- From: David Schinazi <dschinazi.ietf@gmail.com>
- Date: Tue, 18 Oct 2022 14:48:59 -0700
- To: Ben Schwartz <bemasc@google.com>
- Cc: HTTP Working Group <ietf-http-wg@w3.org>
- Message-ID: <CAPDSy+7-XDSSeqFx5FWkSbej6fAGvvMdDKExghgS0DO6BeGL=g@mail.gmail.com>
Hi Ben, I don't think confidential HTTP resources are a solved problem. The unguessable path approach you describe is similar to a shared secret (à la symmetric cryptography) but there is no equivalent for asymmetric cryptography. While I think your draft is interesting and worth discussing, I think the technology overlap isn't big enough to warrant discussing the two drafts together - they're separate proposals with different goals. Thanks, David On Tue, Oct 18, 2022 at 9:14 AM Ben Schwartz <bemasc@google.com> wrote: > I support the goals of the Unprompted Authentication draft. In fact, I'm > so supportive that I recently posted a draft that happens to solve an > overlapping problem in a very different way: "Modernizing HTTP Forward > Proxy Functionality" [1]. > > To step back: confidential HTTP _resources_ are arguably a solved > problem. We can simply place the resource at an unguessable path (e.g. > "capability URLs" [2]). The problem mentioned by this draft occurs when > the HTTP service is origin-scoped (e.g. it is not a resource). The only > non-resource HTTP service that I'm aware of is forward proxy > functionality. Thus, one way to improve confidentiality of proxies is to > make them path-scoped, and this is what the "Modernizing" draft does. > > These proposals are not mutually exclusive. Path-scoped proxies have > other benefits, and unprompted authentication could be useful for other > services with inflexible paths (e.g. .well-known/ resources). However, > given the overlapping use cases, these drafts should probably be discussed > together. > > --Ben > > [1] https://datatracker.ietf.org/doc/draft-schwartz-modern-http-proxies/ > [2] https://www.w3.org/TR/capability-urls/ >
Received on Tuesday, 18 October 2022 21:49:13 UTC