Mega Worlds -- Is it possible?

[Delsus]

"Feasible" in this context means what would they have to gain for the increased operational costs of creating mega worlds.

Granted, I don't know how the client and servers will operate in version 2.0, but right now they're all buckling with ~300 people sitting in one area (north Ul'dah).

The operational costs of a mega server will be lower, there is one server to maintain rather than 10, so they will need less people working on that, the fuel bill will be lower, running the hardware in one box uses less than 10 boxes etc etc, its a big reason why companies are moving towards server consolidation now.

The implementation costs will be very high (new hardware and possible infrastructure upgrades). Operational costs is the cost to run and maintain the servers, implementation costs is what it takes to upgrade existing servers or buy new ones.

[Nix]

While it sounds nice. . . I would feel so overcrowded on a mega world D: I kinda like my servers on the sleepy-side (I'm so antisocial! lol j/k)

Sometimes it can suck farming karakul etc for crafting mats even with the smaller server sizes and I'd rather not have farming made into some dreadful task.

If we had cross-world content, like a dungeon finder type deelio that'd rock for those who can't find a party for instances etc, but I think it'd do more harm than good to have so many people on the one server><

[Wolfie]

The operational costs of a mega server will be lower, there is one server to maintain rather than 10, so they will need less people working on that, the fuel bill will be lower, running the hardware in one box uses less than 10 boxes etc etc, its a big reason why companies are moving towards server consolidation now.

The implementation costs will be very high (new hardware and possible infrastructure upgrades). Operational costs is the cost to run and maintain the servers, implementation costs is what it takes to upgrade existing servers or buy new ones.

As someone who right now is actively working on re-deploying large scale applications from beefy VMs to small Amazon EC2 clouds, this is totally wrong.

Vertical scaling does not work for server infrastructure because you don't factor in just "day to day" operational costs like electricity or "man power." You have to factor in failures, load balancing, overhead costs, etc.

http://en.wikipedia.org/wiki/Amdahl%27s_law
http://en.wikipedia.org/wiki/CAP_theorem

[Wolfie]

PS: That CAP theorem relates to the CPUs and memory usage of beefy servers. A server with 32 CPUs is not better than 8 servers with 4 CPUs each.

[faris]

In 5 years, it might be commonplace.

[NefarioCall]

The in game world structure needs to be better.
If there's 300 people in ul'dah now, there would be 3,000 then.
There is plenty of empty space out there, but that's also because there's nothing of value in said space.
If for example, the markets can be consolidated and accessed from almost anywhere,
and the UI can allow for universal chat channels, like, for example, a party search channel, a sales channel etc etc,
Then it would be i think, far less of a mess than it would with the current world.

[Diadumenos]

If they did do a "mega" server, the entire game would have to be instanced to keep the work load bearable for the servers. Different "phases" would be issued like in other games. It took DCUO less than a year to get its server numbers reduced to one per region/platform.