Slashdot Mirror
New Technologies Attack the One-World Problem
Hugh Pickens writes "An MIT Technology Review article has new details on the challenges of a 'one world design' in Massively Multiplayer Online Games. Most games shard their servers, putting up artificial barriers between friends and family members. Technologies are now being developed to keep lots of players within a single world, some of them based off of the unique PvP-heavy title EVE Online. The best part - the technologies don't just apply to gaming. 'NASDAQ, for example, can be thought of as a very large MMO, supporting very large numbers of 'players' performing billions of transactions daily in a graphically intense environment, all within a single shard. Technologies that solve this problem effectively, says George Dolbier, technical lead for games and interactive entertainment at IBM, will have applications in any industry that requires spotting and reacting to trends, or "anything where behavior is dynamic and you need to move resources around rapidly."'"
", for example, can be thought of as a very large MMO, supporting very large numbers of 'players' performing billions of transactions daily in a graphically intense environment, all within a single shard. "
Jeezs, you mean there are large transaction systems out there? Thank god MMO's brought the technology to the world! Gah. MMO's do nothing technically new regarding transaction.
Also, what is the 'Graphically intense' interface the NASDAQ has? Compared to MMO's it's nothing.
The Kruger Dunning explains most post on
Don't you love it how gamers try to make themselves relevant to the rest of the world? I love gamers... they're so... important... to... um... well... Just keep playing, guys!!
--Ray
http://www.beanleafpress.com
http://blog.heavensdomain.net
WoW isn't the only MMO out there. Not all MMOs have numerous servers each supporting one single "world" with no way to move to the others. Guild Wars lets you play with anyone, anywhere, anytime in the world. Servers are specific to entire continents (with Asia fragmented into Japan, Taiwan, China, Korea), but people can group across those servers on a whim and play together.
Since GW is largely instance-based there might be fewer performance issues to deal with than with WoW in this regard, but my point is I just don't see how one would dedicate a "news" story to this topic when a popular title like GW has already tacked the issue.
Or did I totally misunderstand the topic altogether? It could be ^_^
I like basketball!!1!
And you thought the grind in WoW was boring!
Modern copyright is theft of culture from everyone and it retards the progress of the useful arts and sciences.
Personally, I locked onto to the "graphically intense" part of that comment.
Doesn't seem to me that thousands of stock-trend charts and graphs really count, unless you're making a terrible pun.
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
A beo~ oh nevermind... wrong direction!
The dangers of knowledge trigger emotional distress in human beings.
I don't see how that is even feasible. I don't play WoW because it's a horrible game. However, I have tried it. Now, there's what, 8 million subscribers? Let's say 1/4 of them were logged in at once.
First, can a server even handle 2 million simultaneous logins? I bet they could do something, but it would cost a LOT more than splitting them up into managable chunks.
Second, the game world would have to be enormous in order to give people enough room to move around and do their own thing. Just imagine hunting a single boss, 300 people at the same time trying to kill one monster... it'd make me quit.
Now, the game dynamic would have to be changed and yes, that is possible. But at this point in time, I don't think it's physically feasible to run a virtual world the size something like this would have to be in order to fit everyone in at the same time. Zones lag enough as it is. And updating every user's stats? Unless we all have 1000mbit internet connections, I don't think we even have enough bandwidth. And the travel time in-game for that kind of world? You better give everyone instant teleport to any destination or nobody's going to want to move around...
-SaNo
Obligatory to mention it, but this is of course what Second Life does, and one of the reasons why it's interesting. With SL all assets are stored online, not on your local PC (preloaded from CD or whatever) and everyone is in the same world. Anyone who witnessed the growing pains of SL over the first part of the year when concurrency went from under 10,000 to 30,000 plus will be more than a little aware that what they had didn't scale, although they do seem to have a handle on it now and conccurency of 50,000 is just about bearable.
The article seems to miss that EVE being space based allows a huge environment to be created easily, a few random number generators and a bit of tweaking and you have a whole universe of stars and asteroids and it's easily extendable after that.
The landscape in the likes of WoW is a lot more design intensive, you have features and locations with NPCs and dungeons and so on put in place. To double the population on the server you would need to either double the design/quest writing hours, add in a bunch of fractally generated landscape that would be relatively boring and largely pointless or go with the EQ2 route of opening up instances of zones which always seemed artifcial to me ("Hey are you in Common Lands 1,2,3,4 or 5?").
The author needs to not compare apples to oranges. WoW and EVE might both be round (MMO), but they aren't the same, 200k isn't even in the same ballpark with 9 million. The author needs to do some more research before they try and compare the two. How many WoW players are there per shard, for instance? I'd not be surprised if there were more than 200k players on some of the larger WoW servers.
and that bit about 40-on-40 battle size for WoW is totally bogus. That might be the largest battleground instance, but I guarantee that there have been larger PvP battles in some of the major cities. I've personally seen multi-raid group per side battles in some places. Not to mention the number of people who were on hand for the opening of the Gates of AQ and the Dark Portal.
In EQ, when they would try to run a special event, a large number of players would try to join and crash it.
If you have 100,000 users and some kind of non-instanced shared event (say you tried to have a virtual stadium where spectators in the stands saw an event below) the capacity required is beyond the capacity of current hardware.
She was like chocolate when she drank... semi-sweet at first and then increasingly bitter.
Hell, which ones? Poor performance? Quirky configuration? Conflicts with IBM Director? Event Rules errors? Corrupted objects? Magic failing package deploys?
:)
Oh, wait, you didn't mean JD Edwards. Never mind.
You are not the customer.
1. Many people don't want to play with their family. My son likes to grief, hold grudges, and seek vengeance. I like to Care Bear, focus on skills, and meander about, for example.
...
2. The concept of being forced to play with people who think PvP is great just bores the tears out of me. Just as my RPG style probably does the same thing to PvPers. So, having a fractured community is kind of nice, and it's also good in that, should I totally mess up (as I did when I founded a Squirrelly Wrath guild on one server and "invited" someone that then insulted everyone he could find in multiple lands, making the guild a piece of shyte, well, I just created characters on a different server (sorry, Sisters of Elune, it wasn't me).
3. Economics doesn't need to represent the world. Why should I want it to represent 250 million Chinese Gold Farmers? I'd rather it represent people from Ecotopia quite frankly
-- Tigger warning: This post may contain tiggers! --
We have no choice! Our mod points can't repel flamebait of that magnitude!
Actually, the devil is in the details: when you said that GW is instance based instead, then "one world" doesn't even mean the same thing as in WoW. GW, just like Diablo 2, is the exact opposite. Everything is instanced. It doesn't even _have_ a real shard or world, in the WoW sense.
To illustrate what I mean: in WoW I can for example take a treck from Anvilmar to Ironforge to Stormwind to Goldshire (see for example the funny video with the 40 level 1 gnomes raid on Hogger for a group doing just that) and meet a few thousand players playing in the same world. _That_ is WoW's "world". And going "one world" would mean essentially all 9 million players literally running around in the same world. Not in instanced versions of it.
Does any part of GW have that? No, I don't think so, everyone who isn't in your instanced game, just doesn't exist in your "world".
Technically speaking, GW has lots of smaller shards, not one big world for everyone. It just invented a way to spawn new shards as needed, that's all.
This isn't to say that GW is bad or that WoW is better. I can see the point in instanced content. But let's not go redefining terms for "my game can beat yours" willy-waving. Just having basically a lobby from which you can start an instanced game or join one, does not make a "one world game" in the sense discussed here. It's just not the same kind of "one world", so making the claim that it did it before WoW is meaningless. It's like saying that cats invented leatherworking because your cat has white "shoes".
Wake me up when you can have 36,000 people in your GW game running around independently and actually randomly seeing or meeting each other. That's currently the average population per server in WoW. _That_ is the point at which you can claim with a straight face that GW even does the same thing in that aspect. Do it all in all world? Well, wake me up when it supports 9 million players in the same game, running around and whacking NPCs independently.
And here's another thought, and what the guys in the summary missed:
Chances are you don't even have enough geography for that. If you parked one player per square metre, you'd need a 3 km by 3 km world just to have the players stand there. If you want them to actually have some space to move around and hunt without stepping on each other's toes, you end up needing a world as big as TES Arena. Except at that point you also need a hell of a lot more quests (people won't be happy if they have to run an hour just to get to the next quest giver), and other problems start to creep in too. _That_ is why noone, WoW and GW included, ever tried doing that.
So people coming up with ideas like "hey, look, it's technically possible" have just missed the point.
A polar bear is a cartesian bear after a coordinate transform.
If all of WoW was a single shard, what would the result be? With the existing world size, the population density would be insane. Somewhere around 100x a normal server or so. I've seen AHs with nearly 100 people in them, so how about the same with 10,000 fighting to get to an auctioneer? Or rather than fighting with 2-3 others camping the rare with the drop I need, it'll be 200-300 camping the 24 hour respawn. Or expand the world to a size that gets difficult to meet up with people unless you always get on and off where they are. What the WoW solution would be would to let people move characters around on servers easily. That way your choice of servers wouldn't be a problem when a friend mentions that he's on a different realm. But that would be hard to do now because of character names. And what about items that are hard to get and you not only camp the spawn, but pick certain servers to camp.
I'm not saying that the problem can't be fixed, but it is something that would be hard, if not impossible, to fix in an existing large MMORPG without causing lots of trouble for the users.
Learn to love Alaska
I don't know what "shard" means in this context. I'm going to assume it's a dorky way to say "server", but you're missing the point. The author is saying, "NASDAQ can do it. Why can't MMOs?", not the other way around. MMO's are really nothing new technology-wise. The difference is that no MMO comes even close to be as well built, technically, as say, the NASDAQ and NYSE. Those systems are incredibly high volume, and relatively stable. (Hint: They're not using MySQL).
MMO's could certainly make certainly large worlds. The systems themselves are not all that high volume (hence the comparison to stock markets). It's just that none of them are built well enough to handle it.
I don't respond to AC's.
"... will have applications in any industry that requires spotting and reacting to trends", or "anything where behavior is dynamic and you need to move resources around rapidly."
... a battlefield or even a major military campaign. Eventually war is going to be a matter of software "generals" maneuvering resources and personnel around in order to achieve maximum effect. Something tells me the military may already be far ahead of what the massively-multiplayer folks are doing. Or maybe not: when you think about it, a closed universe interacting with millions of actual human beings is a great place to experiment with this sort of thing, and hey, you even get the players to pay for it.
Like, say
The higher the technology, the sharper that two-edged sword.
What do you call it when a shard server craps out? It sharded.
Previewing comments are for sissies!
You're trying to get me to say the name of that other Sandra Bullock computer movie.
Well, it ain't working, sparky.
Politics is the art of looking for trouble, finding it everywhere, diagnosing it incorrectly and applying the wrong fix.
...most worlds are simply to small. Imagine all WoW servers merged. You couldn't even enter a city. The worlds would have to be designed much larger. Could be truely awesome but I guess they just end up as big bleak planes of boring nothingness. Switching servers should be made as easy as possible. Perhaps like FPS, your character is stored on a central server and you simply choose a server, connect and start playing with your friends.
There are a few items of note:
First, this George Dolbier says that MMOs and massively distributed financial systems share the same problems, and that the financial systems have gone a long way to address them. He says MMOs should adopt solutions applied to the finance sector.
The second thing to note is that he talks about predicting and reallocating server computing resources. He's from IBM, who hawks services and products in this very area.
Everywhere I go I have met people that I like in real life and who also play.
:(
Are any of them ever on the same server? No, of course not
Of course, supporting the density of 9 million people on a single wow server would both require and allow massive shifts in gameplay. No more grinding thirty mobs at once - you've have to zerg rush a single murloc to get him down, or the world would be completely overrun in murlocs.
Is it me or does it seem that all the scandal and internal cheating going on at EVE probably put a dent in their customer base. Now they ramp up the PR machine to protect their cash cow from imploding on itself. Other than a paragraph or two about WOW and the NASDAQ they go into too much detail about EVE.
I mean really some crappy "independent" report and then interviews with the new EVE economist who never once addresses the cheating problems and the fact that EVE is not a closed system as it interacts with the real and parallel world.
For even mentioning that movie, your Geek Factor suffers a -10 hit.
You're RPG equivelant is now "Tunnels and Trolls", and your Star Wars equivelant is set to "Jar Jar".
Are you knocking Tunnels and Trolls? T&T was a great "beer and pretzels" game of yesteryear. The rules were dirt simple in comparison to the fantasy RPGs of its day, and it didn't take itself seriously. What other game has spells like "Take That, You Fiend," giant squirrels as dungeon monsters, and modules with names like "Rat on a Stick" where you could either kill the monsters or set up a fast-food franchise for them. The T&T games I played back in the day were a heck of a lot more fun than the old D&D games.
(Personally, given its age, I consider even knowing about T&T to be bonus geek points, but putting it in the same category as Jar Jar shows failure.)
If it's for-profit but free, you're not the customer -- you're the product (e.g., the Slashdot Beta's "audience").
Just as a point of comparison, here's Wikipedia's architecture:
http://meta.wikimedia.org/wiki/Server_layout_diagrams
P.S. And for fun, here's the most ambitious MMO ever:
http://www.metagovernment.org/
Go one of 2 routes:
1> Instancing of all areas, ala Guild Wars style.
2> Select your character, THEN choose your server, ala FPS game style.
You solve two issues, it's easier to scale up when your game gets popular(no character migrations off of heavy servers to deal with) and it's easy to scale down when the game starts to die off(no annoying server merges).
But on the downside you can't charge people $25 to move their character just so they can play with their friends.
I see this kind of thing pop up on a regular basis, and it always missed the point. This isn't a technology problem! Speaking as someone who's actually worked on multiple massively-multiplayer games, once you've got the server tech to support 10,000 people on a server cluster, there aren't a lot of technical obstacles to scaling that up to 8,000,000 people. Every part of the server cluster can be scaled out more or less infinitely if you apply the correct (and already well-known) engineering solutions. And money, of course.
It's actually a content production problem. If you're going to put 8,000,000 people into a single virtual world, you have to have places for them all to go and not be horribly overcrowded. Ideally you want all those places to be unique, interesting, and compelling to play. The fundamental problem is that we simply don't know how to create that much content. Hand-crafted content is far too slow and expensive to produce at that scale, and auto-generated content is repetitive and boring. Eve Online manages to hold 200,000 players in a single server cluster environment only because all of its environments are the same random-generated solar systems. Once you've seen four or five systems in Eve, you've seen them all. Fortunately Eve's strength doesn't rely on the environments, it relies on PVP action. WoW couldn't get away with that.
The Nasdaq comment caught my eye. I believe that Nasdaq, at least in part, sits on Unisys built mainframes. Has any MMO company talked to Unisys, IBM or any other mainframe manufactures? (are there any others left?)
It seems to me if you wanted "1 world" it might be a good idea to start by having "1 machine".
Of course this presumes the reason why things are on many servers is mainly technical restrictions as opposed to content creation reasons. Maybe Eve works as 1 world is because you only need a map maker; it is the players that create the content, meaning drama. A mainly PvE game would require lots and lots very carefully crafted content to make sure everyone has enough room to spread out but still group up. It is easy to create enough content for 25,000 people is easier then for 200,000; unless the other people are the content.
You DO NOT need to update all clients with all client data. It is NOT a n**2 problem. You only need to update clients that are interacting. The problem with this approach is that is is DIFFICULT. You have to still update positions on the non-interacting clients at some intervals so they are not out of place completely. And the server has to keep track of who interacts with what. Most developers at this point would just use the n**2 solution where you update all with all data and just limit where they can interact.
Also, you do not need to update non-interacting, non-visible characters at all.
EVE is a really easy solution because,
* it is NOT action game - it is a pure strategy (pilots don't dodge stuff, ship flies by itself)
* no need for low lag (see above). So, you can have 1000 clients in one battle.
EVE sometimes has huge lag, but is still playable.
EVE is much easier game to write large battles for than say FPS like Quake. And yes, WoW is like EVE.
In a way EVE is sharded. See each solar system can run on only one CPU. Once one system gets too loaded, CCP will move agents/quests to different systems to spread out the load. At no point will a weapon, ship, etc really go from one system to another. Whenevery they jump, their data is actually just picked and moved to a different CPU. So in a way EVE Online is made up of thousands of shards with the ability to swap data back and forth. Now once we have 10,000 players in one world/shard/star system, that's something to get excited about.
Man is the lowest-cost, 150-pound, nonlinear, all-purpose computer system which can be mass-produced by unskilled labor.
Look at the real world? IN the US we live in a zip code, a city, in a county, in a state, in a country, on a continent. These are physical boundaries. Servers and drives are physical locations. Map the various areas of the fake world to specific servers and drives. What happens when real population grows? We add a zip code. Add a drive or a server.
Rich.
libguestfs - tools for accessing and modifying virtual machine disk images
But will Linux blend?
Even if it did, there's still other possibilities -- compression, for example. (A lot of those movements are going to look the same.)
Let's look at some real numbers, hmm?
And that's for Counter-Strike Source, which, if you don't know, is a very fast-paced FPS which absolutely needs to be as accurate as possible (so people can get headshots, etc). It probably uses several orders of magnitude more bandwidth per player than your typical MMO.
Note, also, that this is server bandwidth needed, not client bandwidth. I'll bet, too, that some of that would go away if we could actually use multicast... *sigh*
Don't thank God, thank a doctor!
While I'm sure there are some interesting algorithms that can be used to manage hot spots and the like the shared world problem is only secondarily a tech problem.
The problem is a design resources and player density problem. The whole reason that players want to play on the same server is socialization but this runs the risk of everyone wanting to be in the same place. While crowds may make it self-controlling in some sense it will scuttle the feel of the small shop in the woods or tiny forest town if it is filled with 1000 adventurers. Worse it will just be frustrating because new shops and new services won't be built by people responding to the market and the difficulty of moving an avatar around compared to walking will be further frustrating.
Ultimately the real problem is that a good virtual world will have a certain optimal number of people per area. The price of running WoW if every server had to be replaced by an equivalent amount of NEW content would be stupendous. You can't just duplicate earlier content since that ruins the story and context. Worse it isn't even clear that with unlimited resources it is doable. After all the fifteenth lord of the underworld or whatever starts to get old. A world with too many ultimate locations gets silly.
If you liked this thought maybe you would find my blog nice too:
Magic: The Gathering Online is another example of a massive multiplayer game (300.000 players) where the transactions weren't very well thought out... and a large part of the game relies on the "trading" aspect of a trading card game.
The "Is the server up" tread on the forums runs to 143 pages since November 2006 and is very scary http://forums.gleemax.com/showthread.php?t=733609
There are three kinds of lies: lies, damned lies, and statistics.
there was a beta out in 2005 for a game called wish (http://www.colossalstudios.com/wish/index.php) it was one huge world with no instances. too bad it never got out of beta before they killed it.
"Now you want them to scale up that content production by two or three orders of magnitude just so that everyone can play in the same virtual world? It ain't going to happen with the current state of the art in content production."
Basically so. Todays tools are the artistic equivalent of working with "ones" and "zeros". Some are a little higher. i.e. primitives, particles, motion paths, premades. But it's still too close to low level. One should start as high as possible, and chip away with simple tools to get what one wants.*
*One starts with stone and envisions what one wants, then removes material. Not the other way around.
Oh I don't know. Modern day dashboards can be pretty graphical.
Name a group of people that purchase the latest and greatest hardware, then push it to the edge, then overclock it, and then push it to the edge again.
Gamers.
A HUGE chunk of the IT world is gamers. So while it isn't as productive raiding a dungeon or fighting in an arena as it is writing a new version of a php/mysql site, gaming pumps more into R&D than people think. It's not only data centers that drive new innovations.
Correct me if I am wrong, but I remember the opening to UO having the globe that was UO being shattered into many pieces that each had the same world in it, thus the invention of shards? Or was this a common practice in other MMOs before it? I guess MUDs that ran the same rule set could have been thought of as shards, but then different people were running them. Of course technically I understand the concept, I am just talking about the term 'shard'.