Amazon's AWS website breaks down the cost of running dedicated vs on-demand servers.
| Cost of continuously running servers | $36,270 | ||
| Cost w/Autoscaling and On-Demand | $18,150 | 50% savings over continuously running servers | |
| Cost w/Auto Scaling and Spot (estimated) | $5,430 | 85% savings over continuously running servers |
EA is saving at least 50% using on-demand servers vs continuously running dedicated servers. They save even more if on-demand spot servers are available, which is unused server capacity Amazon offers at a discount.
https://aws.amazon.com/gamelift/pricing/
It's no wonder EA loves the new matchmaking system where servers are shut down at the end of every round.
@OskooI_007 wrote:Amazon's AWS website breaks down the cost of running dedicated vs on-demand servers.
Cost of continuously running servers $36,270 Cost w/Autoscaling and On-Demand $18,150 50% savings over continuously running servers Cost w/Auto Scaling and Spot (estimated) $5,430 85% savings over continuously running servers EA is saving at least 50% using on-demand servers vs continuously running dedicated servers. They save even more if on-demand spot servers are available, which is unused server capacity Amazon offers at a discount.
https://aws.amazon.com/gamelift/pricing/
It's no wonder EA loves the new matchmaking system where servers are shut down at the end of every round.
I do t think it’s that simple though. I don’t see any reason why what those servers are doing when up and running couldn’t be MM or Dedicated servers.
Or some other configuration I.e a mix.
you can scale servers in all sorts of ways. And quite often th price is also broken down into pending and active. I.e based on how much bandwidth is being used.
also you tend to have to outlay a lot of money upfront when dealing with large numbers like BF. So your already talking dedicated servers farms to your business.
basically cost might play.a part in it. But it really shouldn’t
@RayD_O1 At this stage its probably cheaper to run fewer persistent servers that match up with the current server population compared with autoscaling / on demand if it was set up to allow for the capacity at launch.
For example, I know that auto scaling allows for the resources to be transferred, and thats the whole point, but I would imagine there is still a limit to that contractually.
I don't know, I don't know anything about server rental, but I somehow doubt EA are only paying 5% of what they paid at launch because 95% of the player base has gone.
@FlibberMeister wrote:
@RayD_O1For sure,
I’m just saying server architecture, scaling etc shouldn’t really be determined by the cost of MM vs Persistent servers. They really should be experts in this domain and likely already have to manage there own server farms. Especially at the scale they operate at.
For private servers, the cost can be passed on to the rentee.
For the dedicated dice servers they can spin up persistent servers based on demand. Doesn’t matter if those servers are MM or persistent lobbies.
It’s not like a server is actually turned on and off again after each round. That would be madness. Rather the software might turn over, but the server instance itself will still be up and running.
The MM vs persistent server decision is I think purely a design decision based on how they want people to play and nothing to do with saving money in server architecture.
I understand what you are saying and you make some very good points there.
I would say if the server decision was design based on how they want people to play, then they have seriously miscalculated judging by the response on these forums to MM, and it's borne out by the ever dwindling player population.
Players want / expect to be able to play the game the way they want to, not the way MM forces them to play.
Imho DICE really need to think again and act quickly to reverse their decision by replacing MM with a server browser / persistent server system otherwise BF2042 will be dead in the water.
