Re: A list of issues for RFC2462 update

["JinHyeock Choi" <athene@sait.samsung.co.kr>]

Dear Vijay

Thanks for bring this up. 

Vijay Devarapalli wrote 
> hi,
>
> here is another issue. it involves both 2461 and 2462.
> 
> RFC 2461 says
>
>   Before a host sends an initial solicitation, it SHOULD delay the
>   transmission for a random amount of time between 0 and
>   MAX_RTR_SOLICITATION_DELAY.
>
> RFC 2462 says
>
>   If the Neighbor Solicitation is the first message to be sent from an
>   interface after interface (re)initialization, the node should delay
>   sending the message by a random delay between 0 and
>   MAX_RTR_SOLICITATION_DELAY as specified in [DISCOVERY].
>
> lets assume a Mobile Node moves and attaches to a new
> link. it does router discovery and configures a Care-of
> address. the mobile node cannot send a Binding Update
> until it completes DAD for the Care-of address. these
> two random delays (before router discovery and before
> DAD) contribute a lot to the movement detection delay.
> I think this needs to be fixed.
> MAX_RTR_SOLICITATION_DELAY is 1 second. taking the worst
> case scenario, it could be 1 second before a sending
> router solicitation, one second before sending neighbor
> solicitation for DAD and then 1 second before DAD
> completes. the Binding Update cannot be sent until then.

Indeed these cause long handoff latency. And there is an another random 
delay before sending RA.  
 
> we had a long discussion on the Mobile IP mailing list.
> some argued that this needs to be done only when the
> interface is initialized and not when the mobile node
> attaches to a new link. others argued that this random
> delay is essential to avoid a DAD storm if a bunch of
> mobile nodes move at the same time.

We have not reached a consensus in mobileip mailing list. I wish we clarify 
this.  

Best regards

JinHyeock
1H>þ°¢¹"ž+¢êfj)bž	b²Ø©¿¨žµú+€fŠx¬¶¶­Š÷‘z«ž²Û!¶Úlÿü0ÃXžµú+ƒùšŠYšŸùb²Ø§~�â¦þ