[Webtransport] Re: [Interop] Public WebTransport Test Server (PyWebTransport)

[lemonsterfy <[e-mail]>]

Hello WG,

The public interop server (interop.wtransport.org) has been updated to
track draft-ietf-webtrans-http3-15.

This deployment enforces Draft-15 session admission and flow control
boundaries. Specifically, it executes ALPN negotiation via the
WT-Available-Protocols / WT-Protocol structured fields [Sec. 3.3] and
advertises the SETTINGS_WT_ENABLED and empty reset_stream_at transport
parameters. The state machine now issues immediate
WT_FLOW_CONTROL_ERROR aborts upon data credit violations or
unnegotiated flow control intent [Sec. 5].

The /webtransport/devious-baton endpoint remains operational and has
been synchronized with these updated protocol constraints.

lemonsterfy
On behalf of The WTransport Authors

On Tue, Feb 10, 2026 at 11:51 PM lemonsterfy <[e-mail]> wrote:
>
> Hi Alan,
>
> The Devious Baton endpoint is now live.
>
> - Endpoint: https://interop.wtransport.org/webtransport/devious-baton
> - Spec Compliance: draft-frindell-webtrans-devious-baton-00
>   <https://datatracker.ietf.org/doc/draft-frindell-webtrans-devious-baton/>
>
> Implementation Notes
> This deployment runs on our newly refactored Rust protocol core
> (v0.13.0). Key implementation behaviors include:
>
> - Flow Control: Immediate session termination (DA_YAMN) upon credit
> violation [Section 4.1].
> - Payloads: Randomized padding generation and validation [Section 4.3].
> - State Machine: Strict enforcement of stream type transitions
> (Uni/Bidi) [Section 4.2].
> - Validation Scope: Predictive baton validation is waived for
> Peer-Initiated Bidirectional Streams. Deterministic stream correlation
> in concurrent (`count > 1`) scenarios requires auxiliary state
> tracking. To preclude false positives, the server validates payload
> format but accepts the peer-provided baton value.
> - Datagram MTU: The requirement to echo an identical baton message
> [Section 4.4] conflicts with Path MTU limits if the peer sends
> extensive padding [Section 4.3]. In such cases, our server suppresses
> the datagram response to adhere to transport constraints.
>
> The Devious Baton protocol proves highly effective for verifying
> concurrent stream logic and flow control accounting.
>
> Verification
> To facilitate testing, we provide a standalone client compatible with
> draft-ietf-webtrans-http3-14 and latest WG consensus:
> <https://github.com/wtransport/pywebtransport/blob/main/examples/interop/tests/test_devious_baton.py>
>
> We welcome any interoperability feedback.
>
>
> Best regards,
>
> lemonsterfy
> On behalf of The WTransport Authors
> On Tue, Jan 6, 2026 at 1:07 PM lemonsterfy <[e-mail]> wrote:
> >
> > Hi Alan,
> >
> > Thank you for the suggestion. We recognize the value of the Devious
> > Baton protocol (draft-frindell-webtrans-devious-baton-00) for
> > verifying flow control accounting and stream concurrency management.
> >
> > We will deploy this endpoint on our public interop server to
> > facilitate broader interoperability testing.
> >
> > Our implementation will strictly adhere to the draft's semantics:
> >
> > * **Endpoint Path:** Exposed at `/webtransport/devious-baton` as
> > recommended [Section 3].
> > * **Flow Control:** We will enforce strict credit checks. Sessions
> > will be terminated with the `DA_YAMN` (0x01) error code upon stream
> > limit violations, rather than blocking [Section 4.1].
> > * **Stream Logic:** Baton passing will be handled based on strict
> > stream type and initiator derivation rules [Section 4.2].
> >
> > We will notify the Working Group once the endpoint is available.
> >
> > Best regards,
> >
> > lemonsterfy
> > On behalf of The WTransport Authors
> >
> > On Tue, Jan 6, 2026 at 5:30 AM Alan Frindell <[e-mail]> wrote:
> > >
> > > Hi, this is great!.  Is it possible to add a devious-baton endpoint as well?
> > >
> > > https://www.ietf.org/archive/id/draft-frindell-webtrans-devious-baton-00.html
> > >
> > > Thanks
> > >
> > > -Alan
> > >
> > > On 1/5/26, 11:03 AM, "lemonsterfy" <[e-mail] <mailto:[e-mail]>> wrote:
> > >
> > >
> > >
> > > Hello WG,
> > >
> > >
> > > We have deployed a public WebTransport test server running PyWebTransport.
> > >
> > >
> > > This endpoint tracks draft-ietf-webtrans-http3-14 and the latest
> > > Working Group consensus to facilitate protocol compliance
> > > verification.
> > >
> > >
> > > Infrastructure
> > >
> > >
> > > * Public Instance: https://urldefense.com/v3/__https://interop.wtransport.org__;!!Bt8RZUm9aw!9xqW44OpaWPOdjJpdyJF3byvS5P0h_kM157vmR-MHzocJfQG02-3tI1DLCfRpc3T5YKnYrBJ9hLkR9M$ <https://urldefense.com/v3/__https://interop.wtransport.org__;!!Bt8RZUm9aw!9xqW44OpaWPOdjJpdyJF3byvS5P0h_kM157vmR-MHzocJfQG02-3tI1DLCfRpc3T5YKnYrBJ9hLkR9M$> (IPv4 / IPv6 Dual-Stack)
> > > * OCI Image: ghcr.io/wtransport/interop-server:latest (UDP 4433)
> > >
> > >
> > > Endpoints
> > >
> > >
> > > * /echo: Bidirectional stream and datagram reflection.
> > > * /stats: Session statistics and negotiated transport parameters (JSON).
> > > * /status: Global server health metrics.
> > >
> > >
> > > We welcome interoperability reports from other client implementations.
> > >
> > >
> > > Best regards,
> > >
> > >
> > > lemonsterfy
> > > On behalf of The WTransport Authors
> > >
> > >
> > > --
> > > Webtransport mailing list -- [e-mail] <mailto:[e-mail]>
> > > To unsubscribe send an email to [e-mail] <mailto:[e-mail]>
> > >
> > >
> > >