Draft text for recharter #15
Conversation
There was a problem hiding this comment.
Pull Request Overview
This PR updates the MASQUE working group charter to reflect its current status as a concluding working group. The charter is being rewritten from an active development phase to a wrap-up phase, focusing on completing existing deliverables rather than accepting new work.
Key Changes:
- Restructured from an open-ended charter to a completion-focused charter
- Listed 8 specific remaining deliverables the working group is finishing
- Added explicit statement that no new work items will be accepted
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Co-authored-by: Copilot <[email protected]>
|
In "Status" for draft-ietf-masque-quic-proxy it says "Ready for WGLC", but last call was initiated on October 10. |
Woops, thanks! That's an artefact of us forgetting to update the document state in Datatracker. Now fixed. |
|
|
||
| The MASQUE working group has developed mechanisms that allow configuring and concurrently running multiple proxied stream- and datagram-based flows inside an HTTP connection. These include CONNECT-UDP and CONNECT-IP, collectively known as MASQUE. MASQUE leverages HTTP semantics, multiplexes flows over streams, uses a unified congestion controller, encrypts flow metadata, and enables unreliable delivery suitable for UDP and IP-based applications. | ||
|
|
||
| The MASQUE working group is now in the process of completing its remaining deliverables: |
There was a problem hiding this comment.
This list mixes stuff that's basically done with stuff that has not even been adopted. I don't think this is particularly useful, especially for those not following MASQUE closely.
Suggest we take it as given that adopted documents will be completed and focus on the actual work items this WG thinks are remaining.
| recovery considerations, and intarea for IP Proxying. | ||
|
|
||
| When these deliverables are complete, the working group will either recharter to reflect additional work or close. | ||
| Many network topologies lead to situations where transport protocol proxying is beneficial—for example, enabling endpoints to communicate when end-to-end connectivity is not possible, or applying additional encryption such as in VPNs. Proxying can also improve client privacy by hiding client IP addresses from target servers. Existing proxying technologies (e.g., SOCKS and HTTP CONNECT) have limitations: SOCKS lacks encryption, and HTTP CONNECT is limited to TCP. |
There was a problem hiding this comment.
I don't particularly see the need to edit this text. Practically, its created a large diff with no real semantic difference.
There was a problem hiding this comment.
I would tend to agree that we should leave as much the same as possible just to reduce diffs to review
There was a problem hiding this comment.
Ah, this is a tooling artefact. Reviewing the rich diff (the page icon at the top) will be much more helpful, but I will fix regardless.
| When these deliverables are complete, the working group will either recharter to reflect additional work or close. | ||
| Many network topologies lead to situations where transport protocol proxying is beneficial—for example, enabling endpoints to communicate when end-to-end connectivity is not possible, or applying additional encryption such as in VPNs. Proxying can also improve client privacy by hiding client IP addresses from target servers. Existing proxying technologies (e.g., SOCKS and HTTP CONNECT) have limitations: SOCKS lacks encryption, and HTTP CONNECT is limited to TCP. | ||
|
|
||
| The MASQUE working group has developed mechanisms that allow configuring and concurrently running multiple proxied stream- and datagram-based flows inside an HTTP connection. These include CONNECT-UDP and CONNECT-IP, collectively known as MASQUE. MASQUE leverages HTTP semantics, multiplexes flows over streams, uses a unified congestion controller, encrypts flow metadata, and enables unreliable delivery suitable for UDP and IP-based applications. |
There was a problem hiding this comment.
Similar to above, this is a rephrasing that looks rather large but seems to mostly be, an unfortunate drop of a clear statement of the WGs primary goal.
| * Services that a proxy initiates without any prompt from a client are out of scope. | ||
| * Extensions that solely relate to generic proxying functionality, and are not specific to the core MASQUE documents, are out of scope. | ||
| * Specifying proxy server discovery mechanisms is out of scope. New congestion control and loss recovery algorithms are also out of scope. However, the working group will consider implications of tunneling protocols with congestion control and loss recovery over MASQUE proxies, and may issue recommendations accordingly. | ||
| * IP multicast is out of scope. Designs need not explicitly preclude multicast, but they will not focus on multicast-specific features. |
There was a problem hiding this comment.
This can all be condensed down into the later line of "the MASQUE WG is not accepting any work item not listed above.
|
|
||
| MASQUE will continue to coordinate closely with HTTPBIS, QUIC, TLS, INTAREA, ICCRG, CCWG, and IEEE 802.3 as appropriate. | ||
|
|
||
| The MASQUE working group is expected to conclude once its current documents are published. |
There was a problem hiding this comment.
nit: current is awkward here if no documents have been adopted before the charter is landed. Easier to refer to the work items/documents described in the charter.
|
|
||
| These documents are being finished in line with MASQUE’s previous charter: | ||
|
|
||
| * The intended status is Standards Track unless otherwise noted, but the WG may downgrade if it believes that is appropriate for the ultimate document maturity level. |
There was a problem hiding this comment.
This was useful in the early days of the WG but seems like we can strike this out now. If the intended status is useful, just state that in each of the handful of work items
| 7. An extension to CONNECT-UDP and / or CONNECT-IP to enable compression | ||
| 8. A definition of qlog nomenclature for existing MASQUE mechanisms | ||
|
|
||
| These documents are being finished in line with MASQUE’s previous charter: |
There was a problem hiding this comment.
This seems like some awkward phrasing TBH. Per my comments below, I think we can strike out most of the old stuff and just state things directly and concisely
| 5. An informational architecture document about MASQUE | ||
| 6. An extension to CONNECT-UDP for ECN and DSCP support | ||
| 7. An extension to CONNECT-UDP and / or CONNECT-IP to enable compression | ||
| 8. A definition of qlog nomenclature for existing MASQUE mechanisms |
There was a problem hiding this comment.
As much as I'm a qlog enthusiast, I've never needed to define MASQUE events. This is in despite of having had to implement and debug production MASQUE services. Is there sufficient demand for this such that it can potentially keep the WG from closing until the work item is completed?
Later the charter text says "Any proposals for new extensions or functionality should be directed to appropriate alternative working groups such as HTTPBIS, QUIC, or INTAREA, depending on the nature of the work." Maybe we can just kick the can to those folks to pick this up if someone thinks it would be useful.
This is some proposed text for rechartering Masque to bring draft-schinazi-masque-proxy within scope and begin sunsetting the WG.
The charter keeps the existing WG documents within scope:
The charter also enumerates the recently proposed work as in scope: