QoE-Aware Rendering Service Allocation in Fog-Assisted Cloud Gaming Environments
Athanasios Tsipis, Κωνσταντίνος Οικονόμου, Vasileios Komianos, Ioannis Stavrakakis · 2020
Cloudification of heterogeneous networks along with the virtualization of gaming resources, allow the game providers to move into the cloud gaming era. Cloud gaming offloads computation-intensive tasks to the cloud in order to alleviate the computational burden of end-users’ devices, but introduces delay which deteriorates their Quality of Experience (QoE). To solve the optimization problem of rendering service allocation with QoE guarantees and reduce the operational expenditures of game providers, fog computing is deemed a promising solution. However, the optimization problem is NP-hard since it falls into the category of the well-known Facility Location Problems. In the current paper, a distributed policy is proposed, named QoE-Aware Rendering Service Allocation (QoERSA), which takes advantage of strictly local network information, to reallocate the rendering services inside the fog computing network, towards an optimal delay-sensitive placement, given the QoE gaming constraints and the need for capital investment reduction. To this end, it employs Mean Opinion Score objective QoE assessment and simple mechanics that move, copy and merge services with high elasticity and scalability. QoERSA is evaluated through simulations that highlight its efficiency and showcase its capability of overall service access and deployment cost reduction.