Security-Aware and Time-Guaranteed Service Placement in Edge Clouds

Huaiying Sun,Huiqun Yu,Guisheng Fan,Liqiong Chen,Zheng Liu

DOI: https://doi.org/10.1109/tnsm.2022.3213761

2023-01-01

IEEE Transactions on Network and Service Management

Abstract:Most of the emerging applications such as the Deep Neural Networks (DNN) based smart Internet of Things (IoT) systems need intensive and high-performance computing, which is contradictory to the limited resources of IoT/terminal devices. It is a big challenge to offload all tasks to the cloud due to the bandwidth limitation, processing overhead, and transmission costs. Edge computing as an extension of cloud computing that can provide abundant computing resources near the edge of the network and thereby can potentially improve the QoS of applications. However, offloading tasks to the edge servers is liable to external security threats. How to balance the response time and the security of application services is a big challenge for realizing good application service placement. This paper proposes a time and security efficient task scheduling framework in the edge-cloud environment. The corresponding computing models are established, such as the security-related model, time model, and the risk probability model. Then, a time-guaranteed and security-aware task scheduling algorithm is proposed including the domain construction, the security-aware task ranking, and the task dispatching. Extensive simulation experiments have been conducted. Results show that the proposed method has better performance than the other four compared methods in general.