Improved algorithm of cloud service node path based on cross-border transaction platform under load balancing

Abstract

Cross-border trading platforms study logistics from the perspective of the network, involving all links in the network and the supply chain. The network status is constantly changing, and the acquisition and analysis of the real-time status of each link on the transaction path can further improve the accuracy of the optimal path selection. Meanwhile, traditional networks can no longer adapt to the rapid development of today’s network technology and applications due to their rigid structure and simplified functions. As for traditional networks and their associated network equipment, problems such as inconvenient management and monitoring, poor interoperability, difficulty in expanding network functions, and high operating costs have severely hindered the efficiency of network functions and the pace of innovation. In order to solve many problems faced by traditional networks, in the paper, an SDN network path selection algorithm based on the dual impact factors of cloud service nodes is proposed. In the algorithm proposed in the paper, the final selected path is considered to be composed of multiple links. First, a cloud environment based on the Ryu controller is build, and a corresponding data collection strategy is constructed to complete the acquisition of the link available bandwidth and delay that affect the real-time status of the path. Then, the weighted results are calculated according to the proposed candidate path weight distribution formula to realize the sorting and screening of each transaction link, and provide support for the selection of the cross-border transaction system service path with load balance. The simulation results show that compared with the traditional DLB algorithm, the algorithm proposed in the paper has better load balancing performance, which further shows that the proposed algorithm has a better choice for the path.

Introduction

As an important branch of the logistics industry, cross-border e-commerce can not only bring greater convenience to cross-border trade, but also promote the flow of international funds. However, while cross-border e-commerce makes outstanding contributions to social development, it also brings greater challenges to the development of logistics enterprises [1]. The optimization problems of cross-border trading platforms mainly include the site selection of network nodes and the optimization of network circuit design. The problem of node location and layout is more complicated. Generally, nodes are selected in consideration of the evaluation index system or the problem is optimized through mathematical models. When solving, some software needed to applied, it mainly uses heuristic algorithms such as modular annealing, neural network, tabu search, or other quantitative methods such as operations research. After sorting out the relevant literature on hub-and-spoke cross-border trading platforms, it is found that the current related research is not deep enough. It is necessary to build an emergency cross-border trading platform model based on theory and practice, and use heuristic methods to solve it [2].

To solve many problems faced by traditional networks, the application of cloud computing has redefined the network. The network function is defined and realized through the cloud platform, which is conducive to the expansion, configuration, and update of the network function, and is also conducive to the function deployment and management monitoring of the network by the administrator. In addition, the cost of network operation and maintenance is reduced, the utilization rate of network resources is improved, and the pace of network innovation is greatly promoted. The core idea and advantage of cloud computing is to separate the control plane of the network from the forwarding plane, and have a unified and open programmable interface, which can realize centralized control of network logic and make network management simple and convenient [3]. Nowadays, the convergence of the three networks and the development of emerging network technologies have caused the network load to rise sharply. Meanwhile, the number of cross-border e-commerce companies is steadily increasing, and the requirements for the quality of network service experience are also increasing. The emergence of e-commerce platforms provides new solutions to many network problems including network path selection, load balancing implementation, and improvement of end user experience quality. Gomes et al. designed an effective genetic algorithm to solve the problem in the single-distribution hub-and-spoke cross-border trading platform network model [4]. Besides, R.S.deCamargo et al. have established a multi-hub hub-and-spoke network with no capacity limitation in order to minimize the construction cost, and conducted an in-depth discussion on the positioning of hub nodes [5]. Li Li and Ding Yizhong established a multi-objective hub-and-spoke cross-border transaction network model, and used the simulated annealing algorithm to solve the model to optimize the network [6]. What is more, Liu Sibei added a time window restriction to the traditional hub-and-spoke network, and used genetic algorithm to solve the network model [7].

NOX is the first open source Software Defined Network (SDN) controller released by the Nick team, and it is also the first to introduce the concept of Network Operating System (NOS). NOX is initially dominated by Nicira. However, with the development of research, it is finally divided into NOX and POX. The former supports the Linux platform and uses the C++ language; the latter supports multiple platforms and uses the python language. NOX adopts a modularized design architecture, and this design concept is used for subsequent development of multiple controllers. NOX mainly includes core modules and application modules. The core modules include Discovery module, Topology module, Routing module, Monitoring module, and Authenticator module. These modules call each other to realize a series of functions for controlling network devices. In addition, NOX also allows users to customize modules and allows third-party software to import modules, which greatly improves the programmability and scalability of the controller. POX is a Python version of the SDN controller released by Nick’s team, evolved from NOX. Its bottom-level module is implemented in C++, and the upper-level application is written in Python. Its main function is to provide a platform for rapid development of network control software prototypes. POX is mainly composed of Core and Component. The Core is the gathering place for all components. With the kernel, one component does not need to import through the Import command to use another component, but only needs to be registered with the kernel. The Components can interact through the Core. The main modules of the Core are: OpenFlow and OF$\text{\_}$01 (communication with OpenFlow 1.0 switch).

In the paper, an SDN network path selection algorithm based on dual impact factors (Joint Bandwidth and Delay path selection algorithm based on SDN, BwD-SDN) is proposed to further optimize path selection and improve overall network performance. First, an SDN network environment based on the Ryu controller is build. Then, the corresponding data collection strategy is constructed to complete the influencing factors (performance indicators) that affect the real-time status of the path. Finally, the weighted result is calculated according to the proposed candidate path weight distribution formula to realize the sorting and screening of each link. The simulation results show that compared with the traditional Dynamic Load Balancing (DLB) algorithm based on the single-hop optimal greedy strategy, the proposed BwD-SDN algorithm has a superior choice for the network path, thus showing better load balancing performance.

Research based on cross-border transaction platform load balancing and SDN network path selection algorithm in the paper is the main content, and improving the network load balancing performance and end user satisfaction is the fundamental purpose. The main innovations of the paper include:

(1) A SDN network path selection algorithm based on the dual impact factors of cross-border transactions is proposed. In this algorithm, the final selected path is considered to be composed of multiple transaction links. Compared with the traditional dynamic load balancing algorithm based on the single-hop optimal greedy strategy, the SDN network path selection algorithm based on dual influence factors has better load balancing performance.

(2) A SDN network load balancing algorithm based on ant colony optimization is presented. With a more comprehensive consideration of the influencing factors that affect the state of the network path, a more optimal selection of the network path can be achieved, thereby showing better load balancing performance.

The structure of the paper is as follows:

Section 1 describes the research background of network path selection and load balancing mechanism based on the cross-border transaction platform network architecture SDN, and elaborates the pressure that traditional networks face in terms of path selection and load balancing. In addition, according to the existing network path selection algorithm, the contribution of the paper is analyzed.

Section 2 gives an introduction to the background knowledge related to the paper, it includes the typical architecture of SDN, an overview of cross-border trading platforms and the problems of using traditional network structures.

A SDN network path selection algorithm based on dual impact factors is proposed in Section 3. Through joint consideration, acquisition, and weighted calculation, the dual influencing factors that affect the link status are realized to optimize the network path selection.

A SDN network load balancing algorithm based on ant colony optimization is proposed in Section 4. By increasing the consideration of server-level influencing factors that affect network path selection, the ant colony algorithm widely used in the field of path search is applied to realize the consideration of path-level influencing factors that affect network load balancing, so the determination of the path to the selected server is completed.

Section 5 compares and tests the SDN network path selection algorithm based on dual influence factors proposed in this paper with the traditional DLB algorithm and SPFR algorithm based on the single-hop optimal greedy strategy by establishing a simulation test environment. The results show that the proposed SDN network load balancing algorithm based on ant colony optimization has a better choice of network paths, and thus has better load balancing performance.

Section 6 summarizes the work done in this article and looks forward to the next research work.