A Distributed Social MUD to Enhance Reliability and Scalability

MUD (Multiple User Dungeon) systems are client-server. When the single server goes down, everyone is kicked off and no one can log on. Also, even if there are only a handful of people talking in a given room, a large overall number of users can slow the server and cause lag (higher latency) which makes fluent communication impossible. We introduce Happy Fun MUD, a distributed, peer-to-peer MUD that implements the essential communication commands of simulated online worlds. A gateway allows peers to load the map and to enter the first room. Peers can then move about; the first peer in a given room is designated its innkeeper. The innkeepers coordinate activity between rooms in the MUD. When an innkeeper leaves a room, the responsibility is assigned to someone else, if there is anyone left in that room. Peers send what they say directly to other local peers; they whisper private messages directly to the peer who is addressed in the communication. The system is more robust to the failure of peers or the gateway than the clientserver system is to a server failure. The busiest node is also much less busy than is the server in the client-server system. 1 Why Create a Distributed Social MUD? 1.1 MUDs Are Social Worlds A MUD is an all-text virtual environment. It is called a Multiple User “Dungeon” after a free Fortran version of the interactive fiction Zork that was released in 1979, called Dungeon. Although many people think of puzzle-solving and fantasy quests in relation to MUDs, it is actually the “multiple user” aspect that makes them appealing to many people. Many MUDs serve as social spaces where people have the same sorts of conversations they do “in real life,” talking with friends on the phone or chatting with coworkers at the water cooler. The simulated world that the MUD provides