Zhou Jiaxin, Chen Siyi, Kuang Haiyang, Wang Xu
School of Automation and Electronic Information, Xiangtan University, Xiangtan, Hunan Province, China.
Academy of Mathematics and System Science, Chinese Academy of Sciences, Chinese Academy of Sciences, Beijing, China.
PeerJ Comput Sci. 2024 Sep 30;10:e2350. doi: 10.7717/peerj-cs.2350. eCollection 2024.
To analyze performance in cloud computing, some unpredictable perturbations that may lead to performance degradation are essential factors that should not be neglected. To prevent performance degradation in cloud computing systems, it is reasonable to measure the impact of the perturbations and propose a robust configuration strategy to maintain the performance of the system at an acceptable level. In this article, unlike previous research focusing on profit maximization and waiting time minimization, our study starts with the bottom line of expected performance degradation due to perturbation. The bottom line is quantified as the minimum acceptable profit and the maximum acceptable waiting time, and then the corresponding feasible region is defined. By comparing between the system's actual working performance and the bottom line, the concept of robustness is invoked as a guiding basis for configuring server size and speed in feasible regions, so that the performance of the cloud computing system can be maintained at an acceptable level when perturbed. Subsequently, to improve the robustness of the system as much as possible, discuss the robustness measurement method. A heuristic optimization algorithm is proposed and compared with other heuristic optimization algorithms to verify the performance of the algorithm. Experimental results show that the magnitude error of the solution of our algorithm compared with the most advanced benchmark scheme is on the order of 10, indicating the accuracy of our solution.
为了分析云计算中的性能，一些可能导致性能下降的不可预测干扰是不可忽视的重要因素。为了防止云计算系统中的性能下降，测量干扰的影响并提出一种健壮的配置策略以将系统性能维持在可接受水平是合理的。在本文中，与之前专注于利润最大化和等待时间最小化的研究不同，我们的研究从因干扰导致的预期性能下降的底线出发。该底线被量化为最小可接受利润和最大可接受等待时间，然后定义相应的可行区域。通过比较系统的实际工作性能和底线，将健壮性的概念作为在可行区域中配置服务器大小和速度的指导依据，以便在受到干扰时云计算系统的性能能够维持在可接受水平。随后，为了尽可能提高系统的健壮性，讨论了健壮性测量方法。提出了一种启发式优化算法，并与其他启发式优化算法进行比较以验证该算法的性能。实验结果表明，我们的算法与最先进的基准方案相比，解的量级误差在10的数量级，表明我们解的准确性。
