[midcom] Fail over mechanism in midcom

["Abdallah Rayhan" <arayhan@nortelnetworks.com>]

Fail over mechanisms, IMHO, should be addressed as part of the
midcom architecture. By fail over I mean two things:
    1- external agent going out of service and handing the
       service off to another agent.
       
       Should the middle box trust the new agent and
       impose security challenges on it?
       
       Should state-transfer be assumed between the old agent
       and the new one?
       
       There are issues related to provisioning, such as
       should the middle box be pre-provisioned to authorize
       the new agent to engage in signaling? etc...

    2- middle box going out of service and handing the service
       off to another middle box.
       
       In addition to the issues in 1, we probably need to
       identify how to handle existing pinholes when service
       change occurs.   

Obviously there are issues that needs to be addressed in each
case to make such handing over seamless to the apps/users
that is if we decided to support this feature.

regards
Abdallah

_______________________________________________
midcom mailing list
midcom@ietf.org
http://www.ietf.org/mailman/listinfo/midcom