Re: [dhcwg] static route option for dhcpv6

[Martin Stiemerling <[e-mail]>]

--On Donnerstag, Januar 17, 2002 12:29:00 -0500 Jim Bound 
<[e-mail]> wrote:

> We need the static route option for the dentist office scenarios of IPv6
> where there are two lans and no routers or as VJ pointed out ipv6forwardin
> g is turned on and not doing RAs.

But doesn't an IPv6 router that isn't sending his router advertisements 
violate the IPv6 spec? In this case DHCPv6 should not support this feature 
in this way.

>
> We also want it for configured tunnels and its different than DSTM.

Ok, I agree with using it for configured tunnels. But I would choose 
another name, like IPv4 Tunnel End Point.

>
> We should also keep it simple as a config parameter.

Yes.

>
> I see no pain here and it is needed.  Its a simple option to add and
> necessary for early IPv6 deployment as VJ pointed out in first mails.
>
> regards,
>
>
> /jim
>
>
> On Thu, 17 Jan 2002, Vijay Bhaskar A K wrote:
>
>> Please see my comments inline.
>>
>> ~Vijay.
>>
>> > After thinking about this more and after seeing the other discussion
>> > on this subject, I'm not sure exactly when or why this option would be
>> > needed. But on the other than, it technically isn't needed in IPv4
>> > either because ICMP redirects and other routing table distribution
>> > techniques exist and DHCPv4 does have such an option (and a revised
>> > one to carry classless routes).
>> >
>> > So, we can do one of two things:
>> > 1. Include it and consider DHCPv6 as a toolbox and those people that
>> > want to use it (and those clients that want to support it) do so. For
>> > example, Solaris 8 includes the route command and it supports IPv6
>> > routing table operations. Can anyone who has lots of experience with
>> > IPv6 deployment indicate whether there is a need to statically add
>> > routing table entries? 2. Wait until someone has a clear case of
>> > needing it and have it defined in some future document.
>>
>> Assume the following scenaria.
>> - There are 2 networks A and B.
>> - There is a node n connected  to both the  network,  and it has enabled
>> ipv6-forwarding and not sending router advertisements.
>>
>> Now, a node in network A gets an address from the DHCPv6  server and now
>> it  wants  to  communicate  with a node in  network  B.  In the  current
>> scenario, the route has to be manually configured, then only, it will be
>> able to contact the node in network B.  With static route option, we can
>> autoconfigure   it.  This  will  be  more  helpful  in  getting  minimal
>> configuration   for  smaller  networks,  which  don't  have  any  router
>> advertisements  and for the networks which have not completely  deployed
>> routing mechanisms.
>>
>> It will be useful in the getting the configured tunnels also.
>>
>> >
>> > If we do want to include it, questions to ponder:
>> > - Should any lifetimes be associated with the routes? Either one
>> > lifetime for all routes or each route? - Should this option be
>> > encapsulated within an IA? That way, it would be renewed with the IA.
>>
>> I think, we can treat this as another configuration parameter.  We don't
>> need to mix up with IA.  If there are  multiple  IAs with  same  prefix,
>> then this static route is common for all these IAs.  Whenever there is a
>> change in the  static  route, we can use  reconfiguration  mechanism  to
>> update it.
>>
>> >
>> > I myself am leaning more towards recommending we wait until a need is
>> > found.
>> >
>> > - Bernie
>> >
>> >
>> > -----Original Message-----
>> > From: Vijayabhaskar A K [mailto:[e-mail]]
>> > Sent: Wednesday, January 16, 2002 1:13 PM
>> > To: 'Bernie Volz (EUD)'; [e-mail]
>> > Subject: RE: [dhcwg] static route option for dhcpv6
>> >
>> >
>> > Bernie,
>> > This option format looks ok for me. We can include it.
>> > Vijay
>> >
>>
>>
>> --
>> ____Vijay_Bhaskar_A_K____
>> ______Inet_Services______
>> ________HP_ISO___________
>> ______Ph:_2051424________
>> ____Telnet:_847_1424_____
>> ___Pager:_9624_371137____
>>
>>
>> _______________________________________________
>> dhcwg mailing list
>> [e-mail]
>> https://www1.ietf.org/mailman/listinfo/dhcwg
>>
>
>
> _______________________________________________
> dhcwg mailing list
> [e-mail]
> https://www1.ietf.org/mailman/listinfo/dhcwg
>



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