Behavior and performance of interactive multi-player game servers

Abdelkhalek, Ahmed; Bilas, Angelos; Moshovos, Andreas

doi:10.1109/ispass.2001.990690

Cited by 32 publications

(36 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Previous studies of game server behavior and performance [Abdelkhalek et al 2001] have also concluded that game servers have very different access patterns compared to scientific workloads, and point to game parallelization as a largely unexplored research area.…”

Section: Related Workmentioning

confidence: 99%

Transactional memory support for scalable and transparent parallelization of multiplayer games

Lupei

Simion

Pinto

et al. 2010

Proceedings of the 5th European Conference on Computer Systems

View full text Add to dashboard Cite

In this paper, we study parallelization of multiplayer games using software Transactional Memory (STM) support. We show that the STM provides not only ease of programming, but also better performance than that achievable with stateof-the-art lock-based programming, for this realistic high impact application.For this purpose, we use a game benchmark, SynQuake, that extracts the main data structures and the essential features of the popular game Quake. SynQuake can be driven with a synthetic workload generator that flexibly emulates client game actions and various hot-spot scenarios in the game world.We implement, evaluate and compare the STM version of SynQuake with a state-of-the-art lock-based parallelization of Quake, which we ported to SynQuake. While in STM-SynQuake support for maintaining the consistency of each complex game action is automatic, conservative locking of surrounding objects within a bounding box, for the duration of the game action is inherently needed in lockbased SynQuake. This leads to higher scalability of STMSynQuake versus lock-based SynQuake, due to a higher degree of false sharing in the latter. Task assignment to threads has a second-order effect on the scalability of STMSynQuake, due to its impact on the application's true sharing patterns. We show that a dynamic locality-aware task assignment to threads provides the best trade-off between load balancing and conflict reduction.

show abstract

Section: Related Workmentioning

confidence: 99%

Transactional memory support for scalable and transparent parallelization of multiplayer games

Lupei

Simion

Pinto

et al. 2010

Proceedings of the 5th European Conference on Computer Systems

View full text Add to dashboard Cite

show abstract

“…Previous work [1] on game server architectures identify parallelism as a potential method of increasing performance.…”

Section: Discussionmentioning

confidence: 99%

A transaction execution engine architecture for multiplayer online games

Lintault¹

2004

Proceedings of ACM SIGCOMM 2004 Workshops on NetGames '04 Network and System Support for Games - SIGCOMM 2004 Workshops

View full text Add to dashboard Cite

In the pursuit of massively multiplayer online games, efficient engines capable of processing sophisticated game scenarios are necessary to support large scale virtual worlds. By applying microprocessor microarchitectural design techniques to game server software design, functional parallelism inherent in complex game transactions can be exploited, augmenting server performance.In a given gaming scenario, complex game transactions can be modeled as data flow graphs. An individual graph represents the collective atomic operations of a particular transaction. Analysis of data flow graphs provides a mechanism that identifies opportunities for parallelism by exposing data dependencies.This poster presents an engine architecture that supports spatial and temporal parallel execution of complex game transactions expressed as data flow graphs. The engine takes advantage of data flow enhancement techniques such as dynamic out-of-order execution to enhance performance. Categories and Subject Descriptors SUMMARYThe increasing complexity of large scale game scenarios demands an appropriate increase in game server performance. Previous work [1] on game server architectures identify parallelism as a potential method of increasing performance.Aside from parallelism exposed by executing transactions simultaneously in separate threads common to typical thread Consider a system model that has multiple clients sending requests to a game server system. The servers are responsible for executing transactions received from clients and distributing the simulation state. Transactions vary in complexity and may require further processing by servers in other tiers of the system architecture.In this model we assume transactions are of two types, simple and complex. A simple game operation represents a single atomic operation. A complex transaction contains a collection of atomic operations that must be executed in a specified order due to data dependencies.Complex transactions can be expressed as a type of acyclic data dependence graph (DDG) known as a data flow graph (DFG). The data flow graph nodes represent a single atomic game operation and the graph arcs represent a true data dependency. These graphs can be analyzed for fine-grained parallelism opportunities.A transaction execution engine was developed that analyzes and exploits parallelism to increase the performance of online gaming systems. The engine features a dynamic core [3] which schedules out-of-order execution of atomic game operations. Functional units implemented as dynamically linked libraries (DLLs) execute atomic operations either locally or remotely in an asynchronous fashion. The processing pipeline stages and functional units are realized as object oriented software.

show abstract

“…In [2], the authors are discussing the behavior and performance of multi-player game servers. They find that in the terms of benchmarking methodology, game servers are very different from other scientific workloads.…”

Section: Related Workmentioning

confidence: 99%

LEARS: A Lockless, Relaxed-Atomicity State Model for Parallel Execution of a Game Server Partition

Raaen¹,

Espeland

Stensland

et al. 2012

2012 41st International Conference on Parallel Processing Workshops

View full text Add to dashboard Cite

Abstract-Supporting thousands of interacting players in a virtual world poses huge challenges with respect to processing. Existing work that addresses the challenge utilizes a variety of spatial partitioning algorithms to distribute the load. If, however, a large number of players needs to interact tightly across an area of the game world, spatial partitioning cannot subdivide this area without incurring massive communication costs, latency or inconsistency. It is a major challenge of game engines to scale such areas to the largest number of players possible; in a deviation from earlier thinking, parallelism on multi-core architectures is applied to increase scalability. In this paper, we evaluate the design and implementation of our game server architecture, called LEARS, which allows for lock-free parallel processing of a single spatial partition by considering every game cycle an atomic tick. Our prototype is evaluated using traces from live game sessions where we measure the server response time for all objects that need timely updates. We also measure how the response time for the multi-threaded implementation varies with the number of threads used. Our results show that the challenge of scaling up a game-server can be an embarrassingly parallel problem.

show abstract

Behavior and performance of interactive multi-player game servers

Cited by 32 publications

References 4 publications

Transactional memory support for scalable and transparent parallelization of multiplayer games

Transactional memory support for scalable and transparent parallelization of multiplayer games

A transaction execution engine architecture for multiplayer online games

LEARS: A Lockless, Relaxed-Atomicity State Model for Parallel Execution of a Game Server Partition

Contact Info

Product

Resources

About