Wednesday 28 October 2009

Servers, What’s the Point? – Part 1

We have had cross-server battlegrounds and arenas in World of Warcraft for a long time, now we’re getting cross-server pick-up-groups for dungeons and raids. Games like Aion (and Tabula Rasa and I’m sure others) show us that you can subdivide the population of a region into different channels to make it less crowded while still allowing players to play together. Why are MMORPGs split into servers at all? Why aren’t we all on one big server like in EVE?
Today’s post will deal with the advantages of the traditional server model and try to find solutions for the problems that a one-server world would bring with it. Tomorrow’s part two will then talk about why we would want to condense to one server in the first place and part three (if there is interest in one) will talk about my idea on how such a world could be set up.


Technical Advantages
Handling massive amounts of players requires a high amount of computing power on the server side. Splitting the playerbase into manageable chunks allows the operating company to run several separate clusters of hardware, each taking care of a certain chunk of players. In the traditional model, no communication between these clusters is necessary, solving a severe technical problem.

I am not a hardware specialist at all, but both the WoW and the Aion model show that it is technically possible to transfer a character and all its data from one cluster to another – whether it is the transfer of my shaman to an instance server when I enter the crusader’s coliseum or the transfer of my party in Aion when we switch channels in order to find a region in which the mobs we need aren’t dead yet. It is likely, however, that these transfers don’t span a large physical distance. This means that the instance servers for my World of Warcraft battle group are likely in the same physical location as the hardware that runs the Stormrage world server.

Still, it seems that fifteen of World of Warcraft’s servers can easily access the same data. The technical limit doesn’t seem to be as harsh as it used to be.
Even if it was technically possible, however, to have all the game’s servers in one central location and everyone playing in the same world, latency would become an issue. Australian players will know what it’s like to play on servers that are a large distance away from where you live. The speed of light just isn’t high enough. A certain regionalization therefore seems unavoidable.

Diversification
You often find servers with different rule sets in MMORPGs. Some servers might allow player killing while others do not, some servers are dedicated role-playing servers, and yet others might have an increase in experience gained. If your game requires these kinds of distinctions, you will have to have some sort of separation. I don’t think it would be an insurmountable challenge to game design though to include varying play styles in one world.
The next obvious difference between servers is the language barrier. In Europe you get dedicated English, German, and French servers as well as unofficial servers for other languages. Personally I prefer playing on an English server to playing on one in my native tongue, but there’s clearly a majority that prefers it the other way round. Both German and French servers are highly popular and I suppose the same is true about other languages. The fact the unofficial language servers even exist is proof that many players prefer to play with others that the can talk to in their native tongue. The question here is whether the language barrier would be a big issue in a one-server-world. I daresay that using an Aion-like channel system could alleviate most of these issues. French could play in the French channel(s) and primarily meet people who also speak French. They would also have the opportunity to hop to – say – an English channel if their own one is full of kids (or they want to play with a non-French friend or whatever). More on the channel system later, for now I’d say that language poses an unsolvable problem either.

Have you ever started fresh on another server to have the chance to be one of the top players there? I know I have and I know many others have too. Especially when you are late in joining a game, you’d often rather start on a fresh server with a population full of people like you instead of on a mature one where everyone and their little sister’s dogs are already at the maximum level. In a one-server world you don’t have that chance. If you start late you’ll be a small fish in the pond and potentially unable to catch up. If the game has regular resets however (like certain browser games for example or even the expansions to World of Warcraft), you will just have to wait for one of those for your chance to strike.

Progression
In WoW there are a few guilds fighting for world first kills on raid bosses and other world first achievements. Most players don’t fight in that league, however, but might still be interested in getting server first kills. Being the best guild on your server is one of the few status symbols left in the game at this time. In a one-server world, server-first would be the same as world-first. Everyone else would just be fighting for scraps. That said, my guild is somewhat proud of being the top Alliance guild on our server, even though we are clearly not on top if you take Horde into account as well. This shows us that server-first is not the only first that matters, being the first in a certain subgroup of guilds/people can matter to players as well. A one-server game needs to make sure that those subgroups exist and are clearly distinguishable; otherwise players won’t feel special enough. “You are guild number 2374 on the server” sounds much less impressive than “You are the best guild in [insert region name here].”


Economy
Oh my. Imagine the economy of your World of Warcraft server multiplied by a hundred or so. In the first glance it should scale well as demand will rise in the same amount that supply does – but this isn’t true for every good. If you read the likes of Gevlon and KevMar you will know that the glyph market is often heavily fought over by a couple of players that just offer every possible glyph for sale, relying on the negligible production costs of these glyphs. If the production of a good is so easy that a single player can supply as much of that good as players want to be, introducing a horde of other producers into the market will completely crash it to the point where producing the good isn’t worth it anymore. Right now I can go to the auction house and buy every glyph I want – if the business wasn’t profitable anymore, this supply wouldn’t exist. In general the play style of the auction house monger would probably cease to exist in a one-server world as there would just be too much competition. This is only true, however, if you abstract from WoW’s economy. If you localize trading (with real travel times) and don’t have important items that require next to no effort to produce, you could still have interesting markets and still room for professional traders.

All in all there are issues with creating a one-server world, but except for the latency one, they don’t seem insurmountable. The latency issue means that you would need three or four localized game worlds, but not more. So now that we know that this might be possible, the question obviously is: Why would we want such a world? More on the advantages in tomorrow’s post.
Comments have been disabled because the blog has moved to procrastinationamplification.com. Please head over there, bookmark the new site and comment away :)