All part of the service
Oh, and hosting of servers that can handle 1000's of players will be too expensive. Unless you have financing, and a lot of it.
And, creating a client-server environment like that is
different than your typical peer-peer game. Diablo II has the realms, Asheron's Call and EQ come to mind as well. They have tremendous challenges just keep the service running, let alone the naked expense of hosting. Diablo Realm servers are hosted at Exodus.
I have built datacenters at Exodus facilities. BELIEVE me, you don't have the money yet. Once you start getting into the 1000's of concurrent connections, you start having to care about real time loadbalancing (F5 makes a hardware load balancer that works for
of IP based trafiic, as does Cisco, plan on 6-25 thousand dollars each, and plan on buying 2 of them, and plan on having to modify them to work with non-web UDP traffic.), and then you have to consider the machines that the service actually runs on. Not your average desktops, take the Compaq DL580, 4 Processors, 1 gig of RAM, 5x18gb SCSI disks, with WinNT4.0 enterprise licenses, 20 k a pop. Not to mention the enterprise class databse you will need to build and maintain to keep track of all those players
Sure you can save a little cash by going Linux for the backend, but the hardware doesn't get any cheaper
And I haven't even hinted at the money you need to host at a place like Exodus. Blizzard can afford it because Diablo has topped the sales charts since it's release, and they probably have a colo deal that a new-to-the-field group won't have. AC can do it because they are backed by MS and charge for their service. EQ can do it because EQ is like crack. Ultima Online couldn't do it, so they cancelled.
Build the peer-peer game first, the centralized server things is a whole world of financial hurt you want to avoid your first time out. It will be challenging enough. I'd recommend going for a reasonable player limit per game (if you are really using the Tribes 2 engine, which is a good choice for large areas and large numbers of players, 60 sounds like a good number), and supporting that with AI. Sorta a combination Tribes and Starcraft
Anyway, first things first. Have your 3 programmers grab one of the other engines while we wait for garagegames to notify us, and build some demos. Morfit was the easiest, or try one of the others. Just build a ship-ship multiplayer (2 player) game that fly's over some terrain. Have the ship be stock models. That becomes your tech demo, your "proof of concept" for some of the basic stuff. Take that skeleton game and swap models and terrain out, make sure you are good at that kind of stuff, then add models to it, etc. Will give you a CLEAR indication of what you are getting into and what you need to do to succeed
Of course, write EVERYTING down.
[This message has been edited by Inquisitor (edited 04-07-2001).]
