Jens Müller’s research while affiliated with University of Münster and other places

Massively multiplayer online games (MMOG) require large amounts of computational resources for providing a responsive and scalable gameplay for thousands of concurrently participating players. In current MMOG, large data-centers are dedicated to a particular game title. Such static hosting requires a huge upfront investment and carries the risk of false estimation of user demand. The concept of grid computing allows to use resources on-demand in a dynamic way, and is therefore a promising approach for MMOG services to overcome the limitations of static game provisioning. In this paper, we discuss different parallelization mechanisms for massively multiplayer gaming and grid architecture concepts suitable for on-demand game services. The work presented here provides both a state-of-the-art analysis and conceptual use case discussion: We outline the new European project edutain@grid which targets at scaling real-time interactive online applications and MMOG, including First Person Shooter (FPS) and Real-Time Strategy (RTS) games, in an on-demand manner using a distributed grid architecture. Finally, we describe our experimental online game Rokkatan and report experimental scalability results for this game on a multi-server grid architecture.

Massively Multiplayer Online Games (MMOGs) are an increasingly popular class of real-time interactive distributed applications that require scalable architectures and parallelization approaches. While games of the role-playing genre already allow thousands of users to concurrently participate in a single game session, there are important genres, in particular action and strategy games, which have not been scaled to the massively multiplayer realm so far. These games have hard requirements in terms of scalability, in particular regarding density: many players tend to congregate in small locations. In this paper, we outline our novel approach of replication-based parallelisation for scaling the density of players. The practical impact of our work is demonstrated by porting the popular action game QFusion, based on the famous Quake 2, onto our proxy-server system architecture using the replication approach. The experiments with the ported QFusion demonstrate its high responsiveness and show that our approach allows to almost triple the maximum number of simultaneous players on four servers as compared with a single-server version.

Clayworks is a software system which integrates collaborative real-time modeling and distributed computing. It addresses the challenge of developing a collaborative workspace with a seamless access to high-performance servers. Clayworks allows modeling of virtual clay objects and running computation-intensive deformation simulations for objects crashing into each other. To integrate heterogeneous computational resources, we adopted modern Grid middleware and provided the users with an intuitive graphical interface. We parallelized the computation of simulations using a Higher-Order Component (HOC) which abstracts over the Globus Web service resource framework (WSRF) used to interconnect our worksuite to the computation server. Clayworks is a representative of a large class of demanding systems which combine collaborative modeling with performance-critical computations, e.g., crash-tests or simulations for biological population evolution.

While massively multiplayer online role-playing games (MMORPG) involve large numbers of simultaneous players, two other popular game classes -- first-person shooter (FPS) and real-time strategy (RTS) games - are still only rarely considered for massively multiplayer gaming. A main technical problem for scaling these genres to be massively multiplayer is the absence of a suitable scalable multiserver networking approach. While the commonly used zoning concept performs well for an MMORPG, it is barely suitable for RTS and FPS games. As a scalable networking alternative for these genres, this article summarizes our work on the proxy-server architecture, which uses multiple servers for a single game session and implements a full replication of the game state at all proxies. We present our work on the game Rokkatan, our online evaluator game which enables massively multiplayer real-time strategy gaming using our proxy-server network architecture. We discuss the implementation of Rokkatan and analyze the distributed computation of the proxy-server approach by integrating an analytical scalability model into Rokkatan. Our experimental game sessions demonstrate high scalability of Rokkatan, which allows several hundreds of users to participate in a single, fast-paced game session.

Massively Multiplayer Online Games (MMOG) require novel, scalable network architectures for a high amount of participating players in huge game worlds. Consequently, new and complex multi-server parallelization approaches have been proposed to provide responsive, massively multiplayer gameplay for different game genres. Besides scalability and performance, the issue of correctness of the game state processing is vital for providing a failure-free gameplay as expected by the users. In this paper, we first introduce the concept of correctness for multi-server replication architectures as the ability to preserve the order of user inputs in the virtual processing. We then present two correctness mechanisms optimized for multi-server replication: pessimistic lag and optimistic timewarp. We experimentally show that by implementing the lag mechanism for correctness in our multi-server implementation of the QFusion/Quake2-game the amount of incorrectly ordered actions can be reduced from 50% to 10%.

Multiplayer online games have become a popular class of distributed applications with an enormous amount of running Internet-based game sessions. The basic concept of how to provide game services for users has not changed for years: High-bandwidth, dedicated game servers are statically set up to continuously run game sessions, regardless of how many users actually play. This straightforward approach is inefficient, because it does not take the current user demand into account, thus wasting resources. In this paper, we present a novel system architecture for organizing dynamic, on-demand game services for single-server online games. Our system allows users to book game services for immediate play or some time in advance. The system takes the users’ demands into account and dynamically sets up the required server resources in an efficient way. In contrast to the usually offered flat-rate rental of servers on at least a monthly basis, our system allows to charge users depending on the actual services usage and to incorporate new pay-per-use business models.

We describe how C++ programs that use the Standard Template Library (STL) can be systematically parallelized for shared-memory machines. We present our data-parallel template library (DatTeL) and its use in introducing parallelism into sequential STL programs. As a case study, we demonstrate how an STL implementation of the Barnes-Hut algorithm for solving many-body problems can be systematically transformed into a parallel, efficient version using DatTeL. We present experimental results for the multi-threaded version of the resulting parallel program on a SunFire multiprocessor.

We propose a high-level approach to grid application programming, based on generic components (skeletons) with prepackaged parallel and distributed implementations and integrated load-balancing mechanisms. We present an experimental Java-based programming system with skeletons and use it on a non-trivial, dynamic application – the Barnes-Hut algorithm for N-body simulation. The proposed approach hides from the application programmer many complex details of grid programming and load-balancing, and demonstrates good performance on an experimental grid testbed.

We present HOCs — Higher-Order Components — that provide the Grid application programmer with reusable and composable patterns of parallelism. HOCs can be viewed formally as higher-order functions, i.e. a generic implementation of a HOC on a remote machine can be customized with application-specific code parameters which are supplied by the user and shipped via the network. We take the well-known “Farm of Workers” pattern as our motivating example, present an experimental implementation of the Farm-HOC as a Grid Service using the Globus Toolkit, and report first measurements for a case study of computing fractal images using the Farm-HOC.

Computer games played over the Internet have recently be- come an important class of distributed applications. In this paper we present a novel proxy server-network topology aiming at improved scala- bility of multiplayer games and low latency in client-server data transmis- sion. We present a mechanism to ecien tly synchronize the distributed state of a game based on the concept of eventual consistency. We analyse the benets of our approach compared to commonly used client-server and peer-to-peer topologies, and present rst experimental results.