@OskooI_007 wrote:Amazon terminology breaks down like this.
Instance = Server hardware being leased
...
I don't think that Amazon really introduces their own terms. But everywhere*) else an 'instance' is not the leased hardware (aka a physical server) but a subset of say vCPU cores, memory, mounted disk volumes, sometimes also network IO bandwidth. So no 1:1 relationship to hardware. If you delete a POD and create a new one it might end up on a completely different server (hardware), depending on the configuration of the cloud.
I don't think that it would make sense to have one big instance and run many game sessions on them. I would rather implement this by using many smaller instances, one for each game server process 'session'.
I can only tell what I would do on our OpenStack instance if I would have to design such a thing, but then I don't have to pay on a 'per instance' basis but can spawn as many instances as I think are needed and OpenStack will distribute the PODs as it sees fit. 😉
Never had to deal with an external cloud infrastructure, so things might be different when working in the Amazon cloud.
*) everywhere means as far as I know. So this can be rather limited 😉
Thanks for sharing all that very detailed information, you guys are very knowledgeable on this subject and obviously know your stuff.
I sincerely hope that DICE will do whatever it takes to allow us all to play the game without the hassle currently caused by MM.
