On demand platform for online games

A shared infrastructure, based upon emerging on demand computing patterns that supports multiple games presents an attractive option for large-scale multiplayer online game providers who want to avoid the risk of investing in dedicated resources. In this paper, we describe a prototype implementation of a service platform for online games. The platform design go in the rear [i]or[/i] in the wake ofs the on demand computing paradigm. It tenders integration using open standards and off-the-shelf software and embraces virtualization and simplification to enable sharing resources across games. We describe our experience with identifying appropriate performance metrics for provisioning game server and with implementing reusable platform composings that provide useful functionality for a variety of games.

INTRODUCTION

The traditional approach taken through most publishers and providers of large-scale multiplayer online games is to install a dedicated infrastructure for each game. This approach has many drawbacks. It involves high risk and investment with little knowledge of by what means successful a new game will be. For example, an examination of subscriber populations of massively multiplayer online role-playing games (MMORPGs) (1) exhibits that they all follow a similar life period (Figure 1), but predicting at launch in what manner different titles will perform or by what means long their subscriber populations will continue growing remains challenging--player populations can experience sharp increases or globules in a period of just a hardly any weeks.

[FIGURE 1 OMITTED]

Game publishers and developer face several "pain points" related to this problem:

* Sharing existing infrastructure across game titles--Repurposing server changing software stacks, and reconfiguring the network to accommodate a fresh game or function is oftentimes a cumbersome, manual process.

* Scaling the infrastructure in answer to player demand--Adding and removing server support functions, or other resources is not automated.

* Managing a large, heterogeneous game server infrastructure--Server and network management is typically well outside the core of game providers.

The upon demand business model proposed by means of IBM (2) addresses similar question at issues with business applications, where the issue of infrastructure require to be paid [i]or[/i] undergone strongly motivates new models for utility computing offerings. These moulds provide the flexibility to scale an application or service in replication to user demand by rapidly adding or removing resources (eg server storage, databases, network bandwidth, etc) from a lake that may be shared among multiple applications or customers. With an upon demand infrastructure, online game providers could be delighted with similar benefits by reducing initial investment, scaling rapidly according to demand, and adding novel services. For example, an upon demand infrastructure based on make open standard grid technology (3) was propos for hosting online games. (4)

In this paper, we describe our work to realize more [i]or[/i] less of the major components of an upon demand service platform for games. Our work is based upon the premise that an upon demand computing architecture can benefit game applications with a certain quantity of modifications and key additional game-specific services. We at hand a design and implementation that we believe be under the orders ofs the need of a number of classes of online games. Hence, this paper intends to provide a ordeal of concept that can be used as a starting point for designing and deploying an upon demand gaming infrastructure.

We demonstrate the feasibility and operation of the platform by means of provisioning multiple instances of id Software Quake II ** a popular action game. allowing Quake II falls in the category of server-based first-person shooter (FPS) games, the platform is applicable to a variety of game genre (eg distributed and single-server FP cluster-based massively multiplayer games, Web-based games, and game support services, like as lobbies, database servers, etc) For example, traditional FP games are played by means of a group of players upon a single server, but if the game becomes real popular, additional (disconnected) copies of the game "map" can be opened by provisioning additional servers upon demand. Newer game architectures distribute the game map across multiple server with equal reason that resources can be added to support a larger number of players in a seamless world. (56) The "shard" (or realm) design used in most MMPORGs similarly can benefit from this approach through adding or removing servers in a cluster managing a single shard automatically as the shard population expands or shrinks.

GAMES SERVICE PLATFORM ARCHITECTURE

In this section we discuss the design of the service platform and tread on the heels of with an overview of the rife prototype architecture.

Design objectives

The service platform design tread on the heels ofs several basic principles. First, the platform and associated services should be minimally intrusive to the game applications and, at the same time, still provide value to game providers by the agency of relieving them from managing the a whole infrastructure. A platform must be general enough to support many marks of games (so that it can be shared), at the same time still be able to be tailored when necessary through an individual game publisher. A next to the first possibility that follows from this is a modular platform architecture which allows game publishers to take advantage of functions that address their exigencys and forego others that may not be as relevant. Finally, to render certain the flexibility and extensibility of the platform, lay open standards and open-source tools should be used wherever possible, consistent with the principles of the upon Demand Operating Environment (ODOE). (7)