Talk:Multiverse Platform Architecture

From Multiverse

• The client submits the username and password for login over TCP to the Multiverse master server. In response, the Multiverse server looks in the multiverse database and provides an authentication token to the client.

• The client submits the worldname of the desired world over RDP to the Multiverse master server. In response, the Multiverse server looks in the multiverse database, and returns the hostname and port of the server responsible for that worldname.

• The client submits the authentication token over TCP to the World server (the server responsible for that worldname). That World server then consults a world user database, and returns a list of characters and the proxy servers and port that should be used for each character. Recent scripts call this World server the Login server to distinguish this server from the server used for management of objects (which doesn't ever communicate directly with the client). Unfortunately Login server is a poor name as well, since the client has already provided the login information to the Multiverse server, and is not providing any login information to this server.

• The client submits the character id over RDP to the Proxy server. At this point, the server starts to send over the world information (information about the terrain and player, followed by other world updates).