Re: [Hipsec] HIP Registration problem

[Miika Komu <[e-mail]>]

Hi,

agree with Andrew. We've had already a related discussion on parallel 
connectivity checks which applies to rendezvous as well:

http://www.ietf.org/mail-archive/web/hipsec/current/msg02837.html

On 11/08/2010 05:28 AM, Andrew McGregor wrote:
> Register with several and use DNS round-robin, or just talk to all of
> them at once. That's probably good enough, I think.
>
> Andrew
>
> On 8/11/2010, at 11:03 AM, Zhangfeng HU wrote:
>
>> Hi,
>> yes, that's the truth. Since HIP does provide a mechanism as DAAP of
>> MIP, we are just considering a method of providing fault torlerance
>> for RVS in HIP. For some mobile nodes, the high availability is very
>> important, however the availability of these mobile nodes is mostly
>> decided by the availability of the corresponding RVS who serves them.
>> This RVS fault torlerance can not be easily implemented through
>> updating the entries in DNS, because of the high latency and the cache
>> mechanism used in DNS.
>> 2010-11-08
>> ------------------------------------------------------------------------
>> Zhangfeng HU, PhD candidate of
>> Broadband Network Research Center,
>> State Key Laboratory of Networking and Switching Technology,
>> Beijing University of Posts and Telecommunications, Beijing, China
>> Email:[e-mail] <mailto:[e-mail]>
>> Tel:86+13811892137
>> ------------------------------------------------------------------------
>> *发件人：* Miika Komu
>> *发送时间：* 2010-11-04 18:47:45
>> *收件人：* hipsec
>> *抄送：*
>> *主题：* Re: [Hipsec] HIP Registration problem
>> On 11/04/2010 11:48 AM, zfhu2001 wrote:
>> > Hi all,
>> > As we know, a mobile node registers its HIP and IP address binding to a
>> > correponding RVS server, but now I'm wondering whether the MN will
>> > register to a new RVS if a mobile node moves to another network and
>> > simultaneously its previous RVS server does not work any more. Or does
>> > HIP provide a mechanisim as DAAP of MIP to detect the serving RVS server
>> > of its home area?
>> > Thanks.
>> > 2010-11-04
>> Hi,
>> there's no such thing as "home agent" or "home area" in HIP. This is one
>> of the differences to MIP. The same rendezvous server can be reused
>> independently of the current network using the UPDATE procedure.
>> Remember also that the rendezvous server forwards only the first I1 or
>> the first UPDATE packet, so it's still a light-weight operation even
>> though the latency be longer. The rendezvous server does not forward any
>> application payload.
>> Fault-tolerance for RVS may fall into the category of local policies
>> (?). Changing the "current" RVS requires also an update to the DNS
>> records. Another strategy could be just to publish multiple RVS servers
>> in the DNS and send packets to all of them (which ever happens to be
>> alive).
>> _______________________________________________
>> Hipsec mailing list
>> [e-mail] <mailto:[e-mail]>
>> https://www.ietf.org/mailman/listinfo/hipsec
>> ÔÚ´«ÈëµÄÓʼþÖÐδ·¢ÏÖ²¡¶¾¡£
>> ¼ì²é¹¤¾ß£ºAVG - www.avg.com <http://www.avg.com>
>> °æ±¾£º9.0.865 / ²¡¶¾Êý¾Ý¿â£º271.1.1/3235 - ·¢²¼ÈÕÆÚ£º11/03/10 16:36:00
>> _______________________________________________
>> Hipsec mailing list
>> [e-mail] <mailto:[e-mail]>
>> https://www.ietf.org/mailman/listinfo/hipsec
>
>
>
> _______________________________________________
> Hipsec mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/hipsec