Re: [quicwg/base-drafts] The number of Connection ID is unbounded to NEW_CONNECTION_ID sender (#2403)

[erickinnear <[e-mail]>]

To add to this, I think the fundamental design (which, admittedly, a TP confuses a bit) is that the issuer should issue (and maintain with new ones when retired) as many connection IDs as it’s willing to maintain state for. 

So, if the consumer of the CIDs uses all the new ones and doesn’t retire them, it will eventually run out of new ones to use. 

My understanding of the suggestion to replenish when a new one is observed is simply to lower the chance of having the consumer of the CIDs run out while RETIRE_ frames are in flight (or potentially lost). If the issuer is already at the max they’re willing to maintain, then only issue new ones when old ones are retired. The associated assumption here is that a consumer would not be advised to retire all CIDs at the same time.  

-- 
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/2403#issuecomment-460125027