On Thu, Nov 19, 2009 at 5:23 PM, Dae Young KIM <[email protected]> wrote: > On Fri, Nov 20, 2009 at 12:02 AM, Patrick Frejborg <[email protected]> > wrote: > >> Unfortunately it would require changes to some applications - but it >> is the right place to fix the problem. If the application can not be >> changed and mobility is required, then use Mobile IP. >> A simple session layer would make the networking so much easier. > > Introducing the session ID would require a lot of change in the current > Internet operation. In the attempt of my proposal, I'd tried to minimize the > disturbance to the current Internet infrastructure. So, I chose to make > compromises even if it would not perfectly match a more ideal architecture > of my belief. If this discussion would be with the Future Internet > community, I'd have proposed a more drastically different idea. > Actually, the session ID is not that far away :-) If you look into SCTP, there is a verification tag that have the characteristics of a session identifier. Also Multipath TCP has similar functionality, it is called a token. SCTP is available, MPTCP is under construction. And both transport protocols are supporting multi-homing, i.e. the endpoint can have several subflows to the remote endpoint over different interfaces. If you make some changes to the routing architecture you could ensure that the subflows are routed over different attachment points between the local network and Internet - multi-homing service is no longer provided by the network, the endpoints take cares of that themselves. And the routing architecture becomes simpler, also endpoint mobility is an extra bonus.
-- patte _______________________________________________ rrg mailing list [email protected] http://www.irtf.org/mailman/listinfo/rrg
