It looks like you're new here. If you want to get involved, click one of these buttons!
I wonder if the "MegaServer" concept is related to supporting TESO's more action based combat and latency issues.
Most folks here, I think are familiar with the concept of latency and how it can negatively effect more action oriented - faster paced combat, especialy PvP.
For those unfamiliar, a quick primer. The technical definition of "latency" is essentialy the time it takes for a packet of information to travel from one point on the internet (or any network) to another. In game terms what it really means is how long does it take between the time some action is registered on a server (eg. Player A swings a sword) and it is transmitted to the computers of other players playing that game so that it can be displayed on thier screen and they have the ability to REACT to it...and of course send those reactions back to the server running the game so that they can be registered by it. If you've heard the term "ping" in games, that's a measure of it.
It can be a technical hurdle for more action oriented games (which require a fast pace of play or fast reaction times) played over the internet. For example if the player is required to react to a sword swing in 800ms (milli-second) but the round trip latency for getting that command back to the server where it can be registered is 900ms, you've got a problem...as the player can never possibly react quickly in time, no matter how quick they are on the keyboard.
Latency tends to be determined by 2 major factors....
1) Geography. Although data tends to travel very fast over distances on the internet, it's NOT instantaneous. If a game is processing data in real time on a server based in Chicago then a player based in Indiana is going to have a palable game-play advantage over a player based in Australia.
2) What's happening on the networks/networking equipment in between that server and the client. Latency is NOT constant because conditions can vary on the networks and equipment between the 2 points. For example, an intermediary network may be handling a higher volume of traffic on one day from the next, slowing down the speed at which data transits that network or equipment on an intermediate network can be malfunctioning causing data to have to travel a diffferent (longer) path. For example the normal path for data between a game server in Chicago and a player in Boston may be through a network in Buffalo, but if networking equipment in Buffalo is malfunctioning, that data may get rerouted through Nashville instead....causing that data to travel a longer distance and also causing the equipment in Nashville to carry a greater load of traffic then normal. End result is that if your latency to a server was 90 ms on Sunday it could be 600 ms on Monday.
The way FPS games deal with this is simple, they tend to have multiple servers located on multiple networks in different geographic locations. Since the players don't have persistant characters when they play, it's no big deal...if you have a high latency to a particular server on a particular day, you just choose to play on a different server, problem solved.
Traditional MMO's under the "shard" model can address issue #1 by simply having different "shards" in different data centers across the world. So North American players can play on North American "shards" and European Players Play on European ones. Since players characters are persistant on a particular "shard", they haven't traditionaly had a way to address #2, however. It's one of the reasons MMO's have tended to shy away from the faster paced "actiony" type combat seen in other types of games.
I wonder if Zenimax is using the "MegaServer" idea in part to address the latency issues that could affect the more "actiony" combat we are used to in the Elder Scrolls SPRPG's? Due they perhaps run thier individual "instances" out of different "Points of Presence" located in different geographicaly disperesed Data Centers and part of the decisioning process of what "instance" the game tries to put you on is based upon it's determination of your latency to those different Points of Presence? It has me curious.
Comments
It sounds good on paper because you don't have to worry about spinning server merges though.
Hopefully it will be implemented better than GW2's overflow servers. That was awful.
The gw2 overflow approach is more elegant.
Latency and population management are probably major reasons for the megaserver concept. Unfortunately, instancing zones to high Hell also comes with the issues ShakyMo mentions. Persistance is kind of important for MMOs to retain that "world feel" many are looking for.
I will not play a game with a cash shop ever again. A dev job should be to make the game better not make me pay so it sucks less.
Yeah but that kind of went out the window in 2005 when WoW and its heavily instanced clones started to become literally every AAA MMO.
They've probably lost the daoc fans too when the implications of the server architecture hit home. Slightly easier instanced dungeon grouping at the expense of totally cocking up the pvp endgame.
WoW at least has a huge, open seamless world. And back then the game was much more than standing around in Orgrimmar and queueing for instances.
I don't disagree. Not a big fan of instancing myself....but there are only so many ways you can approach the latency issue for fast paced combat.
Edit: Either you slow down combat pacing (which I personaly wouldn't mind)...or you find some way to assign users to server instances they have good connectivity to....or you live with the fact that the game is going to unplayable for someone that is having a high latency that day....pick your poison.
I'd rather be in an overflow server than a queue.
I can't speak for Darkfall because I haven't tried it myself but I have heard second hand that it was basicaly unplayable for people who weren't experiencing good connectivity (i.e. low ping).
"Latency" isn't really something that is within the hands of a Dev...they can't control how fast a packet of data travels over intermediate networks between them and the end user. The only thing they can do in that regard is place thier servers in different, geographicaly dispersed Data Centers which are closer and have better connectivity to the clients, but then you lose the "single server" thing. They can do ALOT about other types of lag (server/client), resource usage, even bandwidth induced....but network latency is completely out of thier hands.