Hello,

Looking for advice on the behavior to implement in a large-scale
(10'000s or 100'000s of simultaneuous subscriptions) PuSH subscriber
after its own outage (planned or unplanned....):

i.e what shoud this subscriber do when he detects that he's gone
offline for some time ?

Should he  - to make sure that it wasn't unsubcribe of any
subscription by the hub  during the outage - reexecute all
subscriptions previously made ?

It seems to me that this behavior would overload the hub but I don't
see any other to make sure the subscriber doesn't lose any
subscription previously made.

My question is based on draft 0.3 of protocol: both parags 6.2.1
(verification details) & 7.3 (content distribution) leave some room
for interpretation (i.e SHOULD retry for 6 hours for 6.2.1 ;  SHOULD
retry for "reasonable period of time" for 7.3)

My deduction: if you want to be 100% to remain notified for all feeds,
just resubscribe.

Let me know if I misinterpret the specs

regards
didier

Reply via email to