Fine-Grained Scheduling in Cloud Gaming on Heterogeneous CPU-GPU Clusters

Cloud gaming is a promising approach to provide high-quality gaming services to mobile devices. However, existing cloud gaming systems fail to fully exploit hardware resources due to coarse-grained resource scheduling based on virtual machine migration. In this article, we propose a novel cloud gaming design, referred to as FGCG, with fine-grained scheduling to maximize resource utilization on a heterogeneous CPU-GPU cluster. Specifically, we decompose game workloads into small and independent render tasks that can be freely dispatched to different machines. Trace-driven simulation results show that FGCG can significantly improve resource utilization compared to existing cloud gaming systems.

[1]  Jianwei Yin,et al.  vSpec: workload-adaptive operating system specialization for virtual machines in cloud computing , 2015, Science China Information Sciences.

[2]  Emanuel Falkenauer,et al.  A hybrid grouping genetic algorithm for bin packing , 1996, J. Heuristics.

[3]  Lei Guo,et al.  Virtual network embedding for hybrid cloud rendering in optical and data center networks , 2015, Science China Information Sciences.

[4]  Nei Kato,et al.  State-of-the-Art Deep Learning: Evolving Machine Intelligence Toward Tomorrow’s Intelligent Network Traffic Control Systems , 2017, IEEE Communications Surveys & Tutorials.

[5]  Nei Kato,et al.  Hybrid Method for Minimizing Service Delay in Edge Cloud Computing Through VM Migration and Transmission Power Control , 2017, IEEE Transactions on Computers.

[6]  Li Lin,et al.  LiveRender: A Cloud Gaming System Based on Compressed Graphics Streaming , 2014, IEEE/ACM Transactions on Networking.

[7]  Alessandro De Gloria,et al.  Platform for Distributed 3D Gaming , 2009, Int. J. Comput. Games Technol..

[8]  Yin Wang,et al.  VGRIS: Virtualized GPU Resource Isolation and Scheduling in Cloud Gaming , 2013, TACO.

[9]  Cheng-Hsin Hsu,et al.  GamingAnywhere: an open-source cloud gaming testbed , 2013, ACM Multimedia.

[10]  Ryan Shea,et al.  Rhizome: utilizing the public cloud to provide 3D gaming infrastructure , 2015, MMSys.

[11]  Yaozu Dong,et al.  A Full GPU Virtualization Solution with Mediated Pass-Through , 2014, USENIX Annual Technical Conference.

[12]  Benjamin Hindman,et al.  Dominant Resource Fairness: Fair Allocation of Multiple Resource Types , 2011, NSDI.

[13]  Srikanth Kandula,et al.  Multi-resource packing for cluster schedulers , 2015, SIGCOMM.