Hi Guang,
I realized I never acknowledged your responses. Sorry for the delay.
It does clear my concerns.
Thank you!
Eric
On 16/03/12 07:42, Guang Yao wrote:
Hi, Eric
Thank you for the comments. My replies are in the line. We have
updated the text as the attachment. Sorry for it cannot be submitted
because the submit window is closed.
Best regards,
Guang
2012/3/13 eric levy-abegnoli <elevyabe(_at_)cisco(_dot_)com
<mailto:elevyabe(_at_)cisco(_dot_)com>>
Hi,
here are my substantive comments
Look for [eric].
Eric
7.3.1. Timer Expiration Event
EVE_ENTRY_EXPIRE: The lifetime of an entry expires
[eric] 2 minutes sounds very long. DHCP client timeout is 1 sec
for the first
message. Then multiplied by 2, etc. What is the rational behind
this value, which increase the window for DoS attacks?
[guang]
In RFC3315, it reads:
"RT for the first message transmission is based on IRT:
RT = IRT + RAND*IRT
RT for each subsequent message transmission is based on the previous
value of RT:
RT = 2*RTprev + RAND*RTprev
MRT specifies an upper bound on the value of RT (disregarding the
randomization added by the use of RAND). If MRT has a value of 0,
there is no upper limit on the value of RT. Otherwise:
if (RT> MRT)
RT = MRT + RAND*MRT"
Here MRT is 120s. Based on this value, the maximum retransmission
time is in range of 120s(+-)12s. Thus, we think 120s is a favorable
value to remove an entry.
The DoS in this window is a problem, but we think the binding number
limitation on each binding anchor can mitigate the damage.
8. Supplemental Binding Process
[eric] This section is very unclear. The conditional SHOULD
based on "vendor ability" sounds like a "MAY" to me, which is not
what I remember of the WG consensus. In addition, hosts are not
required to (DHCP) re-configure upon link flapping, even when they
are directly attached. The text seems to indicate otherwise.
In practice, in the absence of such mechanism, traffic will be
blocked.
[guang]
We have removed the condition on "vendor ability" . Link flap is
handled through keeping bindings for a period after binding anchor
off-link. We have changed the text to make it clear.
8.1. Binding Recovery Process
[eric] It is unclear what the address is bound to. In the normal case,
the entry is created upon receiving a message (i.e. REQUEST) from
the client, and the anchor is stored by that time. You should
specified where the anchor comes from in this scenario, and where
was it stored (given that the section specifies the binding
entry creattion on LQ Reply)
[guang]
We have changed the text, and specified each step. Tell me if it is
still unclear.
10. State Restoration
[eric] Requiring non-volatile memory sounds wrong. Other techniques
exists such as redundant boxes (switches) synchronizing states. I
don't recall that non-volatile memory was discussed at length in the
WG, especially given that it carries its own challenges: frequency
for saving states, load incurred, etc)
The one technique that was discussed in the WG was Binding Recovery
process. One solution should be enough.
[guang]
There can be a large number of bindings on the savi device. If only
relying on the binding recovery process, there can be a large latency.
Especially, the recovery in this mechanism requires querying the DHCP
server.
Moreover, the storing in non-volatile memory is just recommended but
not mandatory. Using redundant box can be another suggestion. We have
change the MUST to MAY in text.
Eric
On 06/03/12 16:01, The IESG wrote:
The IESG has received a request from the Source Address Validation
Improvements WG (savi) to consider the following document:
- 'SAVI Solution for DHCP'
<draft-ietf-savi-dhcp-12.txt> as a Proposed Standard
The IESG plans to make a decision in the next few weeks, and
solicits
final comments on this action. Please send substantive
comments to the
ietf(_at_)ietf(_dot_)org <mailto:ietf(_at_)ietf(_dot_)org> mailing
lists by
2012-03-20. Exceptionally, comments may be
sent to iesg(_at_)ietf(_dot_)org <mailto:iesg(_at_)ietf(_dot_)org>
instead. In
either case, please retain the
beginning of the Subject line to allow automated sorting.
Abstract
This document specifies the procedure for creating bindings
between a
DHCPv4 [RFC2131]/DHCPv6 [RFC3315] assigned source IP
address and a
binding anchor [I-D.ietf-savi-framework] on SAVI (Source
Address
Validation Improvements) device. The bindings can be used
to filter
packets generated on the local link with forged source IP
address.
The file can be obtained via
http://datatracker.ietf.org/doc/draft-ietf-savi-dhcp/
IESG discussion can be tracked via
http://datatracker.ietf.org/doc/draft-ietf-savi-dhcp/ballot/
No IPR declarations have been submitted directly on this I-D.
_______________________________________________
savi mailing list
savi(_at_)ietf(_dot_)org <mailto:savi(_at_)ietf(_dot_)org>
https://www.ietf.org/mailman/listinfo/savi
_______________________________________________
savi mailing list
savi(_at_)ietf(_dot_)org <mailto:savi(_at_)ietf(_dot_)org>
https://www.ietf.org/mailman/listinfo/savi