Data Security and Privacy Challenges of Computing Offloading in FINs

Abstract

Recently, a variety of novel techniques (e.g. Internet-of-Things, cloud computing, edge/fog computing, big data, intelligence accelerating chip) make a great number of different devices connected for specific purposes. Based on the significant features of techniques, networking technologies have evolved into future intelligent networks (FINs), in which intelligence has been integrated into networks to help generate and optimize policies, freeing network administrators from management and configuration burdens, and improving the efficiency of self-learning from real-time network data. In FINs, low latency is achieved at the cost of computing-complexity which is beyond the capabilities of Internet of Things devices or users' devices. In order to achieve a new generation of computing-intensive, delay-sensitive and function-intelligent services, computing-intensive intelligence tasks are expected to be offloaded to more powerful edge devices with intelligent computing capabilities. However, because the data are copied or divided before being distributed to edge devices, and that the edge devices have heterogeneous computation resources and various purposes, there exist unknown types of security and privacy threats which would possibly crash the network system, break the data privacy of network entities, damage the data property or cause unfairness in incentives adjustment. In this article, we discuss the design issues for data security and privacy in FINs. We present the unique data security and privacy design challenges presented by computing offloading and highlight the reasons why the data protection techniques in current Internet-of- Things, cloud computing, edge/fog computing cannot be directly applied in FINs.