Revenue-Model Learning for a Slice Broker in the Presence of Adversaries

Abstract

Multi-Access Edge Computing (MEC) and network slicing two of the key enabling technologies of the Fifth Generation (5G) of cellular network. MEC helps to reduce latency, offload the cloud, and allow context-awareness. Network slicing allows to create heterogeneous services on top of shared infrastructures. Slice brokers are emerging intermediate entities that take the resources from the infrastructure providers and make slices for the tenants. In this scenario, a slice broker needs to manage the resource and create the slices in order to maximize its revenue to cover the cost and increase the profit. In this work, we consider that the demand of the slice tenant is depending on the price of the slices. Therefore, we formulate a slice allocation problem that consider this demand-price dynamic. Moreover, we consider the presence of adversary that want to compromise the decision process. In order to solve the problem, we propose a multi-agent environment, where some agents cooperate to learn the revenue model and maximize the revenue. Finally, we evaluate the effectiveness of the proposed solution by comparing it with reference solutions. The results highlight that a notable increment of the revenue can be obtained by using our solution.