Hybrid fog/cloud computing resource allocation: Joint consideration of limited communication resources and user credibility

摘要

In this paper, we study the communication and computation resource allocation problem with the assumption of enough computation resources but limited communication resources. This assumption is indeed practical in a hybrid fog/cloud computing system, when there exists a large amount of data to be executed. More specifically, a powerful cloud computing center can help fog nodes (FNs) release the heavy computation burden. Namely, with the assistance of cloud computing, the system usually has enough computation resources to execute such computation-intensive applications. However, since the system has a certain amount of subchannels, the communication resources of the system may be limited at times, especially when the number of subchannels is insufficient. Therefore, with the aim of handling tasks in an energy efficient way, we propose a communication resource-aware cooperated with computation resources (CRACCR) scheme which has two components. The one is called spectral multiplexing computation consideration, where the system multiplexes communication resources under the consideration of computation resource allocation. The other is called FN scale adjustment (FNSA), where the number of FNs in use is influenced by the communication resource allocation. Furthermore, to develop a user-aware CRACCR scheme, we also design a mechanism to sketch users' credibility. Then a limited communication resource allocation problem with the consideration of user credibility is formulated as a mixed integer non-linear programming problem (MINLP). After transforming the problem by l(p)-box constraints and scale conversion, the problem is tackled by the alternating direction multiplier method. Simulation results prove the improvement of energy efficiency achieved by the proposed scheme, and show the variation of FNs' number while considering the communication resource allocation.