Population... underestimation?

[Toppopia]

Originally Posted by AThreatToYou

What I'm more worried about is how a single server can process, run, and host a game tracking 6000 players at once. From my memory, however, a single "continent" was a single server box in PS1, so they could easily chunk a single server into basically 3 servers linked to a final server that processes all of the overflow over a LAN network. Prioritizing the assets can help to where things like map data, resource amounts, experience gains, kill credits or certification gains are all dumped to another machine as soon as the first one gets loaded. Still, other modern engines such as CryEngine 2/3, Unreal Engine, or Source run into serious problems as soon as all of those 2000 players start shooting, throwing grenades, ducking, dieing and re-spawning. The most powerful server boxes and most modern engines these days just can't handle 2000 players, let alone 1000 players, let alone 200 players. Most engines start to die as soon as they get past 100 players--and this isn't network lag, it's where the server just can't keep track anymore. PlanetSide 2's achievement of 2000 players all at once is, on its own, freaking amazing. I honestly believe we just have to leave it to the programmers who know the ins-and-outs of process management to push the player limit as high as it can reasonably go with today's hardware in a situation where it isn't limited by today's software.

Now imagine all 2000 people running to 1 spot, then fighting eachother, we might end up blowing the whole server complex into a million pieces

[Mechzz]

Originally Posted by Toppopia

Now imagine all 2000 people running to 1 spot, then fighting eachother, we might end up blowing the whole server complex into a million pieces

If that happened, they'd have to go to CCP and buy their "time dilation" technology used to handle big space fleet battles in EVE. That would give SOE the chance to use a new strapline:

Planetside - bullet time on a planetary scale.

That might be fun, really!

[TeaReks]

Planetside 1 was able to handle 500? people per cont without too many problems on tech that's at least 9 years old. The other engines you brought up are not built from the ground up for MMO needs unlike Forgelight.

SOE is not some startup company trying something new. They had EQ, the first real 3d mmo. Planetside broke the mold of FPSs. I think these guys understand better than most what it takes to make a game like this.

[Astrok]

Originally Posted by elfailo

2000 per continent isn't going to happen. Beta will show why not. I'm guessing they're going to halve it within 3 months.

100 years ago they told us that flying trough the air wasnt going to happen.and they laugh at the people who wanted to try it.
Right now most people travel by plane.

So why 2000 people cant be happening?
Because u say so?

[Otleaz]

Originally Posted by Astrok

100 years ago they told us that flying trough the air wasnt going to happen.and they laugh at the people who wanted to try it.
Right now most people travel by plane.

So why 2000 people cant be happening?
Because u say so?

That isn't quite the same thing.

[aceshigh]

Relax guys. With their newly discovered VS technologies, 2000 won't be a problem.

[Astrok]

Originally Posted by Otleaz

That isn't quite the same thing.

Tell me the differents then.

for me its the same principle.

Originally Posted by aceshigh

Relax guys. With their newly discovered VS technologies, 2000 won't be a problem.

I think that if they want they can go even higher.

[Krawanan]

what an interesting thread this is

Im kinda into servers and hosting myself, so when i heard of planetside 2 like 6 months ago, i couldn't believe it, since as some posts in this thread state, it is just not possible with the resources we have now.
but since they could handle 500 players before(9 years ago!)

i think SOE might have found a very efficient way to do this.

i think we'll just have to trust them on their word, also they wouldn't have begun a project like this if they had known it would be impossible, would they ?

and i think they have done enough research and simulations(like with bots) to be sure that is possible, because again, sony would not invest this amount of money in something they wouldn't be 100% sure of was doable

[Landtank]

Originally Posted by Astrok

I think that if they want they can go even higher.

I agree, 2000 seems to be the sweet spot in terms of leaving enough open space on the map, but also making it seem like there is a war going on.

They have been promising 2000 players for too long for them to cut it back, that's how I see it.

[Baneblade]

PlanetSide 1 made the 500 player thing work by using CSHD instead of SSHD. PS2 will apparently be using a hybrid system.

[Kayos]

Obviously it can support 2000 or they wouldn't advertise it.

[Xyntech]

Higby has stated that there would be problems if 1000 players try to stand in one tiny area all at once, so clearly they are aware of some the of the major capabilities and possible limitations of their current system. The big questions are how effectively they will be able to keep too many players from going into the same area at once, and exactly how many players can be in the same area before it becomes a problem.

That mock up I did was essentially just for illustrative purposes, to show how players distributed mostly along front line battles would look. It was mostly for my own thought experiment and self education. A lot of players were trying to divide up the 2000 players evenly across all hexes, which just doesn't sound like it jives with the devs battle line style territory system that they are trying to achieve.

While the mockup does show slightly smaller battles than some may have expected out of PS2, don't forget that this is only a generally even spread across the borders. If Forgelight allows it, we may see some of those larger population battles I put in there being consolidated into 2 or 3 even larger battles, with the rest of the border skirmishes taken up mostly with smaller 5-30 vs 5-30 player conflicts.

We really don't know how the player distribution will work out until we get into beta and test for ourselves how well the territory system works. I feel that my mock up is a pretty good representation of the developers intention (at least based on what I could deduce of their intentions), but that doesn't mean that there isn't a lot of room left open for those battles to be quite a bit more spread out or clustered together than I showed.

I'm not going to hold my breath on PS2 succeeding at 2000 players, but I don't think it's outside the realm of possibility. I also don't think that the 200 - 300 player large battles I illustrated will be that tiny either, if that's the largest we end up with. The first game had a limit of 400-500 players per continent, and it was pretty rare that all of them would go to the exact same spot.

I think that Planetside 2 will tend to have similar, if not slightly larger average battle sizes compared to PS1, but it will go a step beyond by having many more of these battles packed relatively closely together. It will be like if you had 4 PS1 continents worth of population all crammed into one continent, but still spread out in a way more reminiscent of how they were spread out in the first game.

[Kurtz]

Originally Posted by elfailo

2000 per continent isn't going to happen. Beta will show why not. I'm guessing they're going to halve it within 3 months.

^this.

I hope for the best, still.

[kaffis]

As you suggest, I don't think Xyntech's number speculations are an unreasonable snapshot of player distribution.

However, take a look at that, and consider... the three-way at the center is 332 players in that one 5-hex region...

E3 was 100 players in a 7-hex region.

So Xyntech's guess at how 2000 would distribute on the continent already has quadrupled the player density of the most ambitious real-world test we're privvy to.

I'd say that given that, it's a bit hasty to declare that the 2000/continent estimate is already too conservative... Performance was good at what we can sort of assume might be a quarter of the load, and even that is assuming that the server and engine do a good job of partitioning space, so all we have to worry about is local density, and not increasing overhead for distant load.


To work off the border and grid distribution in Xyntech's map, we can see that regions adjacent to faction borders comprise 134 hexes' worth of area. We know that the E3 setup was a little more than 7 hexes' worth of space, with, reportedly, around a hundred players. So E3 density is ~15 players per hex, meaning that if the entire frontline supports a similar average density, you're talking about 2010 players.

Gosh, that number looks familiar, doesn't it? Could it possibly be that the E3 population was tuned around their performance goals? And that those performance goals are what yield a good experience, hopefully with a little slack built in to account for less even spreads?

But no, making assumptions like that would imply the developers know what they're doing, or have done some testing...

[Turdicus]

Originally Posted by elfailo

It's a gut feeling.

I can cough up a bunch of rationalizations like people here like to do so much, but I'm not really afraid of being wrong every once in a while, so I can live with the uncertainty.

Lol elfailo I'll take that bet!