Optimization of Resource Management for NFV-Enabled IoT Systems in Edge Cloud Computing

Abstract

The Internet of Things (IoT) has been envisioned as an enabler of the digital transformation that can enhance different features of people's daily lives, such as healthcare, home automation, and smart transportation. The vast amount of data generated by a massive number of devices in an IoT system could lead to a severe performance problem. Edge cloud computing and network function virtualization (NFV) technologies are potential approaches to improve the efficiency of resource use and the flexibility of responsive services in an IoT system. In this paper, we consider the joint optimization problem of gateway placement and multihop routing in the IoT layer, the problem of service placement in the edge and cloud layers of an NFV-enabled IoT system in edge cloud computing (NIoT). We propose three optimization models (i.e., GMO, SP1O, SP2O) that allow an IoT service provider to find the optimal deployment of gateways, the optimal resource allocation for service functions, and the optimal routing according to a cost function with a performance constraint in a NIoT system. We then develop three approximation algorithms (i.e., GMA, SP1A, SP2A) for tackling the problems in a large-scale NIoT system. The evaluation results under a set of scenarios with various topologies and parameters show that the approximation algorithms can obtain results close to the optimal solution with a significant reduction in computation time. We also derive new insights into the strategy for an IoT provider to optimize its objectives. Specifically, the results suggest that an IoT provider should select an appropriate service placement strategy with regard to a charging agreement with an NFV infrastructure provider, and only deploy service functions with a strict delay requirement on the edge of networks for optimizing its cost.

Traffic Engineering Based on Reinforcement Learning for Service Function Chaining With Delay Guarantee

“…Much work has been devoted to finding the optimal traffic engineering solutions in various application contexts and under different constraints [3]- [5], [9]- [12]. For example, Wang et al address the optimization problem of resource management for minimizing the deployment cost of VNF in a hierarchical data center [10].…”

“…Nguyen et al study a quadratic program (QP) for the multiple objective optimization problems of VNF placement and routing with a broad set of key parameters [4]. Recently, Pham et al develop a MILP model for optimizing the VNF placement in NFV-Enabled IoT Systems [5]. However, none of the prior studies provide an optimal traffic engineering solution for delay-guarantee SFC in NFV.…”

“…Hence, the majority of approaches propose approximation solutions for an extensive network. Many papers found in the literature study heuristicsbased approaches, which are generally effective for some specific application scenarios [3]- [5], [10], [13]- [15]. For example, Eramo et al use a heuristic approach based on the Markov Decision Process for the VNF migration problem [3].…”

“…Tajiki et al propose a heuristic for the VNF placement and routing with energy-aware SFC [14]. Pham et al propose a metaheuristic algorithm based on Simulated Annealing for the VNF placement in edge cloud computing [5].…”

Optimizing Traffic Engineering for Resilient Services in NFV-Based Connected Autonomous Vehicles

“…Network Function Virtualization (NFV) has been raised as a potential approach to a flexible and efficient solution for processing a massive volume of data in an IoT system. For the evolution of an IoT system with NFV, we refer the readers to [10]. The reliability of services on the Internet and in an IoT system is a crucial problem that has been widely studied (e.g., [11][12][13][14][15][16]).…”