DSOS: A Distributed Secure Outsourcing System for Edge Computing Service in IoT

Edge computing can help the resource-constrained Internet of Things (IoT) devices to perform some complex tasks. Edge computing has many advantages, such as the distributed architecture, low interaction latency, and good resilience. It can provide fast response and reliable service for the IoT applications. There exists an important application in the edge computing environment is to outsource the computationally intensive problems to nearby edge servers, which has become a research hotspot in the area of industry and academia. In this article, we propose a distributed and secure system DSOS for seeking the least squares solution to the overdetermined system of linear equations (OSLE) with the assistance of multiple nearby noncolluding edge servers. Solving this type of linear equations is one of the most common problems for statistics and data mining in IoT. In our system, the coefficient matrix is divided into multiple blocks according to rows. The different blocks are distributed to the different edge servers. All edge servers can help the user to obtain the correct solution to the OSLE by the interactive computation. Our system can ensure the private information about input parameters and final results is not leaked to the participating edge servers. In addition, the mutual verification between the edge servers can ensure the validity of the final results. The experimental evaluations show that the designed system outperforms the existing ones in terms of fast response between the edge server and the user, low computation overload on the edge server side, and high efficiency on the user side.

Automated summary

Edge computing can assist resource-constrained IoT devices in performing complex tasks. Its distributed architecture and low latency provide advantages such as fast response and reliable service for IoT applications. This article proposes a distributed and secure system that utilizes multiple nearby noncolluding edge servers to find the least squares solution to overdetermined systems of linear equations. Experimental evaluations demonstrate that the designed system outperforms existing ones in terms of response speed, computation overload, and efficiency.