[DiscordArchive] Back in the day weren't there other server handoffs e.g. to BG servers?
[DiscordArchive] Back in the day weren't there other server handoffs e.g. to BG servers?
Original source
I think nowadays there is no real use case for this
rektbyfaith
02-05-2024, 02:37 PM #21Archived author: robinsch • Posted: 2024-02-05T14:37:57.447000+00:00
Original source
I think nowadays there is no real use case for this
Original source
When you can have machines with 80 CPU cores
rektbyfaith
02-05-2024, 02:38 PM #22Archived author: robinsch • Posted: 2024-02-05T14:38:11.960000+00:00
Original source
When you can have machines with 80 CPU cores
Original source
For dungeons and battlegrounds there arguably could be, it's much cheaper to spin up a small container whenever your load gets too high compared to a dedicated massive host
rektbyfaith
02-05-2024, 02:38 PM #23Archived author: Foe • Posted: 2024-02-05T14:38:51.478000+00:00
Original source
For dungeons and battlegrounds there arguably could be, it's much cheaper to spin up a small container whenever your load gets too high compared to a dedicated massive host
Original source
If you lean into a micro service architecture there's definitely a place for automatic up and down scaling
rektbyfaith
02-05-2024, 02:39 PM #24Archived author: Foe • Posted: 2024-02-05T14:39:32.894000+00:00
Original source
If you lean into a micro service architecture there's definitely a place for automatic up and down scaling
Original source
I mean BGs and dungeons take way less time to update compared to large continent maps
rektbyfaith
02-05-2024, 02:39 PM #25Archived author: robinsch • Posted: 2024-02-05T14:39:34.300000+00:00
Original source
I mean BGs and dungeons take way less time to update compared to large continent maps
Original source
For sure, but let's say a node can handle x amount of dungeons, once you hit a threshold you just spin up another small shard
rektbyfaith
02-05-2024, 02:40 PM #26Archived author: Foe • Posted: 2024-02-05T14:40:01.988000+00:00
Original source
For sure, but let's say a node can handle x amount of dungeons, once you hit a threshold you just spin up another small shard
Original source
It's a lot cheaper with dynamic resource allocation
rektbyfaith
02-05-2024, 02:40 PM #27Archived author: Foe • Posted: 2024-02-05T14:40:13.674000+00:00
Original source
It's a lot cheaper with dynamic resource allocation
Original source
From a cloud perspective obviously
rektbyfaith
02-05-2024, 02:40 PM #28Archived author: Foe • Posted: 2024-02-05T14:40:33.476000+00:00
Original source
From a cloud perspective obviously
Original source
it literally costs 1000x to write this compared to renting a big machine for 100 years
rektbyfaith
02-05-2024, 02:40 PM #29Archived author: robinsch • Posted: 2024-02-05T14:40:45.368000+00:00
Original source
it literally costs 1000x to write this compared to renting a big machine for 100 years
Original source
Definitely, not arguing that
rektbyfaith
02-05-2024, 02:41 PM #30Archived author: Foe • Posted: 2024-02-05T14:41:24.988000+00:00
Original source
Definitely, not arguing that