Player warning: Crowd Limiting System coming up

[Beach]

More on the lowering of the mesh detail.... when an area gets crowded.... the server sends a message to EVERY client that is within the "crowding zone".... then every client who get that, lowers the detail of ALL the meshes down A GREAT DEAL....


Then, this will solve the lag.... there is no need to prevent players from entering a particular zone.... if we are talking about chamlot, the current system of GS zones is more than fine.... Once u start to limit players going into a "crowding" zone, and force existing players to exit, it will get so rediculous i don't even dare to think of the consequences....



And, if lowering down the detail of the mesh (and I emphazise again... by a lot..... maybe lower down the detail by 80-90%) doesn't help.... and u insist on limiting the number of players in a "crowding zone".... then, all I can say is that....




Your FC Server Cannot Handle The Load, Not The Player's Clients.

[________________Luke]

The whole thing makes me laugh a lot.

Imagine this scenario in FC's offices:

Gaute: So, Sales .... how are we looking over the next 12 months? I guess we should be ok as we're making about $4.5mill a year based on 30,000 players subscribed. Profit after tax and overheads etc, should leave us with about 1.5mill. Great

Sales: Actually, not that great really. It seems that we have a steady flow of customers leaving.

Gaute: Oh, why?

Sales: Well, we aren't selling many more memberships. I reckon the actual player base may be disgruntled and telling others not to bother. In fact, a lot of the forums talk about our current players waiting for something called SWG and AC2?

Gaute: Whaddaya mean?

Sales: Well, it seems we kinda promised a lot, but er, the Developers/Programmers just haven't delivered Gaute. A storyline was promised and even started. But we erm, couldnt keep it goin. The vids looked cool though! A lot of players hadnt even realised that the storyline had even finished... thats how much they were following what little of it there was.

Gaute: Oh right, i see. Well ok. I've gotta turn this around and get more players/more revenue and streamline this whole thing. What we got in the pipeline?

R&D: Well at the moment we're trying to get some patches out to do with the client - crashing issues. We're also trying to get pathing sorted out and some of the professions need some major fixing too. Player base is mighty annoyed at a lot of stuff.

Gaute: Hmm... well im afraid that your people are gonna have to work harder as im cutting personnel to save some more money.

R&D: Bugger. :/ ( well there goes our commitment/incentive )

Gaute: Look you guys, we need more players and more money. so.... what do you suggest?

Sales: Well we could always do an expansion pack.

Gaute: I thought we already did those every other month as part of 'delivery' for the monthly fee?

Sales: Sure, but we could add lots of patches together and add some more content etc.

Gaute: Hmm... dunno

R&D: But some of our patches, including the graphics download, have been almost 850Mb. How could we sell a patch to them that we would normally have them doanload for free?

Sales: We'll just call it something different and say it's too big..... also that to get it to the player, we need to box it up and put it on shelves for them to buy. plus we get great advertising for something that our player base will have to buy to basically keep playing the game!

R&D: Um... why dont we just sell at cost to existing players and mail it out to them? We have their details on file... easy done

Sales: don't be stoopid
Gaute: don't be stoopid, we need a cash boost, simple. anyway, i want you R&D people to come up with something new and i want it developed ASAP!

R&D: What about our current commitments/promises?

Gaute: Tell them... hmm.......... i know, get them to create a player council where they put all of their wants into a BIIIIIG list to be presented to us. that will buy us easily 3-4 months!

6 MONTHS LATER>>>>>>>>>>>

R&D: Erm, Gaute?

Gaute: Oh hi, hey yeah, love the expansion pack theme! Excellent ideas! We've promoted this so much it's gonna be a massive hit!! I can almost see the reviews now!

R&D: Er, well, we kind have a few problems. Basically our original coding would need to be completely re-worked and our graphics engine should ideally be re-designed altogether. We also have a serious problem with 'What players term as lag' in some major areas of concern and it causes players some serious frustration or client crashes. :/

Gaute: what???? are you crazy?????? we don't have the time to re-work! Oh you mean LAG? why have we got LAG?

R&D: Well, opur graphics basically re-draws the same shapes/colours everything 6 million times a second and some of the players Pentium9 12.6Ghz 5Gb RAM machines are having serious problems. Unfortunately ours is the only game that seems to emphasise this as a problem.

Gaute: Bugger! well, can't we blame them instead. Tell them they need to turn down their graphics etc and also that we have to implement some restrictions?

R&D: erm, sure you could, but that really won't help. In fact, that might just annoy people.

Gaute: Nonsense you fool! I'll word it just so that it sounds really sincere. then they will be happy and i will even suggest they discuss it so that we get their feedback Jeez im great arent i

Sales: Erm, Gaute?

Gaute: Yeah?

Sales: Well, it wouldn't really be a 'discussion' would it? I mean, well, we would just watch what happens right?

Gaute: Of course not you fool.... i'm more than happy to leave the players chatting about it for as long as they like. Either way, we're gonna HAVE to implement this or we won't get our $$ back dammit!

[Engelen]

Uhm...please explain to me...what will this mean for us "Non PVP" players? I enjoy hunting outdoors, running around and looking at monsters, and generally having a good time in the freedom. How will this affect gameplay?

Those zones where there are towers, there are a certain limit of how many players in that zone? Or..?

If there is a war going on in...lets say Mutant Domain, and I want to go there to look at the mobs..or do a mission, will I be denied access to the whole zone?

/me Scratches his head*

[________________Luke]

not sure how it affects non pvp'ers like you or i...... However, you should still be able to go into a zone, not into the area where the pvp battle is going on ie, within 80 metres of that battle. so you can still zone in there, you just wont be able to go closer than 80 metres to the battle area.

[Engelen]

Ahh okay! Sounds okay to me, I was not too interested in spending all my time building towers anyway..=)
But then it is a good idea to make sure the battles will be played out with even odds.

[Gizzed]

I've been thinking about this and there is a way it could work.

Let's assume the limit is 60 people in any 40m area.

20 Clanners
20 Neutrals
20 Omni

Defenders get towers to help them in the fight, attackers get... Mercinaries! Neutrals!

The fight is between Clanners and Omni, the attacking force has time to plan and know they can not go in 20 to 20 with the towers helping the defenders, so they hire 10-20 neutrals as cannon fodder! An idea? The defenders might have a contract with certain neutral orgs that if they get attacked that org will come to their defense, if the battle is won with their help, they get a handsome sum! An idea?

If the attackers win they get that land, and therefore pay their mercinary companions a handsome sum! A win win situation! Omni and Clanners get the land, neutrals get the money!

The Free Falcon

[Animosity]

I have played on a machine that really is below the requirements of the game. I then I bought my uber machine. There is quite a difference, BUT.

I dont think FC should be catering to both. I would prefer to be able to shut off more and more features to make my low end machine work. I would actually like to be able to drop to wireframe on my low end machine (anyone remember mech warrior 2). In all honesty, if you've seen one atrox in a thong, you dont want to see any more of them.

If I had my way, you would create a second stream of clients for those of us who have the machines to handle it, let me overload my options.

I would prefer the problem to be solved by specializing the clients rather than limiting gameplay.

I would prefer there to be an ability to have massive seiges, and hundreds of people attacking at once. If anyone has seen starship troopers, you will remember the attack on the colony, I would love to play that out again in Tir, or Harries.

Frankly FC has gotten scared of large scale events because of this problem, and it is what we really want to do.

Ani.

[Animosity]

Originally posted by SM
Firstly I _am_ a software developer both win32 DX and Linux/BSD.

Only 5 years or so as a professional but I have a fair idea of how things work.

I can understand the bandwidth choke from system mem to card mem, but the solution these is simple, reduce texture rez when the play area fills up (mipmap then disgard parent).

Too many different meshes? simplify! Design some generic ones (the existing one should be generic inside breed anyway) and use the textures to differentiate (the UVW mapping in this game is so simple that this would be easy to do)

Thrashing from disk to mem when you zone? then don't zone like that, zone like this:

1. Client issues move order
2. Server initiates zone, removes player from playfield
3. Server issues rez manager command to client: load new playfield + all meshes in area (switching to more generic meshes/lower rez textures if overcrowded)
4. Client thrashes (Screen still blank, or maybe playing a pretty and simple animation)
5. Server prepair to add player to new playfield but doesn't do it yet.
6. Client issues "all done" order
7. Server adds player to playfield

True, It'll take longer to zone, but what would you prefer? the proposed "zone full" message? The server could even wait for a team to be ready before zoning them, if turning up together was considered important (like the Q3A level loading waits for all players to be up and ready)

One thing about the rez manager, I understand that IDE blocking can really screw up framerate when loading stuff, but why-oh-why doesn't AO cache properly? I have 1GB of ram, and _still_ the HDD thrashes after zoning.

People were asking how Luclin does in in EQ? well the caching works, and you need lots of RAM.

I still this that a proper non-blocking trickle loader is the only proper way to do this (the console can do this, but they don't have to contend with IDE)

Rather than limiting the battles please implement generic fallback meshes/textures (use vertex shading to get the general colour rather than textures, most people are wearing black/white/red/blue anyway) then trickle the textures from mem to GFX card as you get the oppertunity, and cache everything you can in mem before you log.

Please if anyone in FC thinks I haven't thought this through I'd love to hear their thoughts. I don't want to argue or be akward, I'd just love to hear their side.

This guy is cool. I am a tester, not a designer. I tend to couch what I hear in non technical terms, because it makes the designers think a bit.

Here is what my response is, Get everyone on the screen in basic gear. Then fill in the details as your processor can handle. That would be my preference. So, frame rate woudl stay high, my low end pc would see everyone in basically wireframe, which would improve if i sat still. My uber machine would not be affected much, except for those rare events where there are many people on the screen at the same time. I would prefer to lose the tactical advantage of knowing what someone's armour is rather than being shut out of their playfield.

Ani

[_____________Crymcyn]

As someone who writes internet apps for a living - as obvioiusly many of the player base also are - I have a little bit of an idea what I'm talking about here.

First - lets forget about, for a moment, placing blame as to where the actual problem resides.

We start of with a problem, incorrectly and generically referred to as LAG, so that's how I'll reference it here.

Now, as the provider of an application, it's FC's job to solve said problem. As the client, we each expect FC to fulfill this obligation as we are paying them to do so.

In most application providing business models, making excuses for why the client can't have what s/he was promised and providing less would result in loss of said client and likely millions of dollars in revenue. Unfortunately FC is in the business of farming massive numbers of clients with a far lower dollar value attached to each. So instead of being forced to follow the ethic that the customer is always right and correct the problem they feel comfortable is saying "Sorry the app doesn't perform to spec. If you don't like it, tough, we'll just get 2 other clients to replace you if you leave." The problem is, as seen through repeated examples in the past, eventually your poor track record in dealing with what you assumed to be an unlimited supply of clients will become common knowledge and your ability to replace the clients you chase away will vanish leading to the ultimate demise of both AO and likely FunCom.

Gaute, take some responsibility for the mess that AO is in. Quit making excuses and trying to lay off the blame on client machines. Poor application performance is almost ALWAYS a result of poor programming. It's either that or play the numbers game you're currently playing and lose. MMORPGs are rapidly increasing in number and as the choices increase your failure to follow anything resembling acceptable business ethics will come back to haunt you.

[___________Redrumsam]

im sick of everyone trying to be Politically Correct, so here is my rant:

Is the game is supposed to be a virtual reality? - no reality i know goes to such lengths to equalize conditions for all involved. In reality, life is NOT FAIR. In many many many ways ( i cant count all the ways on both hands) the game is already NOT FAIR, why start asserting fairness now?

why not balance the maximum amount of money, the strongest weapons, the best armour? the game is unfair in so many ways - to say that crowd limiting has anything to do with fairness is an insult to our intelligence.

if 100 people attack 10 people - LET THEM. how else are you going to provide a simulated environement which pushes weaker factions to search out strong members, or ally with stronger orgs?

it is unrealistic, fake, phony, a farce, to implement any system that arbitrarily balances battles, if anything, you should:

1) have a maximum number of players in an area
2) assess the RATIO of players at start of battle
3) base crowd limiting on the starting RATIO of battle

why was this not an option?

it is not fair, and it should not be. What is the #(*$(#*$ point of fighting for something if every battle is going to be fair?

my friend quit the game, and gave me his account becaus the game was too "weak"

if you make the land ownership battles "fair" im going to quit too, and NOT give the account to anyone.

[_______________Omeki]

You are running for a zone, because there is a not so friendly mob or two or three chasing you . . . and you reach that zone . . . but cannot zone because of the battle raging on the other side - that you are not stuck there getting killed while choosing between alternate places to go or reading a sorry - denied message.

[___________TreeTrunk]

Some alternate strategies to mitigate the extent to which a "fair" system would make too many decisions against player will are:

a) Organizations RANK their tower defenders within their organizations. Defenders are allowed into zones and kick out other players based on defender RANK assigned by organization, not based on level.

b) Organizations RANK their attackers. Rank is used to order of access to above ground tower-attack.

This strategy allows organizations to keep a doctor who is low level, for example, and ensure he is the last to be kicked. Ofcrouse, implementing Ranking would take some time (but muight add interesting politics to orgs )

c) Kicked or ported players who are in the battle are sent into the notum mines beneath the tower. They must fight against their opponents down there. Players are sent in queue style, and if suddenly room is available on surface, they port back up again. This would allow 'lower rank' players to participate in PvP and serve reserve positions.

[_____________Chubula]

Ok... I understand wanting to let everyone in on the upcoming battles. But to put it simply, You can't. Take UT2003. That game looks so pretty on 640x480 low setting, and my bros computer /w a GeForce 4 MX 420 chunks with more than 3 people on the screen in a LAN game with a 1.2 ghz cpu. You can't cator to everybody without pissing off the people who actualy bit the bullet to spend money to keep up with the times. I'm getting a new cpu in the mail today. I bought it in hopes of improving alot of things including this game. There goes that idea. How about these suggestions...

- Turn down your player veiw distance. Hello. Thats what the option is there for.

- Add an option o show just a certain amount of players at one time.

- Add an option to make everyone of one race appear the same(use the same mesh and clothed the same)

- Continue to optimize the engine and leave the thing alone.

Anyone have any others? Anyone starting to feel like the cash they spent on their computers was a waste?

[________The Engineer]

Everyone seems to forget that on the internet none of us seem to have a direct line to the AO servers. So real lag versus frame rates are a huge factor. To be sure become a tester! Sign up and play the test server and walk through trade when no one else is in zone even......Still you hesitate and freeze. I agree about the engineer's we seem to become less and less group value. Our Sym Nano's and warps where a start. How about making Engineers very very important in hacking or bringing the def down of land based towers? This way we would be saught after........................ Just a thought :-) Pathing of Bots still a problem too. getting stuck to right of door

[Solstyce]

I'm sure it has been mentioned before on this thread because of the many posts here. Forgive me if it has, but I'll state it, if it hasn't.

LOD and Swapping: I think a client that monitors swapping and predetermines whether there is going to be an excessive amount to load, or only a little, would help a good deal. If there's going to be a huge number of textures and onjects loaded at once, and the loading queue gets out of hand, use default textures on people until everything can be loaded in an orderly fashion. This can also be used to enable a working LOD (level of Detail) system. One that drops global (not just people) poly-counts, global texture sizes and drops to 16-bit low rez textures if needed. As framerates can handle it, the higher the settings return to what they should be...the highest being what you have set on the client's options. (So it doesn't ever go higher than what you set.) The last LOD sysem didn't seem to do much of anything. Some textures would distort because of slightly lower poly-counts, and I believe this is because it also distorted UV coords when those polys were removed. Also, as with most LOD systems, the farther an object is from you, the lower the texture size, and the lower the poly-count should be, and it should gradually ramp up as you get closer. You should, of course, be able to turn this LOD/Swapping system on or off from the client.

Another thing to mention... on a graphics standpoint, when you lower the view distance, the clipping plane ends up being closer than the fog-line. If the fog (or desert haze if you want to be creative) were closer to me than the clipping place (so I couldn't see the polys or objects being drawn in as I move) I wouldn't mind dropping my view distance to raise framerate and take some load off my video card.

As for there not being server lag, how about explaining how going link dead is related to the graphics handling capabilities of my computer? I can log into Tir, get a decent framerate (decent being about 10 fps...enough to haul arse my outta there) and even hang around there without it being too painful. But eventually, I'll LD. Sounds like a network issue to me. This all comes down to crowding, I know, but shouldn't a faster server-side network be able to handle sending and recieving positioning data and whatnot faster...which would cause less client drops and timeouts? Graphics are my forte, networks are not, forgive my ignorance and my allowing common sense to take over.

Just some ideas off the top'o'my head.

Sol

[Gridfan]

Btw what he ment by LD/lag due to your own systems is this:

Lets say your gfx card got 32MB gfx ram on it.
If you sit in tir and well alls fine..
Then more and more people pop up.
with different armor, weapons and stuff, some in veichles and yalms etc.
Then lets say your 32mb gfx card is filled up completly.
Now when the next person shows up the oldest textures etc in your gfx card is flushed out into system memory or flushed completly to make room for the new ones.
Now a texture is not in the card that should be there so the system
immidiatly reloads it and hence causing the oldest texture to be flushed.

You you end up with a cycle that flush the oldest, reload the flushed ones from system mem or harddrive.
And ends up circling thru all the textures.
Now add one more texture
And the system need to flush out 2 textures to add 2 more into gfx mem.

Memory/diskswapping mess up the systems bandwidth,
the gfx card and engine cant keeep up with the changes needed.
At some point the server notice YOU (your machine)
no longer respond fast enough, often enough.
(remember gfx and peoples movements and youor movement is synced to some extent) so if the gfx system lags behind (lag as in i.e gets slow, um clientlag rather than networklag or serverlag or internettlag)

Then timeouts will pop up. At some point the server consider you unresponsive enough and boots ya cause you are no longer "there".

People isnt the only things building textures are loaded as well etc.
There is still the odd unexploained sudden drops.
But I blame that on either the internet itself, or so far not found bugs in the client/server syncing code.

Gaute so must have hated to say that. If he were a politician before an election I guess you'd say he commited political suicide. (is that the proper term?)

Imo there is only 2 ways to no 3 ways to fix this "issue".

1. Gaute and FC's solution.
2. Or a total rewamp of gfx engine and some client and server code. (I think Expansion (Shadowlands) have some of this)
3. Dumb down the graphics so much that um AO quality wise gets the same as all the 1st gen graphical mmorpgs...

I think a combo of 1 and 2 is best and is what FC is actually doing.
As hardware gets better, as their servers and the code and client engine etc gets better. the higher and higher they can increase the Crowd control roof.

Also, guys. do this please.


DRAW A 360 degree circle.
Put a spot in the center of it.

That spot is YOU. the circle is tha 80 click radius around you.
if within your 80click radius there is
more than the allowed limit for your clients load a you enter the area
you get moved outside.

That the personal Crowd Limit.
Then there is additional higher limits set on zonez within a pf and pf's themselfs.

Point is the personal circle crowd limiot will try and mak things run smooth enough to avoid issues like playfields going down due to too many people.

We arent talking 25 or 50 peeps in a pf here.
You can have up to 200-300 people in a playfield/zone and the servers should in theory handle it.

What FC need to fix tough is the 80 click thing.
You cant "see people" 80 clicks away in visual nor on radar really.
But the server still tells you they are out there to some extent.
Tweaking this more could allow more than 200-300 peeps in a pf.
Sorry I dont have any code examples on this tough...

Point is in a huuge PvP guild war the Crowd Control all the way up to the zone/pf/server limit,
would ensure that nobody crash or go LD.
In fact it would "move" people slightly to ensure they are out of sight.

So effectivly you would end up with groups of people fighting 20-50
spread around the towers or town or zone etc.


It may not be the ideal solution but it sure beats going LD
or getting framerates of 1FPS causing you to stare at reclaim 2secs before your even able to target peeps.

And what is morre amusing is my netconnection isnt the fastest out there, nor is my pc barely mid-level. Yet I get less lag issues
than people with awsome connections and top of line systems.
And I'm in Norway, which means the game servers is on the other side of the atlantic from me. (in the US)

If ANYONE got some nifty code or mathematical ideas, PLEASE TELL FUNCOM. I'm sure theyd love to avoid doing this.

But I agree with Gaute, right now this seems the only current solution. Next to a complete engine rewrite.
Which would fedinetly postpone Shadowlands, Booster etc until 2004.

Dont worry tough, FC is tweaking the engine. But it steps and areas only. (hence the messups when they remove old no longer needed code and remove a bit too much old code (i.e quick fixes).

And who knows. FC may just be working on a AO2 engine that may pop up sometime next year. But I doubt theyd hint to anything lie that until the last minute

[SilentMobius]

Originally posted by MORB


I don't think that the various meshes they have currently for the players are the reason for the disk access thrashing - they probably don't weight that much in total. I think that the textures are much more of a problem.

I agree I was planning to add "but I don't think meshes are reall y this issue here" but I didn't, I wanted to make suggestions across the board.

My point is that must of the post-whompa thrashing could and should be avoided and that FC really need a decent texture reduction strategy.

[______________Insitu]

For the most part I would agree with this limiting of clients especially in PvP Combat areas, however I would think this would give an unfair advantage to some of the larger orgs taking each other on in the Notum Wars. Say you have 200 users that log on in the notum wars for PvP such as the Dojo on RK2 fighting against another Org of simlar size. This would not allow the full forces of the Org defending to be brought into play. I think this would limit game play and PvP some in that instance though I do realize that this won't be a big issue. I do know that this could also cause some annoyances however I will admit this is a mutch better alternative than crashing and having to re-log.

Also I would think that if running the basis for Client limiting at mid level computers Say PIII 750+ with a 64MB Video Card At the Mid levels of Graphics would probably cover the vast majority of people even though it would trigger the Client Limiting Sooner.

[Avarael]

I'd rather have it scaling back the client's looks - even quite drastically - than forcing players out of the fight.

[________________Luke]

Some really valuable stuff here, really hope FC pays attention to all of this.

As Crymcyn mentioned earlier, FC really should take some responsibility for this.

Chubula... i beta tested Neocron along with many others. We were engaged in many a battle outside the city gates with hordes of NPC's and there was easily 100+ players battling it out too. There was granted some mild 'lag', but that lasted no more than 3 secs... AO 'lag' can last a lot longer and with far fewer players in the zone. I don't know how the developers of Neocron got around it, but they seemed to have achieved it.

Forgot to mention something in my earlier post (which i thought was amusing but no-one else did.. lol):

MMORPG = Massively Multiplayer Online Role-Playing Game
MMOPG = AO