DOF servers?

[Darvin]

The difference in playerbase and income between Reverie and Blizzard is one thing, but I think more important is to look at the underlying game. I do not think it's feasible to scale up a conventional MMORPG server system to deal with the size of a MMORTS. If it is possible, there are some cunning optimizations running in the background, but I somehow doubt it.

In a RPG, there is essentially a master server with a master state. If someone cheats to forcibly alter the game state, then the master server will detect a desync and "correct" this state. Barring some sort of bug (many older engines, such as Diablo II, can enter invalid game-states or timestamps that allow you to break the rules) this is very difficult to cheat. The problem is, someone needs to be the dedicated master server, and in a MMO that has to be "the server".

That system is simply too inefficient for a RTS. The entire game-state needs to be analyzed and resent fairly regularly. That's doable with a RPG, but in a RTS the gamestate can be hundreds or even thousands of times larger, so it's harder to detect a desync and harder to repair it. Unless Reverie has some cunning optimizations, I don't think this is feasible at all.


More likely is that the servers will take a hands-off approach, doing little more than keeping tabs on where you are and what you're doing. This would be far less intensive (per player) than a regular MMO. The problem with this is that you could easily cheat while the servers weren't looking. Now, there are ways to work around that. My approach would be a simple audit. Have each player's instance leave a simplified paper trail of data, which the server can sniff through for anomalies. If something doesn't smell right, it does a full audit of the player, silently watching the gamestate and waiting to catch the player in the act of cheating. Not foolproof, but it would work.

[Pilgrim]

Originally Posted by Darvin

Alright, this isn't going to be World of Warcraft. Even presuming that DoF is an awesome game that's very successful (which I hope it is), it can't hope to have a comparable player base. World of Warcraft was the perfect storm of the right developers making the right game at precisely the right time. Dawn of Fantasy, on the other hand, is coming from a new developer, will have to compete with Starcraft II for RTS fans, and its genre is new and experimental. Let's be serious here: 40k people online at once is a pipe dream. DoF would have to be one of the highest selling PC games of all-time to even have a shot at this.

Remember, most people are singleplayer only gamers. Even in MMO's, the vast majority of gamers avoid human contact and would prefer an offline mode if it were available. The vast majority of people who buy DoF will never touch online mode, and only a small fraction of those who do will stick around for the long-run. It may be that we only have 1000-4000 people online at peak hours.

Secondly, the MMO term for "server" and the literal term diverge a little. In reality, one server is not managed by a single machine, but rather a network of custom-built machines. So in reality, one server is actually a network of servers in and of itself. The real question is how Reverie is building their server system.

It may be that from the user's perspective, there is only one "server". This is not to say that there is only one machine managing everything, but rather that all of these machines are part of a coherent network and are indistinguishable to the user. How Reverie decides to build their architecture will determine what it looks like in the end.

Thanks for the reality check!
To be honest, Reverie. is putting an impressive amount of effort into a revolutionary multiplayer considering the company's own realistic player forecast.