Re: [savi] WGLC: draft-ietf-savi-dhcp-22

["Eric Levy- Abegnoli (elevyabe)" <[e-mail]>]

Hi Guang,

On 22/04/14 05:56, "Guang Yao" <[e-mail]> wrote:

>Hi, Ted
>
>Thank you very much for your comments! I agree with you. It's safe to
>assume
>all the SAVI devices have layer-3 stack. But it seems Eric also concerns
>the
>implementation of DHCP leasequery and NDP snooping.
>
>On the MLD problem mentioned by Eric, I wonder should SAVI-DHCP be
>consistent with RFC6620, or be different?
>
>Hi, Eric
>
>On the DAD problem:
>
>I read the doc again and find DAD NS will not be sent to the tentative
>node.
>Whenever probe should be sent to the tentative node, we use plain NS
>instead
>of DAD NS. Thus would it be OK?

I am a little bit confused on the scenario we are talking about (apologies
if this is my reading of the text).
The scenario I have in mind is
 1) SAVI device gets a data packet sourced with an address not found in
its table 
 2) the SAVI device creates the binding (in DETECTION)
 3) The SAVI device send and ARP/DAD to look for a conflict (I guess to
the entire vlan but the receive interface).
 4) If no conflict was detected, it does the LQ
I'll tend to argue that if LQ is required, the DETECTION state is not
necessary and should be removed.
If LQ is optional, and DETECTION failed to detect a conflict, then the
entry should not move back right away to NO_BIND?

DETECTION in general is problematic. The state in the SAVI device and on
the end-node which has the conflicting address are independent. When he
SAVI switch sends a DAD, while the end-node is right in TENTATIVE (bad
luck), this DAD would shutdown the end-node interface. Maybe that is a
problem we have to deal with, but these probe packets are causing exactly
this type of issues, seen in real life. I thought I should mention it.

In your response, you said it could also send a regular NS, which I assume
is an NS lookup. I could not find a reference in the text about sending NS
lookup instead of DAD. This is certainly a good practice (we have
implemented it that way) provided that you have an address to source it
from. However (my usual objection) the access switches often don't have a
l3 address per link/vlan, not even a link-local address. So NS-lookup
should be an option. Which leaves only DAD or LQ.
 

Thank you.
Eric




>
>We are looking forward to your further comments, thanks!
>
>Best regards,
>Guang
>
>-----Original Message-----
>From: Ted Lemon [mailto:[e-mail]]
>Sent: Tuesday, April 22, 2014 12:31 AM
>To: Guang Yao
>Cc: Eric Levy- Abegnoli (elevyabe); Jean-Michel Combes; SAVI Mailing List;
>[e-mail]
>Subject: Re: [savi] WGLC: draft-ietf-savi-dhcp-22
>
>Do we really think there are modern layer 2 devices that will implement
>SAVI-DHCP that will not have IPv6 addresses?   This seems highly doubtful
>to
>me-the devices that would only have layer two addresses would be unmanaged
>switches.   I have a cheap managed switch, and it has an IPv4 address and
>a
>web server in it.   I think this is a non-problem.
>
>
>