CALM: QoS-Aware Vehicular Sensor-as-a-Service Provisioning in Cache-Enabled Multi-Sensor Cloud

Abstract

This work addresses the problem of QoS-aware cache selection for provisioning vehicular sensors-as-a-service (Se-aaS) in the presence of multiple sensor-cloud service providers (SCSPs), vehicle owners, and roadside units. In cache-enabled sensor-cloud, each SCSP serves the applications from its internal and external caches, which may not be feasible in vehicular networks due to lack of resources. Moreover, in vehicular sensor-cloud, the SCSP may not have access to the required data; hence it loses revenue. To address these issues, we propose a game-theoretic cache orchestration scheme, named CALM, for multi-sensor-cloud. CALM is performed in two stages. In the first stage, after receiving the end-users’ service request, the requested SCSP identifies the optimal subset of internal caches using an expected utility theory. If the service is not executed in the first stage, the SCSP identifies the optimal subset of external caches at the RSUs in the second stage. In this stage, we use a single-leader-multiple-followers Stackelberg game, where the SCSP and the RSUs act as the leader and followers, respectively. We evaluate the performance of CALM theoretically and through simulation, while comparing it with the existing schemes. Simulation results depict that using CALM, the profit of SCSP and service delay improve by 33.91% and 20.79%, respectively.