www.Tobold.com
Tobold@GMail.com
A blog with my thoughts regarding games I am playing and other stuff in life. Please read my Terms of Service
RSS Feed
Sunday, September 01, 2013
World of Warcraft connected realms
Per reader request, this is the thread in which to discuss World of Warcraft's connected realms. I think pretty much everybody agrees that this is just a minor variation of server mergers, necessitated by WoW's shrinking player numbers. The one thing that appears a bit nebulous in the announcement and FAQ is what exactly happens when there are multiple players or guilds with the same name on the realms connected that way. Because on the one side you could have names marked with the original server name to make the difference, but on the other side Blizzard said the connection would be "seamless".
The connected realms appear to be a better solution than simply closing under-populated servers. We know that WoW subscription numbers have spikes and valleys. While the general trend certainly is one of decline these days, there will be upward spikes from major patches and expansions. Or an eventual Free2Play conversion. So not closing the door permanently makes sense to me.
But ultimately the question is whether the server model for MMORPGs is still the best available technology. In how far is a single-server model like EVE Online possible for other games today? Why do we in 2013 still have games releasing with a lousy launch experience because the servers are full, only to be followed by news of server mergers some time later? If we KNOW that numbers of players of MMORPGs fluctuate a lot, can't we find a more flexible solution?
The connected realms appear to be a better solution than simply closing under-populated servers. We know that WoW subscription numbers have spikes and valleys. While the general trend certainly is one of decline these days, there will be upward spikes from major patches and expansions. Or an eventual Free2Play conversion. So not closing the door permanently makes sense to me.
But ultimately the question is whether the server model for MMORPGs is still the best available technology. In how far is a single-server model like EVE Online possible for other games today? Why do we in 2013 still have games releasing with a lousy launch experience because the servers are full, only to be followed by news of server mergers some time later? If we KNOW that numbers of players of MMORPGs fluctuate a lot, can't we find a more flexible solution?
- posted by Tobold @ 11:37 AM Permanent Link
Comments:
<< Home
Newer› ‹Older
I think Cory Stockton said how identical names will be handled in wowcrendors interview: http://www.youtube.com/watch?v=TMimojASq48
The idea is to add another char, like the "(*)" in LFG, to indicate that it is a different guild/player.
One problem with servers is that many start out with an even player balance but over time players are going to move to the "better" ones, leaving behind near empty servers.
I can't think of a way to prevent that.
A single-server model like EVE can only work with a low overall population. Gevlon posted a while ago about a massive battle that pushed the server boundaries with 4000 player players in one system.
Now imagine what the Orgrimmar and Stormwind would look like if just a fraction of the whole EU or NA population gathers there, even with todays "low" numbers it is goind to be a couple of hundreds of thousand people.
I mean there was the Swifty incident a while ago that caused server problems by having too many players in the same place.
The idea is to add another char, like the "(*)" in LFG, to indicate that it is a different guild/player.
One problem with servers is that many start out with an even player balance but over time players are going to move to the "better" ones, leaving behind near empty servers.
I can't think of a way to prevent that.
A single-server model like EVE can only work with a low overall population. Gevlon posted a while ago about a massive battle that pushed the server boundaries with 4000 player players in one system.
Now imagine what the Orgrimmar and Stormwind would look like if just a fraction of the whole EU or NA population gathers there, even with todays "low" numbers it is goind to be a couple of hundreds of thousand people.
I mean there was the Swifty incident a while ago that caused server problems by having too many players in the same place.
I wonder if a combination of EVE online's single server and Guild Wars instanced towns would work. Guild Wars used dynamically spawning instances for towns and just spawned more instances if the population got full. Use the single server model for wilderness and other unpopulated areas and then use Guild Wars dynamic spawning instances for towns, dungeons and other heavily populated areas. The beauty of Guild Wars implementation is that it was pretty seamless for the player. On a typical day there might be five version of Lions Arch but during a festival it might jump to 20 or 30. Players could instantly switch between instances to be with their friends.
"I think pretty much everybody agrees that this is just a minor variation of server mergers"
No, I think cross-realm zones were a minor variation of server mergers. This is just flat out server mergers, I don't see even the slightest difference.
I think City of Heroes handled this problem the best with having several instances of the same zone, such as Atlas Park 1, Atlas Park 2, etc. If you really wanted a specific one (to be with your friends) you could select that one. So at launch when you have 5 times as many players, each playing 3 times as much as a player would a year from now, and all of which are clumped at the starter zones, you will have lots of instances of those zones.
Developers don't want to do that though, because it makes it too easy to see accurate activity levels. Nearly every MMORPG should be expecting a big decline from launch, and an exponentially bigger decline in starter zone activity. This will have people screaming, "there used to be 25 instances of [starter zone], now there's only 4!"
No, I think cross-realm zones were a minor variation of server mergers. This is just flat out server mergers, I don't see even the slightest difference.
I think City of Heroes handled this problem the best with having several instances of the same zone, such as Atlas Park 1, Atlas Park 2, etc. If you really wanted a specific one (to be with your friends) you could select that one. So at launch when you have 5 times as many players, each playing 3 times as much as a player would a year from now, and all of which are clumped at the starter zones, you will have lots of instances of those zones.
Developers don't want to do that though, because it makes it too easy to see accurate activity levels. Nearly every MMORPG should be expecting a big decline from launch, and an exponentially bigger decline in starter zone activity. This will have people screaming, "there used to be 25 instances of [starter zone], now there's only 4!"
This is just flat out server mergers, I don't see even the slightest difference.
You shouldn't be nitpicking when you are wrong. There ARE slight differences, like you still logging in using your old server name and not having to remember the new one.
You shouldn't be nitpicking when you are wrong. There ARE slight differences, like you still logging in using your old server name and not having to remember the new one.
"If we KNOW that numbers of players of MMORPGs fluctuate a lot, can't we find a more flexible solution?"
I ask myself the same thing. This problem screams for a cloud-based solution. Think AWS, Google Compute Engine, etc. Blizzard should create this type of infrastructure, rather than rely on commodity hardware.
I ask myself the same thing. This problem screams for a cloud-based solution. Think AWS, Google Compute Engine, etc. Blizzard should create this type of infrastructure, rather than rely on commodity hardware.
This is server mergers with a huge advantage - you don't have to change your name. Returning players also log into the same realm name.
----
I think the SWTOR/TESO/et al superservers give you more flexibility. If you need 30 instances on launch and nine months later early morning midweek in a lull, you can handle it.
Realms have the problem, with activity being so cyclical, how can you have a playable realm at peak still feel full at the minimum? VR was needed and better than mergers, but it does not address the cycle.
----
I think the SWTOR/TESO/et al superservers give you more flexibility. If you need 30 instances on launch and nine months later early morning midweek in a lull, you can handle it.
Realms have the problem, with activity being so cyclical, how can you have a playable realm at peak still feel full at the minimum? VR was needed and better than mergers, but it does not address the cycle.
I ask myself the same thing. This problem screams for a cloud-based solution. Think AWS, Google Compute Engine, etc. Blizzard should create this type of infrastructure, rather than rely on commodity hardware.
I'd be surprised if Blizzard didn't already have their own in-house cloud service. Whether they use for more than just hosting virtual servers is up for debate, though.
I'd be surprised if Blizzard didn't already have their own in-house cloud service. Whether they use for more than just hosting virtual servers is up for debate, though.
"You shouldn't be nitpicking when you are wrong."
YOU shouldn't be nitpicking when YOU are wrong. Login screen? Are you serious? You might as well claim the authenticator somehow means this isn't server mergers.
It's server mergers. There may be individual server merging issues while they are merging servers during this server merger, but you always expect server merging issues when merging servers during the server merging of a server merger.
Server merger.
YOU shouldn't be nitpicking when YOU are wrong. Login screen? Are you serious? You might as well claim the authenticator somehow means this isn't server mergers.
It's server mergers. There may be individual server merging issues while they are merging servers during this server merger, but you always expect server merging issues when merging servers during the server merging of a server merger.
Server merger.
You said there wasn't the SLIGHTEST difference. Being able to keep your name, your guild's name, and your server name IS a SLIGHT difference. So you're wrong.
Samus - I noted your tone yesterday when you first mentioned the subject and were keen to get Tobold to discuss it and I am noticing a similar tone in this thread too.
I get the impression that your interest in this topic is one of axe grinding?
Exactly what is your angle here? Why do you have a beef?
It is as if you are annoyed that Blizzard didn't go for straight mergers?
Almost as though you feel cheated at not being able to post on forums and say "they are merging servers, wow is dying lol".
I get the impression that your interest in this topic is one of axe grinding?
Exactly what is your angle here? Why do you have a beef?
It is as if you are annoyed that Blizzard didn't go for straight mergers?
Almost as though you feel cheated at not being able to post on forums and say "they are merging servers, wow is dying lol".
The problem is that what you want in a good MMORPG is a reasonably small, reasonably stable population, so that you can get to know other characters in the realm. For a long time Blizzard achieved that in WoW, but now it's gone. It's hard to achieve unless lots of players contentedly play in the same realm for years. Single-server games can achieve it, but only if they are small.
Anything else is a second-rate substitute for the desirable situation. The problem is that it is only when the stars come together in the right configuration can the ideal situation be achieved. We might have to accept that there is no way of forcing an MMO into this ideal unless you really have the players by the short and curlies, i.e. your product is so good or for some other reason you can control where and when your customers play.
Maybe, like the '60s, you just had to be there, and now it's too late.
Anything else is a second-rate substitute for the desirable situation. The problem is that it is only when the stars come together in the right configuration can the ideal situation be achieved. We might have to accept that there is no way of forcing an MMO into this ideal unless you really have the players by the short and curlies, i.e. your product is so good or for some other reason you can control where and when your customers play.
Maybe, like the '60s, you just had to be there, and now it's too late.
Didn't GW2 overflow servers prevented most of the server problems at launch and post launch?
Everyone was able to play (aside a few hours in the first day) and there has been no server merges.
And the overflow servers keep working fine in cases of people flocking into new content, just like in this last month of August.
Everyone was able to play (aside a few hours in the first day) and there has been no server merges.
And the overflow servers keep working fine in cases of people flocking into new content, just like in this last month of August.
Didn't GW2 overflow servers prevented most of the server problems at launch and post launch?
Yes, although there were issues with individual party members being shunted off into Overflow while the other members stayed on the realm. Then there is the general feeling (at least when I played) that "my server" didn't exist outside of WvWvW at all. Between zerged events, instant teleports every 500m, the Overflow, and lack of general interaction (aside from rezzing), I was lucky to even catch the name of other people around me. When I cared to.
I have always thought that it'd be better for MMOs to not have named servers at all, but I do see the downsides now. The trick, I think, would be to have no named servers but being able to "tag" certain individuals you run across such that they are more likely to appear in your version of the countryside. Nothing so formal as "friending," or inviting to a guild, but just... associating with people that are positive influences on the game.
Yes, although there were issues with individual party members being shunted off into Overflow while the other members stayed on the realm. Then there is the general feeling (at least when I played) that "my server" didn't exist outside of WvWvW at all. Between zerged events, instant teleports every 500m, the Overflow, and lack of general interaction (aside from rezzing), I was lucky to even catch the name of other people around me. When I cared to.
I have always thought that it'd be better for MMOs to not have named servers at all, but I do see the downsides now. The trick, I think, would be to have no named servers but being able to "tag" certain individuals you run across such that they are more likely to appear in your version of the countryside. Nothing so formal as "friending," or inviting to a guild, but just... associating with people that are positive influences on the game.
Azuriel said: "The trick, I think, would be to have no named servers but being able to "tag" certain individuals you run across such that they are more likely to appear in your version of the countryside"
That's a good idea.
Post a Comment
That's a good idea.
<< Home
