Re: SRH insertion vs SRH insertion + encapsulation

[Brian E Carpenter <[e-mail]>]

Hi Fernando,

On 08-Sep-19 11:11, Fernando Gont wrote:
> On 8/9/19 01:52, Ron Bonica wrote:
> [...]
>>  
>>
>> IMHO, EH insertion modifies the semantics of the IPv6 source address.
>> Today, the IPv6 source address indicates the source of an IP packet and
>> **ALL** of its contents. If transit routers are allowed to insert
>> extension headers, downstream routers can no longer identify the source
>> of a packet and all of its contents.>
>>  
>>
>> Granted, in some cases, transit routers are allowed to modify a packet
>> (e.g., Hop Count, DHCP, mutable options). But there is a big difference
>> between changing a field whose value is know to me mutable and inserting
>> a new option.
> 
> Indeed, it's a major modification (besides the associated issues that
> have been pointed out a number of times by several folks).

It's explicitly forbidden by RFC8200 plus the fact that option length is
immutable according to IPsec. So the issue is pretty black and white
in terms of the standards track.

> That's why I think it's not even in the scope of 6man.

6man "is not chartered to develop major changes or additions
to the IPv6 specifications." Of course it's a matter of judgment,
but I agree that changing this is "major".

Look, this is one of the reasons we wrote
https://tools.ietf.org/html/draft-carpenter-limited-domains
which is currently being considered as an Independent Submission.
If the community wants to standardise mechanisms that are *not*
intended to work across the open Internet, we need to make that
clear, and specify how it works safely and securely. That's a
community choice, not a WG choice, IMHO.

Regards
   Brian