CLO: Controller load optimization using multi-step load prediction for software-defined internet of things

IF 6.2 2区计算机科学 Q1 COMPUTER SCIENCE, THEORY & METHODS

Future Generation Computer Systems-The International Journal of Escience Pub Date : 2025-08-26 DOI:10.1016/j.future.2025.108104

Yuanhang Ge , Yong Liu , Qian Meng , Zihang Chen

{"title":"CLO: Controller load optimization using multi-step load prediction for software-defined internet of things","authors":"Yuanhang Ge , Yong Liu , Qian Meng , Zihang Chen","doi":"10.1016/j.future.2025.108104","DOIUrl":null,"url":null,"abstract":"<div><div>Software-Defined Internet of Things (SD-IoT) is a novel network architecture that integrates Software-Defined Networking (SDN) with Internet of Things (IoT) technologies. As the network scales up, increasing service requests impose a heavier processing burden on the control plane, resulting in load imbalance among controllers. Existing switch migration mechanisms have been proposed to optimize controller load. Unfortunately, most current approaches rely solely on real-time network information or the network state in the next time period, which fails to identify overloaded controllers effectively. Moreover, they overlook the load variation trends of switches in the process of switch selection, leading to suboptimal results. More critically, existing methods often struggle to balance load balancing effectiveness and migration cost when selecting target controllers. To address these issues, we propose controller load optimization using multi-step load prediction (CLO) scheme. This scheme adopts the decomposition-based linear model (DLinear) for multi-step load prediction, which helps avoid unnecessary migrations. We further incorporate the Weighted Least Squares (WLS) method to analyze the load trend of each switch, enabling intelligent identification of candidate switches for migration. In addition, we propose a target controller selection algorithm based on an improved Zebra Optimization Algorithm (ZOA), which significantly reduces load imbalance and migration cost. Our approach is based on two assumptions. Firstly, all controllers cannot be overloaded simultaneously. Secondly, each switch can only be connected to one master controller. Under these assumptions, we conduct experiments using Mininet as the emulation platform and Ryu as the controller. Experimental results show that, compared with existing approaches, CLO scheme reduces the average load imbalance rate by 21.3 % and the average response time by 14.1 %.</div></div>","PeriodicalId":55132,"journal":{"name":"Future Generation Computer Systems-The International Journal of Escience","volume":"175 ","pages":"Article 108104"},"PeriodicalIF":6.2000,"publicationDate":"2025-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Future Generation Computer Systems-The International Journal of Escience","FirstCategoryId":"94","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S0167739X2500398X","RegionNum":2,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"COMPUTER SCIENCE, THEORY & METHODS","Score":null,"Total":0}

引用次数: 0

Abstract

Software-Defined Internet of Things (SD-IoT) is a novel network architecture that integrates Software-Defined Networking (SDN) with Internet of Things (IoT) technologies. As the network scales up, increasing service requests impose a heavier processing burden on the control plane, resulting in load imbalance among controllers. Existing switch migration mechanisms have been proposed to optimize controller load. Unfortunately, most current approaches rely solely on real-time network information or the network state in the next time period, which fails to identify overloaded controllers effectively. Moreover, they overlook the load variation trends of switches in the process of switch selection, leading to suboptimal results. More critically, existing methods often struggle to balance load balancing effectiveness and migration cost when selecting target controllers. To address these issues, we propose controller load optimization using multi-step load prediction (CLO) scheme. This scheme adopts the decomposition-based linear model (DLinear) for multi-step load prediction, which helps avoid unnecessary migrations. We further incorporate the Weighted Least Squares (WLS) method to analyze the load trend of each switch, enabling intelligent identification of candidate switches for migration. In addition, we propose a target controller selection algorithm based on an improved Zebra Optimization Algorithm (ZOA), which significantly reduces load imbalance and migration cost. Our approach is based on two assumptions. Firstly, all controllers cannot be overloaded simultaneously. Secondly, each switch can only be connected to one master controller. Under these assumptions, we conduct experiments using Mininet as the emulation platform and Ryu as the controller. Experimental results show that, compared with existing approaches, CLO scheme reduces the average load imbalance rate by 21.3 % and the average response time by 14.1 %.

查看原文本刊更多论文

CLO：基于多步负载预测的软件定义物联网控制器负载优化

软件定义物联网（SD-IoT）是一种将软件定义网络（SDN）与物联网（IoT）技术相结合的新型网络架构。随着网络规模的扩大，越来越多的业务请求增加了控制平面的处理负担，导致控制器之间的负载不均衡。现有的交换机迁移机制已被提出以优化控制器负载。不幸的是，目前大多数方法仅依赖于实时网络信息或下一时间段的网络状态，无法有效识别过载控制器。而且在选择开关的过程中，忽略了开关的负载变化趋势，导致了次优结果。更为关键的是，在选择目标控制器时，现有方法往往难以平衡负载均衡的有效性和迁移成本。为了解决这些问题，我们提出了使用多步负载预测（CLO）方案的控制器负载优化。该方案采用基于分解的线性模型（DLinear）进行多步负荷预测，避免了不必要的迁移。我们进一步结合加权最小二乘（WLS）方法来分析每个交换机的负载趋势，实现智能识别候选交换机进行迁移。此外，我们提出了一种基于改进斑马优化算法（ZOA）的目标控制器选择算法，该算法显著降低了负载不平衡和迁移成本。我们的方法基于两个假设。首先，所有控制器不能同时过载。其次，每个交换机只能连接到一个主控制器。在此假设下，我们以Mininet为仿真平台，Ryu为控制器进行实验。实验结果表明，与现有方法相比，CLO方案的平均负载不平衡率降低了21.3%，平均响应时间降低了14.1%。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

Future Generation Computer Systems-The International Journal of Escience 工程技术-计算机：理论方法

CiteScore

19.90

自引率

2.70%

发文量

376

审稿时长

10.6 months

期刊介绍： Computing infrastructures and systems are constantly evolving, resulting in increasingly complex and collaborative scientific applications. To cope with these advancements, there is a growing need for collaborative tools that can effectively map, control, and execute these applications. Furthermore, with the explosion of Big Data, there is a requirement for innovative methods and infrastructures to collect, analyze, and derive meaningful insights from the vast amount of data generated. This necessitates the integration of computational and storage capabilities, databases, sensors, and human collaboration. Future Generation Computer Systems aims to pioneer advancements in distributed systems, collaborative environments, high-performance computing, and Big Data analytics. It strives to stay at the forefront of developments in grids, clouds, and the Internet of Things (IoT) to effectively address the challenges posed by these wide-area, fully distributed sensing and computing systems.