RE: Proposed way forward with the transition mechanisms

["Karen E. Nielsen (AH/TED)" <karen.e.nielsen@ericsson.com>]

> -----Original Message-----
> From: Brian E Carpenter [mailto:brc@zurich.ibm.com]
> Sent: Tuesday, August 03, 2004 1:20 AM
> To: Karen E. Nielsen (AH/TED)
> Cc: 'Alain Durand'; Karim El-Malki (AL/EAB); 'Soininen Jonne ';
> 'v6ops@ops.ietf.org '
> Subject: Re: Proposed way forward with the transition mechanisms
> 
> 
> Karen E. Nielsen (AH/TED) wrote:
> > Hi Alain,
> > 
> > I find it difficult to argue based on 
> draft-ietf-v6ops-assisted-tunneling-requirements-00.txt given 
> that this draft so clearly focuses on the full-fledged 
> > tunnelling functionality demanded by (perhaps ?) the ISP environment
> > and not the lightweight tunnelling function requested for 
> the 3GPP environment.
> > Examples - just to pick a few: NAT traversal, prefix 
> delegation, extensibility requirements of 
> draft-ietf-v6ops-assisted-tunneling-requirements-00.txt. 
> Neither of which can be found in the 3GPP v6ops documents, 
> that is RFC 3574 and draft-ietf-v6ops-3gpp-analysis-10.txt.
> > 
> > I opt for standard track standardization of Isatap as it 
> serves the needs of the 3GPP environment. In particular:
> > 
> > * It is simple and easy to deploy,
> 
> Manual configuration is easy to deploy at 3GPP scale?
> 
>  From the ISATAP draft:
> 
> 8.3.2 Potential Router List Initialization
> 
>     ISATAP nodes initialize an ISATAP interface's PRL with 
> IPv4 addresses
>     discovered via manual configuration, a DNS fully-qualified domain
>     name (FQDN) [STD13], a DHCPv4 option, a DHCPv4 vendor-specific
>     option, or an unspecified alternate method. FQDNs are 
> established via
>     manual configuration or an unspecified alternate method.

I assume that you're hinting at manually configured FQDNs - ?
The solution anticipated for 3GPP (at least) would be to rely on
FQDNs on the form isatap.<domain_name>", where
<domain_name> is the DNS suffix configured on the underlying IPv4 interface.

Other tunnel discovery possibilities, which I believe have been considered for 3GPP2 usage, is to use a special option in DHCP. It would be preferable to agree on relying on the same solution in both 3GPP and 3GPP2, but not an absolute requirement.


> 
> > * It is lightweight network architecture wise, i.e. it
> > requests one additional infrastructure element only, namely 
> the Isatap router.
> 
> And an extra module in every host stack that wants to use it.

So would all tunnelling mechanisms, I assume.

Karen
>