ID:1859815
 
I know nothing about servers except for the most basic way they work, so I wanted to ask a question. If I was purchase a shell server of a pretty good capacity, how many players would it hold without lag/with max lag? And is there a way to get more expensive shell servers that can hold more people?
Lag wont be the issue, the issue is how well the game is done so it wont use alot of CPU, the 10 dollars Shell servers with a very well optimizated game can hold 150-200 people. If the game its very unpolished and buggy, maybe 25-50.
In response to Zasif
Zasif wrote:
Lag wont be the issue, the issue is how well the game is done so it wont use alot of CPU, the 10 dollars Shell servers with a very well optimizated game can hold 150-200 people. If the game its very unpolished and buggy, maybe 25-50.

It's hard to believe that a game can reach >150 players, I have seen 125, and the Shell Server was quite expensive.
Even if the game is pretty well optimized, it'd be impossible due to BYOND's performance. I'm not saying it's a bad engine, but it's not meant to hold that amount of players in a single player due to the few client-thread procs.

*huhuhuhuhuhuhuhuhuhuhuhuhuhuhu*
http://www.byond.com/games/S10Games/DragonKi
In response to Zasif
Zasif wrote:
*huhuhuhuhuhuhuhuhuhuhuhuhuhuhu*
http://www.byond.com/games/S10Games/DragonKi

Well, yeah, I just saw that one.
But it's a simple rip, with a bunch of commands that do nothing, and all you have to do is train. There are no complex systems or anything going on...

But yeah, you got a point: games can actually hold 150 players.
Look at the complexity of SS13. I've seen servers reach upwards of 120 people on them.
Hmm I see. Is there a way to purchase more expensive servers that can hold maybe a thousand people? Or would it not be possible due to the way BYOND is set up
In response to AnimeBeyond
AnimeBeyond wrote:
Hmm I see. Is there a way to purchase more expensive servers that can hold maybe a thousand people? Or would it not be possible due to the way BYOND is set up

Not possible due to BYOND.

Contact me here - http://byondpanel.com I have a range of Dedicated servers ranging from Enterprise CPU's Xeon E3, E5's to i5 - i7's all with DDoS protection.

Let me know what you need/want and I'll work out a nice deal for you.
It might be possible to hold 1000 people per server, but the biggest games have typically only reached up to a little over a hundred. I think that some of them could handle more with more efficient code, and of course not increasing the frame rate helps a lot.
In response to Lummox JR
Not true Lummo, unless its a Chat Window Maybe? and Low FPS really? we are in 2015 anything below 30-60 shoudnt be played, and if you have alot of things that sucks out all the CPU I dont see any game no matter How awesome coded it is will reach over 200players per server.
In response to Zasif
Truthfully I don't know what the realistic limits are. The games that have pushed those limits the most have sometimes had in excess of 100 players. However, a lot of those games also have been hamstrung by inefficient design ('cause let's face it, a lot are rips); the flip side is that not many of them have higher frame rates.
When I had about 175 people CPU was stedy at 30-60% usage depending on what actions people was doing, once we reached 215 players it started going -10%-20% Fps thats when we had to reboot. xD
I think it's possible, depending on the game in question. The problem, like Lummox suggested, is inefficient practices: spawn() abuse, loop abuse, useless variables, using Del() instead of the GC - the list goes on.

Since programming efficiently takes time, I think you'll find that most people choose the alternative (rapid development) and almost never consider efficiency until it becomes a problem, versus meticulously designing their game to run at its best from the get-go.
In response to FKI
I would suggest that fixing a program for efficiency takes time. Designing it to be efficient from the outset isn't all that hard; it just takes mindfulness.

Even projects designed for efficiency might discover down the road that a different way of doing something might improve performance, so refactoring is inevitable. But the less you have to do when the time comes, the better--and the later that time will come.