On Wed, Nov 29, 2000 at 03:28:56PM -0600, Brian E Carpenter wrote:
> We would have to evaluate the risk (which I guess is traffic analysis)
> versus the benefit (ability to provide QOS). That sounds like a
> user decision to me.
Of course.
Which is, why I think intermediate routers will have to interpret flow labels
as magic cookies. Maybe changing them (I have no strong opinion about this),
maybe internally sorting them according to a flow label - address/protocol/port
mapping for unencrypted packets; but as protocol designers, we can't enforce
a static mapping, else we would make hiding of flow properties through ESP
impossible to the users.
Regards,
-is
- Re: Usage of IPv6 flow label Brian E Carpenter
- Re: Usage of IPv6 flow label Alex Conta
- Re: Usage of IPv6 flow label Brian E Carpenter
- Re: Usage of IPv6 flow label itojun
- Re: Usage of IPv6 flow label Brian E Carpenter
- Re: Usage of IPv6 flow label itojun
- RE: Usage of IPv6 flow label Elwyn Davies
- Re: Usage of IPv6 flow label Brian E Carpenter
- Re: Usage of IPv6 flow label Ignatios Souvatzis
- Re: Usage of IPv6 flow label Brian E Carpenter
- Re: Usage of IPv6 flow label Ignatios Souvatzis
- Re: Usage of IPv6 flow label Brian E Carpenter
- Re: Usage of IPv6 flow label Jim Bound
- Re: Usage of IPv6 flow label Francis Dupont
- RE: Usage of IPv6 flow label Thomas Eklund
- RE: Usage of IPv6 flow label Michael Thomas
- Re: Usage of IPv6 flow label Brian E Carpenter
- Re: Usage of IPv6 flow label Jim Bound
- Re: Usage of IPv6 flow label Jim Bound
- Re: Usage of IPv6 flow label Brian E Carpenter
- RE: Usage of IPv6 flow label Christian Huitema
PGP signature