Re: [Softwires] Port Set Definition Algorithms Analysis

2012/3/7 Rémi Després <despres.remi@laposte.net>

> Hi Tina,
>
> Le 2012-03-07 à 09:58, Maoke a écrit :
>
>
>
> 2012/3/7 Tina TSOU <Tina.Tsou.Zouting@huawei.com>
>
>>
>>
>> Sent from my iPad
>>
>> On Mar 6, 2012, at 2:24 AM, "GangChen" <phdgang@gmail.com> wrote:
>>
>> > Hello Tina,
>> >
>> > Could you help to clarify the sentence:
>> >
>> >   *1 MAP and divi-pd provide better security than 4rd-U, because they
>> >   provide more variation in port set definition.
>> >
>> > I guess they should have similar features on security
>> Agreed. No much difference between MAP, divi-pd, and 4rd-U with regard to
>> security. So I propose to remove this sentence.
>>
>
> +1 (see below)
>
>
>
> well, *better security* might be an inaccurate expresssion. i suppose MAP
> provides better flexibility than 4rd-U. personally, i understand the reason
> of 4rd-U requiring fixed offset (4bits) is to support the longest-match of
> port for CE-finding in the mesh mode (i.e., the MAX PSID). it is a
> tradeoff. without the fixed offset, longest-match fails for the PSID
> matching and there must be a way of distributing CE's PSID to other CEs.
> however,
>
>
>
> it does also confuses me that the newest 4rd-U draft doesn't mention the
> MAX PSID feature (or maybe i missed something).
>
>
> The max PSID feature, present in 4rd-u-00 and -01 (i.e. before IETF 82),
> was abandoned during the Taipei meeting (November 2011)
> It no longer appeared in -02 (December 29).
>
> Reasons of the 4rd-u fixed offset are (ref. sec 4.3 of -04):
> "NOTE: The choice of the PSID position in Port fields has been guided by
> the following objectives: (1) for fairness, avoid having any of the
> well-known ports 0-1023 in the port set specified by any PSID value (these
> ports have more value than others); (2) for compatibility RTP/RTCP
> [RFC4961], include in each port set pairs of consecutive ports; (3) in
> order to facilitate operation and training, have the PSID at a
> fixed position in port fields; (4) in order to facilitate documentation in
> hexadecimal notation, and to facilitate operation and training, have the
> PSID at a fixed position in port fields."
>
>

Remi,

point (1) is qualified for the individual use case but not available for
the enterprise environment, where surely it is possible that one put all
well know services under a same CE. but if MAP or 4rd-U states enterprise
use case is NOT to be supported, it is fine. point (2) is qualified but not
a sufficient condition to derive the fixed position of PSID. point (3) and
(4) is about the human reading on the PSID. however, the MAP or 4rd-U
address has been human-unfriendly. i don't see training and operation on
easy-to-read PSID is significant.

therefore i think MAXPSID, though it is abandoned, is a convincing reason
of having fixed position of PSID.

- maoke

>
>
> Hope it clarifies.
>
> Regards,
> RD
>
>
>
>
>
>> >
>> > Gang
>> >
>> > 2012/3/2, Tina TSOU <Tina.Tsou.Zouting@huawei.com>:
>> >> Dear all,
>> >> You may be interested to comment on
>> >>
>> http://datatracker.ietf.org/doc/draft-tsou-softwire-port-set-algorithms-analysis/
>> >>
>> >> Abstract:
>> >>  This memo analyses some port set definition algorithms which
>> >>  encode port set information into IPv6 address so as to support
>> >>  stateless IPv4 to IPv6 transition technologies, e.g. 4rd-U and MAP.
>> >>
>> >>
>> >> Tetsuya, Simon and Tina
>> >>
>> >> _______________________________________________
>> >> Softwires mailing list
>> >> Softwires@ietf.org
>> >> https://www.ietf.org/mailman/listinfo/softwires
>> >>
>> _______________________________________________
>> Softwires mailing list
>> Softwires@ietf.org
>> https://www.ietf.org/mailman/listinfo/softwires
>>
>
> _______________________________________________
> Softwires mailing list
> Softwires@ietf.org
> https://www.ietf.org/mailman/listinfo/softwires
>
>
>