Sam,
(2012/11/07 23:14), Sam Hartman wrote:
"Yoshihiro" == Yoshihiro Ohba <[email protected]> writes:
Yoshihiro> As I spoke today, I prefer having informative text on the
Yoshihiro> retransmission issue without using a requirement language
Yoshihiro> such as MUST/SHOULD. This is because a lower-layer can be
Yoshihiro> designed in multiple different ways to deal with silent
Yoshihiro> discarsion of EAP messages, including immediate
Yoshihiro> termination of the lower-layer session, and it is
Yoshihiro> difficult to determine what the best way is. In general,
Yoshihiro> it depends on the characteristics of the target
Yoshihiro> application. For example, an application that is capabile
Yoshihiro> of server-driven retransmission can simply rely on
Yoshihiro> EAP-layer retransmissions to deal with this issue.
I strongly agree with the above.
I'd like to use normative text in the following wway:
Applications specifications SHOULD describe how they handle EAP
retransmission.
Use of normative text in this way is good. On the other hand, I think
there is a room to improve the text.
Two comments:
- I think the text is applicable not only for applications but for any
EAP lower layer.
- I suggest to be a bit more specific about what actually should be
described. IMO, what should be described in each lower-layer
specification is how to handle the case where an EAP message is
siliently discarded by EAP peer or authenticator layer.
Suggested text:
Lower-layer specifications SHOULD describe how they handle the case
where an EAP message is siliently discarded by EAP peer or authenticator
layer.
Regards,
Yoshihiro Ohba
I think it would be undesirable to use normative language when
describing which retransmission strategy applications should use.
--Sam
_______________________________________________
abfab mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/abfab
