@OzzLink wrote:
@Zhukov211 wrote:Honestly it defies all logic! It’s as if they want the game to fail.
From a spending POV I think it's logical for them. I think the model they have at the moment is something like: rise a cloud vm with the BF server and the moment the match is done it shuts down, and that repeats itself for the rest of the servers we get matched in. Instead of keeping the vm alive nonstop it will close it thus EA wont have to pay for the resources the vm is using. if persistent. In short you pay what you use and EA thinks it will lose money if the servers will be persistent. I'm 100% sure they could optimize costs if they put a little effort into it even if the servers are persistent but I have little hope with DICE.
100% this. !
@RayD_O1 wrote:
@OzzLink wrote:
@Zhukov211 wrote:Honestly it defies all logic! It’s as if they want the game to fail.
From a spending POV I think it's logical for them. I think the model they have at the moment is something like: rise a cloud vm with the BF server and the moment the match is done it shuts down, and that repeats itself for the rest of the servers we get matched in. Instead of keeping the vm alive nonstop it will close it thus EA wont have to pay for the resources the vm is using. if persistent. In short you pay what you use and EA thinks it will lose money if the servers will be persistent. I'm 100% sure they could optimize costs if they put a little effort into it even if the servers are persistent but I have little hope with DICE.
100% this. !
Well, yes and no. Are they trying to save money? Possibly. Is it justified? Not really..
you reserve say 100 servers and then at peak you have servers you can scale to. I.e 150 servers.
the 100servers you have two pricing options.
resting cost. Price per minute to keep the servers ready.
serving cost. Price on top, for streaming data. I.e people playing.
you pay for the 100 servers regardless.
during peak hours you then have servers spin up based on demand. And they will cost the full price.
Now dice /ea likely have 1000s of servers in there own data centres. Or under agreement with say Amazon.
so they are paying at least a fixed price for N servers
What you do with those servers is entirely up to you. There is zero justification from an server architecture point of view to not have the server run two rounds per match swapping sides.
there is no pricing justification for not having an in intelligent permemant lobby that groups and manages users and passes them of to game servers.
@FlibberMeister wrote:
@RayD_O1 wrote:
@OzzLink wrote:
@Zhukov211 wrote:Honestly it defies all logic! It’s as if they want the game to fail.
From a spending POV I think it's logical for them. I think the model they have at the moment is something like: rise a cloud vm with the BF server and the moment the match is done it shuts down, and that repeats itself for the rest of the servers we get matched in. Instead of keeping the vm alive nonstop it will close it thus EA wont have to pay for the resources the vm is using. if persistent. In short you pay what you use and EA thinks it will lose money if the servers will be persistent. I'm 100% sure they could optimize costs if they put a little effort into it even if the servers are persistent but I have little hope with DICE.
100% this. !
Well, yes and no. Are they trying to save money? Possibly. Is it justified? Not really..
you reserve say 100 servers and then at peak you have servers you can scale to. I.e 150 servers.
the 100servers you have two pricing options.
resting cost. Price per minute to keep the servers ready.
serving cost. Price on top, for streaming data. I.e people playing.you pay for the 100 servers regardless.
during peak hours you then have servers spin up based on demand. And they will cost the full price.
Now dice /ea likely have 1000s of servers in there own data centres. Or under agreement with say Amazon.
so they are paying at least a fixed price for N servers
What you do with those servers is entirely up to you. There is zero justification from an server architecture point of view to not have the server run two rounds per match swapping sides.
there is no pricing justification for not having an in intelligent permemant lobby that groups and manages users and passes them of to game servers.
This is not how cloud service work my friend. IaaS or PaaS is pay what you use and they have dynamic scaling. I have no idea what kind of infra they have setup but the reasoning to cloud service is that that you pay for what you use, I use both Azure and AWS at work. You dont pay for 100 servers slots, this is not the 1990 to reserve vm slots, the data centers these day can accommodate a lot. The main factor for the price is the dynamic range for the resources you will use (ram / cpu / bandwith) and the type of disks used and the amount ofc
@OzzLink wrote:
@FlibberMeister wrote:
@RayD_O1 wrote:
@OzzLink wrote:
@Zhukov211 wrote:Honestly it defies all logic! It’s as if they want the game to fail.
From a spending POV I think it's logical for them. I think the model they have at the moment is something like: rise a cloud vm with the BF server and the moment the match is done it shuts down, and that repeats itself for the rest of the servers we get matched in. Instead of keeping the vm alive nonstop it will close it thus EA wont have to pay for the resources the vm is using. if persistent. In short you pay what you use and EA thinks it will lose money if the servers will be persistent. I'm 100% sure they could optimize costs if they put a little effort into it even if the servers are persistent but I have little hope with DICE.
100% this. !
Well, yes and no. Are they trying to save money? Possibly. Is it justified? Not really..
you reserve say 100 servers and then at peak you have servers you can scale to. I.e 150 servers.
the 100servers you have two pricing options.
resting cost. Price per minute to keep the servers ready.
serving cost. Price on top, for streaming data. I.e people playing.you pay for the 100 servers regardless.
during peak hours you then have servers spin up based on demand. And they will cost the full price.
Now dice /ea likely have 1000s of servers in there own data centres. Or under agreement with say Amazon.
so they are paying at least a fixed price for N servers
What you do with those servers is entirely up to you. There is zero justification from an server architecture point of view to not have the server run two rounds per match swapping sides.
there is no pricing justification for not having an in intelligent permemant lobby that groups and manages users and passes them of to game servers.
This is not how cloud service work my friend. IaaS or PaaS is pay what you use and they have dynamic scaling. I have no idea what kind of infra they have setup but the reasoning to cloud service is that that you pay for what you use, I use both Azure and AWS at work. You dont pay for 100 servers slots, this is not the 1990 to reserve vm slots, the data centers these day can accommodate a lot. The main factor for the price is the dynamic range for the resources you will use (ram / cpu / bandwith) and the type of disks used and the amount ofc
It is how some set ups work. I don’t disagree that there are alternatives.
having priced up servers to secure availability and meet expected peak demands, this is a common pricing strategy.
my point is, there isn’t really a justification for a rubbish poorly executed game flow. You can set up the most optimised server architecture you want and pricing shouldn’t really make much of a difference.
For Example, pre-game lobbies don’t need to be actual game servers. They could/should just be a database. I.e a list. There is zero reason why you couldn’t bring all previous players from a game into the next game, and send them to a newly spun game server together.
