Agents-based modeling for a peer-to-peer MMOG architecture

  • Authors:
  • Abdennour El Rhalibi;Madjid Merabti

  • Affiliations:
  • Liverpool John Moores University, Liverpool, U.K.;Liverpool John Moores University, Liverpool, U.K.

  • Venue:
  • Computers in Entertainment (CIE) - Theoretical and Practical Computer Applications in Entertainment
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

Massively Multiplayer Online Games (MMOGs) are becoming a very important part of computer entertainment business. With the recent development of broadband technologies, the increase in the number of players is putting a strong pressure on this type of application. Commonly used clients/server systems don't cope well with scalability, limiting the number of players who can interact with each other, are not robust enough, and might be subject to bottlenecks due to their centralized infrastructure. These systems also force developers to invest enormous amounts of money in hardware and time to design complex software systems. To solve these problems we propose a fully distributed, peer-to-peer architecture for MMOGs.We discuss the issues surrounding MMOGs, the limitations in terms of network infrastructure and the lack of a simulation environment to study and evaluate network architectures and protocols. We use a peer-to-peer (P2P) based architecture and protocol to provide a more scalable, flexible, and robust technology solution than do currently used infrastructures. We conducted the design and implementation of a modular MMOG, called "Time-Prisoners," using a P2P protocol developed in Java and JXTA. The characteristics of P2P overlays enabled us to organize dynamically, and in transparent way for the users, the group of players according to their locations in the virtual world, and allowed the design of a scalable mechanism to distribute the game state to the players and to maintain a consistent world in case of node failures.