What would really make a difference: the client-server model

[CMDR Cosmic Spacehead]

How about, people can rent a "private server" that basically acts as a host for players to instance too.... If that makes sense....

So say 10 players want to meet up.
Player 1 has a private server that basically follows them around "somehow", allowing players to instance with their server, instead of them.
Players 2-10 connect to the server instead of player 1.
And the server handles everything else.
Or something.



I know nothing of networking. Lol

 

[Drelthar]

If I had to pay per month that would sadly mean me not playing the game anymore. I'd hate that.

 

[Eidolon]

If I had to pay per month that would sadly mean me not playing the game anymore. I'd hate that.

Yes, me too I'm afraid.
Already backed the game and bought the horizon dlc. Along with a sizeable amount of cosmetics from the store.
I don't see why I should pay every month to keep playing as I do now.

 

[krylite]

I think FDev could change the infrastructure to accommodate more CS "nodes" rather than the other way around focused on p2p. It could help with more persistent objects, player assets, not just how many players in an instance. But as has been said before it would cost more. To have the galaxy dreamed about with everything implemented, spacelegs with player objects dropped on location, apartments in stations, cities, etc. , where all this is persistent on tens of thousands of habituated worlds without relying on procedural generation would require buying out a server farm(s) or a data center and also require closer to terabytes of storage on a player client computer. I think they had to choose whether to make the scope of 400 billion systems work or restrict the persisten to a few thousands visitable systems and I'm glad they chose the former otherwise it would just be another 10 to 100 star system game with fantasy space backdrops masquerading as a open space galaxy.

 

[Sylveria]

I'd definitely pay a subscription fee for a client-server solution... IF PvP/PvE flagging were available, that is.

Hell, I'd even pay to rent a private server for PvE. What I won't do is finance other people's concept of EVE-in-Cockpits.

 

[Cosmo]

Persistent NPCs can be done with a peer to peer system.

USS's has nothing to do with a peer to peer system.

You can get instancing issues with client to server games too.

Yes it could help with combat logging, but I can't see that as a major issue anyway.

Seconded.


We won't get client-server, because it costs too much and people don't want to pay a subscription fee. The P2P code will continue to be improved.

 

[Blackcompany]

and the water is wet.
client-server model cost a lot of money to mantain. Maybe you could tell us how FD would cover the cost of such an infrastructure.

Maybe YOU can tell US why a company would bother making an always online, multiplayer game that REQUIRES an internet connection at all times...if they cant afford to do it PROPERLY from the get go? Seems to me, thats the more important question here. Otherwise, the always online is basically, for Solo players, just constant DRM.

 

[Riverside]

Maybe YOU can tell US why a company would bother making an always online, multiplayer game that REQUIRES an internet connection at all times...if they cant afford to do it PROPERLY from the get go? Seems to me, thats the more important question here. Otherwise, the always online is basically, for Solo players, just constant DRM.

Easy to answer, it's a choice with pros & cons either way. Some games do it one way, others do it another way. Just depends on what the devs consider to be a priority.

Sometimes there is no single obviously right answer & you just make a choice. In this case, time has told they made a good decision but it's not perfect. If they'd gone with client/server other stuff would be better but (probably) most of the game would be a little worse.

 

[Chris Simon]

Easy to answer, it's a choice with pros & cons either way. Some games do it one way, others do it another way. Just depends on what the devs consider to be a priority.

Sometimes there is no single obviously right answer & you just make a choice. In this case, time has told they made a good decision but it's not perfect. If they'd gone with client/server other stuff would be better but (probably) most of the game would be a little worse.

Agreed. Considering the Elite is primarily designed as a singleplayer (despite what some try to believe) with some very limited player interaction added on a top, the P2P system is the most efficient way about it.
The only true PvP feature is that we can blow one another up, really. Even the features that are considered "kinda PvP", like Powerplay and BGS aren't really. Most of the interaction is indirect - kind of PvE vs. PvE, meaning destroying your opponent doesn't change anything, you have to simply do PvE against them to succee - for wich the P2P with a matchmaking server is the right choice.
Yes, it would be nice to be able to gather hundreds of players at one spot, but how many times would that be put to use, and I bet PvPers would love to be able to destroy opponent's ship even after they have logged out, but that is about it. Adding a client-server functionality isn't worth the effort, imho.

 

[Bortas]

" - Combat Logging"

Not solving this one either.

Please elaborate. Virtually every other client-server based online-game successfully solved this by simply leaving ungracefully disconnected clients in the instance for a while. Why should this not be possible in ED?

 

[Agony_Aunt]

If any move to a C/S model would require a monthly subscription, then i would be against. Also all of your points OP can be resolved with the hybrid model FD uses.

P2P is probably better for the long term future of the game, and ultimately its transformation into a self-hosting game when the servers shut down, or maybe will allow FD to keep servers running longer due to lower running costs.

So, personal opinion is that best if FD don't change it. (although presume they have no intention of doing so anyway).

Please elaborate. Virtually every other client-server based online-game successfully solved this by simply leaving ungracefully disconnected clients in the instance for a while. Why should this not be possible in ED?

I think he means it can be solved with FD's model as it is. You don't need full C/S to solve the issue. FD has severs. With C/S the server takes over when someone disconnects. FD could add a server side service to do the same.

 

[Cmdr Eagleboy]

Please elaborate. Virtually every other client-server based online-game successfully solved this by simply leaving ungracefully disconnected clients in the instance for a while. Why should this not be possible in ED?

Not really, no one has solved logging that way Real life examples are very few and most of them aren't really suited for ED.

It is possible for FD to do this but they see it as can of worms - veey hairy, bug prone approach. And with potential loss of ships is just not worth it at the moment. Better invest in punishment of logging so it wouldn't happen in first place.

 

[Bortas]

And with potential loss of ships is just not worth it at the moment. Better invest in punishment of logging so it wouldn't happen in first place.

Potential loss of ships is the very point of the method. It is the punishment for logging and prevents it very reliably.

 

[Cmdr Eagleboy]

Potential loss of ships is the very point of the method. It is the punishment for logging and prevents it very reliably.

Not if there's honest disconnect.

 

[MarkAusten]

Not if there's honest disconnect.

Indeed and that is the heart of the entire combat logging problem. As far as I know it is currently impossible to determine the root cause of the disconnect. Basically was it deliberate or accidental? Until that can be reliably determined it doesn't matter if the architecture is peer to peer or client server. You either have to just ignore combat logging or punish every disconnect whether deliberate or accidental. I'll wager that there are more accidental disconnects than deliberate ones in the which case you would be fixing one problem by substituting it with a bigger problem. That doesn't sound like a solution to me.

 

[Bortas]

You either have to just ignore combat logging or punish every disconnect whether deliberate or accidental.

Exactly. It's an all or nothing thing. You can either ignore cheating or accept the odd errenous rebuy screen.

I'll wager that there are more accidental disconnects than deliberate ones in the which case you would be fixing one problem by substituting it with a bigger problem. That doesn't sound like a solution to me.

True, but there's a way to mitigate this... the mechanics can be limited to PvP combat, where deliberate disconnects are probably several orders of magnitude more common than accidental ones. Would be a good compromise in my book.

 

[Surprised Brony]

Exactly. It's an all or nothing thing. You can either ignore cheating or accept the odd errenous rebuy screen.

Usually in games we also accept that some erroneous cheating happens. Not always all or nothing.

Worst part? This doesn't directly hurt you or your in-game avatar at all. I want it fixed, but going to such an extreme and potentially punishing innocent players for something that doesn't directly affect you? Come on.

 

[sollisb]

Effort: Immense
Risk: Can't prove who disconnected and why. FDev servers disconnected me yesterday for example
Reward: Absolutely nothing.

Aint worth it. Move on

 

[Brrokk]

Groan. So this is actually a disguised combat logging thread?

I'd like to give an example to show how little I care about clogging. There's a long-standing bug which causes station announcer audio to be slightly out of step with events if you're using a docking computer. (Basically, they're late telling you which pad number you've got; it's only said after your DC engages). It's something many people don't even notice, but it spoils things a bit for me. I consider that annoyance on each docking much more significant than combat logging, which I've never seen.

Clogging is cheating, but it never affects me so I don't care. I would be emphatically against FD taking the risk of making fundamental changes to the net architecture to deal with something minor that I don't care about. Not that it matters, there's no reason why FD would need my approval.

 

[TheSynopticVision]

Agreed. I wanted to ask an hypothetical question about a possible solution to a number of problems. Maybe these could be fixed by CS structure? Would they consider it? (for the record: I forgot the question mark when writing the title )

What the thread actually became was a collection of snobbish "yawn, you know nothing, moving on" remarks or yet another boring CLogging debate (and to think that I put it last in my original list, in order not to emphasize it too much!).

A mod can now put this thread out of its misery.