Re: [RFC net-next 00/15] add basic PSP encryption for TCP connections

[Willem de Bruijn <willemdebruijn.kernel@gmail.com>]

Steffen Klassert wrote:
> On Tue, May 28, 2024 at 09:49:34AM -0400, Willem de Bruijn wrote:
> > Steffen Klassert wrote:
> > > On Wed, May 22, 2024 at 08:56:02AM -0400, Paul Wouters wrote:
> > > > Jakub Kicinski wrote:
> > > > 
> > > > > Add support for PSP encryption of TCP connections.
> > > > > 
> > > > > PSP is a protocol out of Google:
> > > > > https://github.com/google/psp/blob/main/doc/PSP_Arch_Spec.pdf
> > > > > which shares some similarities with IPsec. I added some more info
> > > > > in the first patch so I'll keep it short here.
> > > > 
> > > > Speaking as an IETF contributor, I am little surprised here. I know
> > > > the google people reached out at IETF and were told their stuff is
> > > > so similar to IPsec, maybe they should talk to the IPsecME Working
> > > > Group. There, I believe Steffen Klassert started working on supporting
> > > > the PSP features requested using updates to the ESP/WESP IPsec protocol,
> > > > such as support for encryption offset to reveal protocol/ports for
> > > > routing encrypted traffic.
> > > 
> > > This was somewhat semipublic information, so I did not talk about
> > > it on the lists yet. Today we published the draft, it can be found here:
> > > 
> > > https://datatracker.ietf.org/doc/draft-klassert-ipsecme-wespv2/
> > > 
> > > Please note that the packet format specification is portable to other
> > > protocol use cases, such as PSP. It uses IKEv2 as a negotiation
> > > protocol and does not define any key derivation etc. as PSP does.
> > > But it can be also used with other protocols for key negotiation
> > > and key derivation.
> > 
> > Very nice. Thanks for posting, Steffen.
> > 
> > One point about why PSP is that the exact protocol and packet format
> > is already in use and supported by hardware.
> > 
> > It makes sense to work to get to an IETF standard protocol that
> > captures the same benefits. But that is independent from enabling what
> > is already implemented.
> 
> Sure, PSP is already implemented in hardware and needs to be supported.
> I don't want to judge if it was a good idea to start this without
> talking to the IETF, but maybe now Google can join the effort at the
> IETF do standardize a modern encryption protocol that meets all the
> requirements we have these days. This will be likely on the agenda of
> the next IETF IPsecME working group meeting, so would be nice to
> see somebody from the Google 'PSP team' there.

Sounds good. We'll try to have someone join the next WG meeting
during IETF 120.