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

[Igor Lubashev <[e-mail]>]

@kazuho I would say that reachability/fitness-for-service is a shared problem for all _client migrations_, not a specific SPA problem.

We introduced `disable_active_migration` to warn clients of situation where there is a high risk to reachability/fitness-for-service due to migrations after network attachment changes.

SPA is an attempt to improve reachability.  It does help with reachability after migration for some important types of deployments, and it helps with reachability even without migration by removing load balancers as possible points of failure (I include BGP-based anycast in "load balancers").  SPA does _not_ help with fitness-for-service after a migration.

This discussion is about deployment where there is a high risk to fitness-for-service due to migrations after network attachment changes.  SPA can be helpful there (see above), but migrations after network attachment changes is still inadvisable.


In terms of practical considerations, embedded CDN nodes are certainly examples of deployments with high risk to reachability _and_ fitness-for-service due to migrations after network attachment changes.  However, even public-facing anycast CDN nodes can have a similar fitness-for-service risk due to attachment changes in areas with poor network interconnections (some parts of Asia, for example).

-- 
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-657903203