[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: questions about draft-wen-ipv6-rsra-opt-multihoming-00



hi Lawrence,

I'm glad to receive your comments.

I can explain to you one by one.

1. The secnarios in the draft are using the ISP name sub-option. It does not 
mean that it's a "MUST" sub-option in the multi-homing information option for 
RS message (and other sub-option can be used if needed). Of course, if we
want to solve the problem described in the charter of shim6 w.g., it is a "MUST"
sub-option for RA message. Without this information along with the Prefix 
Information option, host doesn't know how to choose the appropriate prefix and 
the associated exit router. 
I think multi-homing environment doesn't mean that some hosts in a multi-
home site just in the VLAN connected to a specified ISP. All the hosts in the 
multihome site have been connected to all the exit routers that belongs to 
different ISP. That is, the host can obtain all the periodical RA messages from
different routers. Our goal is to solve prefix selection and exit router selection in
this environment.

2. First, that's just an example of how to use multi-homing option in access
network. It can also be used in LAN with multiple exit routers.
The layer2 access node I mean in this example is not just a pure layer2 devices.
It can snoop some special layer3 packets, such as RS/RA messages, then it
can forward the RS to the correct router without flooding it to other parts of 
access netework. 
If the access node doesn't have this ability, RS can be sent to all routers, only
the corresponding will respond this RS message. This will not add any bad 
impact.

Thanks.

Best regards,

Haibo



> -----Original Message-----
> From: Lawrence Zou [mailto:zou.rong@huawei.com]
> Sent: 2006年6月6日 12:04
> To: CTO WEN Haibo
> Cc: multi6@ops.ietf.org
> Subject: questions about draft-wen-ipv6-rsra-opt-multihoming-00
> 
> 
> hi,wen:
> 
>      i have read you draft.  Although there are some strong technical
> points in this document,  i still think there are some big problem:
> 
> 1. I noticed  that in all 3 Scenarios, the RS message must include ISP
> name sub-option,so,is it a "MUST"  sub-optinon that be include in the 
> Multi-homing Information option? if it is true,what will it happen if
> the host don't know the name of the ISP? I think in the Scenarios of
> stateless 
> addres autoconfigue, it is not necessary for host to know the topology
> of the network and the name of the ISP.
> in your draft ,all hosts know clearly which ISP they belong 
> to ,I think
> we can distinguish them using some kinds of  VLAN techonoly.
> 
> 2.In Scenario 3,you mention the equipment of "layer2 CPE ".so what the
> difference is that with "layer3 CPE "? In my understand ,the 
> layer2 CPE 
> is a layer2 equipmnet ,so in the section 3.3 
>    Step (b) : Base on the information in Multi-homing 
> Information option
>               , layer 2 access node can forward this RS to the correct
>               edge router of the desired ISP without sending 
> it to the 
>               router that will not respond this RS.	
> 
>  can layer2 equipment do this kind of thing?	
>   
> 
> 
>