurf-5000 will implement TanNSurf-Disconnect-Key at some point. I guess I just don't like contributing to the de-facto standardizationificationism of vendor specific attributes. And besides which, XAscendSessionSvrKey is a biatch to type. :) Your call. You (one of) Da Man.
Peter Nixon
-- hugh