Re: [MEXT] TLV header in DSMIP

["Vijay Devarapallli" <dvijay at gmail.com>]

On Thu, Apr 3, 2008 at 1:44 AM, Hesham Soliman <hesham at elevatemobile.com> wrote:
> Folks,
>
>  During one of the reviews for DSMIP the issue came up regarding the
>  TLV header Type field values and how the current spec doesn't describe
>  how it can be used for IPsec.

What was the issue?

>  After looking into this a bit I don't see why we would need IPsec for
>  this. We originally defined because people wanted to use GRE so the
>  format would be:
>  IP
>  UDP
>  GRE
>  IP
>  ...etc
>
>  It seems like it would be sufficient to only define this now for GRE
>  and if someone comes up with a need for using IP/IPsec directly after
>  the TLV header they can define that behaviour and request a new type
>  value.
>
>  So the new draft would only allocate one value for GRE.
>
>  Any objections to this ?

Yes. When we specified the TLV header, it was felt that the mobile
node and the home agent need can use the same UDP port for
both traffic tunneled with ESP encryption and for traffic without ESP
encryption. The same TLV header would tell the home agent whether
there is a GRE, IPv4, IPv6 or IPsec header that follows the TLV
header.

Without this, you end up using UDP port 4500 for IPsec protected
data traffic between the mobile node and the home agent and the
DS-MIPv6 UDP port for non-IPsec protected data traffic between the
mobile node and the home agent. I would prefer to avoid this.

Perhaps you could explain what the issue is with using the TLV
header followed by the ESP header.

Vijay
_______________________________________________
MEXT mailing list
MEXT at ietf.org
https://www.ietf.org/mailman/listinfo/mext