Towards Scalable Cloud Gaming Systems: Decoupling Physics from the Game Engine

  • Saulo Soares De Oliveira UFG
  • Carlos Henrique R. Souza UFG
  • Jefferson Carvalho Silva UFG
  • Sérgio T. Carvalho UFG


By leveraging cloud computing resources, cloud gaming, also known as Games-as-a-Service (GaaS), has emerged as a new computer game delivery paradigm that promises gaming anywhere, anytime, on any device. In this regard, game engine architectures are one of the main objects of study when adapting digital games to cloud gaming. However, most of the proposed cloud gaming architectures cannot take full advantage of cloud computing resources to offer system scalability due to a monolith design. In this paper, we evaluate the impacts of a distributed architecture in terms of system performance by employing the service-oriented architecture paradigm to offload the game physics calculations as a decoupled system. As a differential, our work contributes to the discussion by implementing a proof-of-concept to compare the game engine performance between the monolith and distributed approaches by utilizing a modern public cloud provider, which as far as we know, has not yet been done in the literature. As a result, the distributed approach had a better performance in computation-intensive physics calculations scenarios. In this case, the communication overhead is outweighed by performance gains, as the physics engine was deployed on a compute-optimized VM. This further indicates the benefits of decoupling game engine systems, not only better game performance but also taking a step further for a scalable cloud gaming system, which can scale better both vertically and horizontally.
Palavras-chave: Cloud computing, Cloud gaming, Distributed architecture, Games-as-a-Service, Scalability
OLIVEIRA, Saulo Soares De; SOUZA, Carlos Henrique R.; SILVA, Jefferson Carvalho; CARVALHO, Sérgio T.. Towards Scalable Cloud Gaming Systems: Decoupling Physics from the Game Engine. In: SIMPÓSIO BRASILEIRO DE JOGOS E ENTRETENIMENTO DIGITAL (SBGAMES), 22. , 2023, Rio Grande/RS. Anais [...]. Porto Alegre: Sociedade Brasileira de Computação, 2023 . p. 151–160.