Re: [quicwg/base-drafts] disable_active_migration with SPA (#3765)

[Kazuho Oku <[e-mail]>]

@igorlord 
> To summarize, it seems that people are generally not philosophically opposed to fixing this issue, but the solution needs to be simple and be compatible with the current behavior.

FWIW, I am fine with adding a field to the SPA TP indicating if active migration is disabled on that address.

At the same time, I am wary about recommending endpoints to use the current local address when migrating to a server-preferred address. That is because it changes how QUIC endpoints can be implemented.

At the moment, developers can implement a QUIC client that follows all recommendations (i.e. SHOULD), by letting the operating system choose the local address. You can simply create a local socket that is bound to 0.0.0.0, and let the operating system choose the source address of the packets that will be sent.

That approach would no longer be considered as recommended, if we are to have a SHOULD (or MUST) that suggests client to use the same address when migrating to SPA. Clients would be recommended to track and intentionally select the source IP address of outgoing packets.

Therefore, while I am not necessarily opposed to suggesting clients to continue using the current local address, my preference goes to *not* using a RFC 2119 keyword for the purpose.

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/quicwg/base-drafts/issues/3765#issuecomment-665594676