One of the things I wanted to bring up in the last post, but then completely forgot when it came time to write, was the mechanic in EverQuest II that allows the game to spawn additional versions of a given zone if the population of that zone gets to a pre-defined threshold.
I would link to one of my character histories on EQ2 Players to show that I leveled up in zones with names like “Antonica 3,” but you have to pay for access to that sort of data on EQ2 Players, so you’ll just have to take my word for it.
Anyway, given that multiple versions of a zone mechanic, theoretically you should be able to run all of the EQ2 servers as one consolidated server. Theoretically.
Of course, I know nothing of the actual mechanics and limitations of server usage, database operations, network traffic, and all the other fine details that generally gang up on theory and bring it back down to reality.
But let’s just sweep that aside for now. Let’s assume that SOE could and would consolidate all of its servers down to four. One standard PvE, one PvP, one Role Play PvE, and one for those Station Exchange people, call it the RMT server.
So now if you roll up a character on a standard PvE server, like most people do, you know that any of your friends who join the game or are already playing the game will be on the same server as you. If you meet fellow PvE players at, say, fan faire, you will know they are on your server almost immediately and you can find them later in game. And if you are a real jerk to people, your reputation will span the whole game, as it is played on your server.
Would all of that balance the zone based world of EQ2? Would having six, or eight, or a dozen versions of a popular zone running make knowing that, with a little work, you and your friends could always link up because you are all on the same server?
Is this preferable to the sharded existence? Would you give up a “seamless” world for that?
If your answer is “no” or “maybe” what issue or issues stand in the way?
The Ancient Gaming Noob 
In a game which includes both server and gamewide discoveries, as well as contested named and raid targets – imagine what happens when you shoehorn 100,000 – 200,000 players into one server.
I’ve already seen guilds game the zone spawning technology to give themselves additional raid targets after they’d cleared out the target. Just imagine for a moment what happens when you take the 2-5 high end raid guilds from each of the servers and put all 100+ of them into one giant uber server.
LikeLike
There are certainly game mechanics that would have to be worked on if it were a plan for EQ2.
But you will note that my actual question was more general that just EQ2, referring just to accepting the multiplicity of zones for a fantasy MMORPG. EQ2 gets the headline to give them their due for having the basic mechanic required already in place.
So, EQ2 specific issues aside, yes, no, or maybe?
LikeLike
Yes, I’d prefer one server per ruleset (you’d still need a separate servers for PVP, PVE, and so on.) Sort of like how it is in Guild Wars: the cities and outposts are shared areas, but hundreds of instances of them can spawn to handle all the people running around them, and people can switch from one to another at will. (In EQ2, you can switch to another instance of the same zone if you are grouped with someone in that instance.)
I think that would help a lot with population issues; you could never really have an “empty” server until the game itself was dying.
There are probably technical reasons why this isn’t done often today but I imagine in the future it will become the norm.
LikeLike
I like the idea of one giant server. One of the complaints a lot of people have in MMO (or at least when i play eq2) is that they can’t find a group. With everyone together there would be more people looking and forming groups, and getting a group should be easier.
It is also technology feasible as well. Though it most likely would call for a redesign of eq2 databases to get something like that to work.
LikeLike
I run a massive machine, so the “one giant server” idea is fine for me. However, that said, how many of your guildmates don’t run top end PC’s.
Obviously, if we’re talking only from a philosophical standpoint, I don’t think there’s ANY studio that’s against the idea of “one server per ruleset”. However, I’d argue strongly that it’s not the philosophy which ends up winning out once we start to see server queues, etc.
LikeLike
Certainly its feasible (Guild Wars has done it already), but there are a number of issues due to the shared server environment.
1) Names. Every single name will have to be unique. That won’t matter for some people, and can be partially avoided by giving players a last name at outset, or even the ability to use three words in a name, etc. However, a lot of people ‘reserve’ names or hop servers until they can find one to make their favorite character name. With unified servers, this just isn’t possible.
2) Auction Houses. Auction houses can’t be instance specific or they’d be useless, which means a colossal central database sending and receiving information from every instance of the Auction house zone. If the game has a low population this may be feasible, but if your game has a huge population it would be a nightmare. I suspect this is why Guild Wars doesn’t have an AH system.
3) Lost in the mirror worlds. Someone would need to build damned robust social networking tools for a highly popular unified game. If you took something like World of Warcraft and put it on a single instanced server structure you could easily end up with 157 instances of Orgrimmar running at any given time (this is assuming there’s a western and Chinese server, otherwise it would be even worse). Navigating that would be . . . most likely torturous. “Ooops, sorry guys, I went to Orgrimmar 117 instead of 118. Brt . . . oh, 118 is full, I can’t get in, umm, meet me in 130?”
There are more issues, but this post is long enough. In short, its possible, and for a low population game could be good, but there are a LOT of quality of life issues with a single server solution in most MMORPGs. EVE is a special case, and most people hate the fragmented feel of Guild Wars’ world.
LikeLike
@Kendricke: I’m not sure how your own gaming rig enters into it. I assume that is the “massive machine” to which you refer. There were certainly points of time where there were enough people running around Kunark to have the game spawn multiple versions of zones. Did some of your guildmates run into issues with that? Was the threshold for the number of people set too high? I am not suggesting that some company dump 100,000 players in the same zone.
I would say that it is more than a philosophical question, just pointed as a design idea for the next generation of MMOs, as opposed to trying to make it work within the context of EQ2. Of course one server per rule set should always be a winner if there are no consequences. Give me next gen graphics and intelligent NPC AI while you’re at it, if it doesn’t cost anything.
What I am wondering is, is it worth the price? Would people put up with many discreet zones, often many versions of the same zone, along with requirements like instancing of some group and probably most raid content, to get everybody on the same server.
Given how often I have heard “seamless world” put forth as a requirement for a “modern” MMO, and how much annoyance I have seen associated with zoning, I would expect that most people would say, “No.”
LikeLike
I’m in favour of a single server world – seems to work okay for GW and EVE Online – although as mentioned variously above, there would be many issues in taking an existing game and making the switch.
However, there are always solutions. Speaking as a non-programmer of course, but the massively multiple zone issue could be aided by allowing groups to custom name the zone they go into. Instead of struggling with “Hey, we’re in Antonica 76… no wait, 77, er…” you could have, ‘Antonica_Prez’ or suchlike? Of course, each zone would be pretty empty but that seems to work for GW. Maybe combine it with larger-access ‘public’ versions of Antonica zones, say?
The other issue of course, is one of public relations/marketing. As soon as an existing game declares “Hey, we’re merging servers!” instantly people think “Uh oh, Game X Online is dying”.
LikeLike
Wilhelm,
Two responses for the gaming rig question: Unless you literally plan on setting up the game to spawn new mirrors for EVERY zone (and Everquest II does not – you can’t get Qeynos Harbor 2, for example), then you’re setting yourself up for a situation where you CAN have thousands or tens of thousands of players in the same zone.
That said, that leads me to response 2, which is “if it can happen, assume it will happen”. If you allow players to move freely between zones, then you can and likely WILL have instances where players are stacking zones to be with friends/most populated instance/get away from the guys in instance 3 camping target X, etc.
Of course, all of this assumes zones, which obviously, as you’ve pointed out, isn’t so obvious a requirement after all.
LikeLike
I am also from the camp that enjoys Guild Wars style of instancing. The centralized meeting places in the various towns gives the game an overly “busy” feeling.
This makes for a game that looks to be well populated as well.
The ability to move from one instance to the next to meet friends makes the game much easier to adapt to.
I believe Age of Conan has this, but the people are too spread out. Move everyone to one server for PvE and PvP and there would be less of an issue finding people. This would give a feeling of a highly populated game, which it has been noted to have, but for this very reason, no one sees.
LikeLike
I guess I’m really torn on the issue. In a game such as Everquest II, I’m actually against giant servers, but that has more to do with the fact of the game’s actual ruleset.
In future games, I’d be much more open to “uber servers”, but I’d be hesitant to play in a Guild Wars or DDO style “everything’s instanced” MMO. If I wanted to play in a giant world server, I’d want it to likewise feel “giant” – not as if I’m playing on one big mirrored instance.
LikeLike
The first concern that comes to mind is how such things as avatars and other contested mobs would be timed and placed, as well as what opportunities there would be for farmers to, say, head to four different Sebilis zones and kill The Myconid Spore King and Kotiz fours time within the space of a couple hours (roughly a three hour respawn for both mobs). I guess what I am trying to work out is how farmers and high end guilds would be able to take advantage of numerous incarnations of a zone. That is, unless there’s just one mob to be had among them all, strangling the market for contested gear.
Speaking of the market, this would also be put in a state of upheaval, due the populations of many servers suddenly having one central market. I suppose this would die down quickly, and see no reason that it wouldn’t turn out fine.
Overall, I’d love a chance to group with people from other servers without the cost of transfer, but the developers would have to make some difficult decisions in terms of addressing supply and demand.
Numbered zones (which I believe are already due to overpopulation) do take away from the roleplaying experience, if only by a degree. Still a worthy suggestion.
LikeLike
I would answer yes, but with the caveat that the goal itself is not a single realm/shared/game world instance, but rather the ability for anyone to play together with anyone.
To accomplish that goal, one option is certainly a single world with many zone instances. Guild Wars is a good example.
But one could also have many realms/shards/game world instances like there is now – just do not tie a character to a specific instance of those.
The whole thing requires some rethinking in other areas also, like separating character name from identity (e.g. City of Heroes/Villains global names vs character names).
In the end it will be a matter if it is worth the effort, will players reward the game developers to “fix” this issue?
For any existing game I do not think any developer will think it is worth the effort and perhaps not if they will mainly reuse the capabilities of the existing infrastructure that they have already built for previous games.
In that regard I think that if we see anything in this area it will be from a game developer that has already done it (e.g. CCP, ArenaNet/NCSoft) or have been tossing around with the concepts (e.g. Cryptic) or of course someone who is developing entirely new MMO software.
LikeLike
One thing is contested mob mechanic (hello timezones), as mentioned above; other biggie is localization (names, quests, just think about it).
LikeLike
I think that this is a fantastic idea, having had lots of experience with EQ2. Specifically relating to that game, there are a lot of dead old world zones that few people play in (often except confused newbies) and therefore having lots of people on one server would help keep them popular. A better idea would be just to improve the ‘old world’ zones up to a par with ones from the current expansion, but hey (the fact they haven’t done so is my biggest pet hate with EQ2, as anyone who reads my blog will no doubt know).
Generally, though, grouping and social interaction are the biggest forces keeping people playing MMOs over offline games that feature better graphics and usually better combat. Any way to facilitate either of those two features would be a good idea.
LikeLike
I’m absolutely in favor of this. As has been said, the technology is there for every zone, and only requires “switching on” for each zone as necessary. As was mentioned, some of them are set deliberately not to instance (most city zones) but changing that is a simple matter, and not a re-design of the game.
As for technical limitations, I can see there being two main issues: 1) Load balancing the zones/servers (I don’t know how it’s being done now, so I don’t know how scalable that is), and 2) The Broker. The first is something that’s purely on SOE’s side, and we really have very little insight into how that’s done. The Broker is also there side, but we can all see how throwing 10,000s more sellers into a market could possibly cause slowdown of the interface at the least. Again, just how scalable their architecture here is (I suspect since it’s ONE market, it’s not AS scalable as spreading zones around between servers) is the main question.
As for playing, as contested mobs is a very small component of the game for most people (most people don’t raid) I don’t have a huge issue with that, and besides, the “top guilds” would probably thrive on the increased competition, rather than being the “guaranteed to the top guild” as is the case on at least some servers now.
The rest is really the social aspect. I almost don’t want to know what would happen to the level-based channels (70-79 in particular) when/if this would happen, but the rest of the effects would be positive. Maybe a bit more of a hassle when zoning (I suspect MOST times you zone you’d have to pick an instance) but other than that, it would be highly beneficial for finding groups, which I find especially at the low-end to be severely lacking.
LikeLike
What really needs to be done is to get rid of zones altogether, present a truly massive amount of content, and do not used named mobs to provide hot loot drops. Make loot provisioning truly random among mob level, so there becomes no benefit to camping a given spot of mob’s level X thru X + 5, since any spot in th eworkd with mobs level X thru X + 5 can be used to attain similar if not identical loot (may even have identical stats just different name that uses the NPC name in the item name).
Then make the world large enough and rich enough to support say 500k simultaneous users, and add more via expansions when necessary. If the game is ever THAT successful, then they company has enough money to crank out more content asap. It isn’t like that kind of success happens overnight.
Zones are not the only way to split load among servers. Grouping game objects by geographic locality is often (but not always!) going to be the best approach, but the lines to not need to be static. If designed to handle it from the get-go, a valley full of 10k players should be able to run just fine across a cluster of servers. It doesn’t have to all sit on just one box in order to work. 10 years ago maybe, but not today. The limiting factor should typically be the client bandwidth.
If architected from the beginning to scale to these extremes, it is completely doable. But you have to plan for it. If you don’t design your persistent storage from the beginning with this goal in mind, you’ll never scale to that point, and you’ll end up having to chunk things up.
The first game that does this and does it well is going to be the WoW killer. Unfortuantely building MMO’s is so hard to begin with that most studios aren’t up to to challenge of taking on the risk of trying something like this which is unprovent and groundbreaking. In game development, what has worked in the past is what is built, 90% of the time. So you keep seeing the same type of game being built again and again. But someday it’s going to happen…
LikeLike
I’m a little late to this show, but I also support the idea.
I’m playing WAR at the moment and it sucks trying to find people to play with, especially in the off-peak times, and especially with people strung out so much in terms of level range. Why they went with ~ 30 servers for launch I’ll never know.
If they “zone-ified” it similar to EQ2 it would be much easier to find people to play with, which is what these games are all about after all.
Rhino, your thoughts about having 10k people in one valley are nice but are a pipedream for the foreseeable future due to the epic lag that would ensue.
LikeLike