Re: [Monami6] Comments on draft-petander-mip6-mbb-00.txt

[Christian Vogt <[e-mail]>]

Hi Henrik,

 > Actually we assume only that MN performs MIPv6 handoff at this time.
 > L2 handoff and IPv6 link attachment can be performed before the MIPv6
 > handoff, if MN is capable of connecting to both the old and new access
 > router simultaneously, which is required for make-before-break
 > handoffs, which our draft focuses on. In successful make-before-break
 > handoffs the delays of L2 handoff and IPv6 link attachment do not lead
 > to packet loss. The handoff procedure for a make-before-break handoff
 > would be:

ok, if the MN has multiple interfaces, it can set up the 2nd interface
at any time btw. BU transmission and BA reception.

However, you can do a make-before-break handoff even with a single
interface.  This is actually what we did in
draft-vogt-mobopts-early-binding-updates-00.txt.

Nonetheless, the scope of your draft is two interfaces, I see this.  And
given that you are using two interfaces, you can do a care-of-address
test on the 2nd while communicating with the 1st.

I agree with the rest of your comments.

Take care!

- Christian

--
Christian Vogt, Institute of Telematics, University of Karlsruhe
www.tm.uka.de/~chvogt/pubkey/


Henrik Petander wrote:
> Hi Christian,
>
> Thanks for reading the draft and commenting it. Replies to your
> comments are inline.
>
> On 10/19/05, Christian Vogt <[e-mail]> wrote:
>
>>Hi Henrik, hi Eranga,
>>
>>I took a good look at draft-petander-mip6-mbb-00.txt.  These are very
>>valuable considerations indeed.  A few comments below.
>>
>>You assume that the MN affects the L2 handoff at the time it receives
>>the (first) Binding Acknowledgement (BA).
>
>
> Actually we assume only that MN performs MIPv6 handoff at this time.
> L2 handoff and IPv6 link attachment can be performed before the MIPv6
> handoff, if MN is capable of connecting to both the old and new access
> router simultaneously, which is required for make-before-break
> handoffs, which our draft focuses on. In successful make-before-break
> handoffs the delays of L2 handoff and IPv6 link attachment do not lead
> to packet loss. The handoff procedure for a make-before-break handoff
> would be:
>
> 1) Find a suitable next AR using idle interface A, while using
> interface B for active communications via old AR.
> 2) Perform L2 handoff and IPv6 link attachment on interface A with new AR
> 3) Perform MIPv6 handoff to switch traffic to new interface.
>
> In the draft we focus only on 3), since 1) and 2) can be done with the
> existing protocol specs for IPv6 (in handoffs to foreign networks). I
> hope this clarifies the intent of our draft.
>
>
>>If the L2 handoff takes a long time, initiating it as early as this
>>would be ok, since the MN would loose packets during the L2 handoff anyway.
>
>
> This is true in the case of break-before-make handoffs, in which MN
> loses connectivity with its old AR before establishing connectivity
> with a new one. However, for make-before-break handoffs L2 handoff
> delay does not have a direct connection with packet loss, only on the
> probability of being able to finish the handoff before connectivity at
> the old CoA ends.
>
>
>>You may want to take a look at the following draft, however:
>>
>>   draft-vogt-mobopts-early-binding-updates-00.txt
>>
>>This draft---in particular section 4, 7th paragraph and
>>onwards---defines how a single-interfaced MN can use Early Binding
>>Updates and Credit-Based Authorization for handoff anticipation and
>>proactive Mobile IPv6 signaling.
>
>
> Thanks for the pointer. If your draft is used with make-before-break
> handoffs, it will reduce the time required for handoff. Shortened
> handoff time increases the probability of a make-before-break handoff
> finishing before the connectivity at old CoA ends.
>
> Best regards,
> Henrik

_______________________________________________
Monami6 mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/monami6