#39: Address and OII collision text --------------------------------+------------------------------------------- Reporter: [email protected] | Owner: [email protected] Type: defect | Status: new Priority: major | Milestone: Component: nd | Version: Severity: - | Keywords: --------------------------------+------------------------------------------- Improve the address and OII collision text, and the coverage of cases.
From list discussions, there are some special cases where the current nd-03 OII collision detection text does not work well. List mail from Pascal: The text can definitely be improved. But Mostly we need something that works with all cases covered. Section 7.6 Address collision detection and resolution Seems we are missing text when the binding is for the same edge router different OII. This is a collision and causes a reject exactly like NA (different OII) during DAD. The text also says that upon positive re-registration, the edge router SHOULD set O in NA. This is misworded, in fact it SHOULD send NA(O) as a refresher and eventually a tie breaker if for some reasons (missed DAD) incompatible states have setup. An edge router with a registration should clean up at that point. Section 7.7 Duplicate OII detection As Richard said we need to be more precise on equal TID. That must be treated as a collision otherwise we'll flip flap forever. But then is if there is a risk of a packet getting duplicated over the network we might be doing the wrong thing. Would we need a nonce to discriminate that? Also, in the most pathologic case, if a packet sent to anycast is duplicated, can it reach 2 different packets received by 2 different edge routers? -- Ticket URL: <http://trac.tools.ietf.org/wg/6lowpan/trac/ticket/39> 6lowpan <http://tools.ietf.org/6lowpan/> _______________________________________________ 6lowpan mailing list [email protected] https://www.ietf.org/mailman/listinfo/6lowpan
