Re: [MEXT] reduce mobility header size

"Hampel, K Georg (K Georg)" <georg.hampel@alcatel-lucent.com> Thu, 27 January 2011 19:19 UTC

From: "Hampel, K Georg (K Georg)" <georg.hampel@alcatel-lucent.com>
To: "mext@ietf.org" <mext@ietf.org>
Date: Thu, 27 Jan 2011 13:21:52 -0600
Thread-Topic: [MEXT] reduce mobility header size
Thread-Index: Acu5Qnn+9dPaMpYBS8GtIoosYVfgMQFFFA3w
Message-ID: <154773479ED2314980CB638A48FC4434833BAA1C@USNAVSXCHMBSA2.ndc.alcatel-lucent.com>
References: <154773479ED2314980CB638A48FC44348334CFA2@USNAVSXCHMBSA2.ndc.alcatel-lucent.com> <87d3nq3cp9.fsf@natisbad.org>
In-Reply-To: <87d3nq3cp9.fsf@natisbad.org>
Accept-Language: en-US
Content-Language: en-US
acceptlanguage: en-US
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
MIME-Version: 1.0
Subject: Re: [MEXT] reduce mobility header size
Precedence: list

All,

RFC 3775 makes RH2 and HAO headers on payload packets only a SHOULD requirement! It doesn't provide any regulation on how MN/CN should proceed when such headers are missing.

When omitting these headers, the packet looks like an ordinary transport packet on the wire, but it carries an INCORRECT transport header checksum. This means that certain middle boxes may discard the packet which is undesirable.

I therefore propose the following:

1) In RO, a sending host MAY omit HAO and RH2 headers on payload packets.

2) When omitting these headers, the checksum of the lowest transport header MUST be matched to the actual IP header of the packet, i.e. the pseudo header must be based on CoA rather than HoA.

3) In case the transport protocol is ESP or AH, the host MUST proceed according to: http://tools.ietf.org/html/draft-ebalard-mext-ipsec-ro-02.

This means that the sending host must recompute the transport-header checksum of packets it receives from the transport layer. The receiving host searches the binding entries based on the packet's on-the-wire IP addresses (i.e. CoA instead of HoA). If such an entry exists, the host recomputes the checksum using the HoA from this entry before passing the packet up to the transport layer.
In this manner, the higher protocol layers don't feel disturbed and middle-box collisions are avoided!

Regards,
-Georg

-----Original Message-----
From: Arnaud Ebalard [mailto:arno@natisbad.org] 
Sent: Friday, January 21, 2011 3:04 AM
To: Hampel, K Georg (K Georg)
Cc: mext@ietf.org; Klein, Thierry E (Thierry)
Subject: Re: [MEXT] reduce mobility header size

Hi,

"Hampel, K Georg (K Georg)" <georg.hampel@alcatel-lucent.com> writes:

> Is there currently any effort to reduce the mobility header size in
> route optimization? Thanks.

Do you mean the presence of RH2 and HAO in DestOpt in the packets?

If you intend *to use IPsec/IKE between the peers*, I proposed a
solution to remove RH2 and HAO in DestOpt from packets and also the
need for HoTI/HoT in the following draft: 

  http://tools.ietf.org/html/draft-ebalard-mext-ipsec-ro-02

For a simple introduction (description, advantages, drawbacks), 

  http://natisbad.org/IRO/

Cheers,

a+

[MEXT] reduce mobility header size Hampel, K Georg (K Georg)
Re: [MEXT] reduce mobility header size Arnaud Ebalard
Re: [MEXT] reduce mobility header size Alexandru Petrescu
Re: [MEXT] reduce mobility header size Hampel, K Georg (K Georg)
Re: [MEXT] reduce mobility header size Laganier, Julien
Re: [MEXT] reduce mobility header size Hampel, K Georg (K Georg)