混合云环境面向安全科学工作流数据布局策略  

Security-oriedted data placement strategy of workflows in hybrid cloud environment

作　　者：苏明辉 林兵 卢宇[3] 王素云 SU Ming-hui;LIN Bing;LU Yu;WANG Su-yun(College of Physics and Energy,Fujian Normal University,Fuzhou 350117,China;School of Electronics Engineering and Computer Science,Peking University,Beijing 100871,China;Concord University College,Fujian Normal University,Fuzhou 350117,China)

机构地区：[1]福建师范大学物理与能源学院,福建福州350117 [2]北京大学信息科学技术学院,北京100871 [3]福建师范大学协和学院,福建福州350117

出　　处：《计算机工程与设计》2024年第7期2004-2012,共9页Computer Engineering and Design

基　　金：国家重点研发计划基金项目(2018YFB1004800);国家自然科学基金项目(62072108、61672159);福建省高校产学合作基金项目(2022H6024);福建省自然科学基金项目(2019J01244)。

摘　　要：为解决混合云环境下科学工作流数据布局问题,在考虑数据的安全需求的前提下,以优化跨数据中心传输时延为目标,提出了一种混合云环境下面向安全的科学工作流布局策略。分析数据集的安全需求以及数据中心所能提供的安全服务,提出安全等级分级规则;设计并提出基于遗传算法和模拟退火算法的自适应粒子群优化算法(adaptive particle swarm optimization algorithm based on SA and GA,SAGA-PSO),避免算法陷入局部极值,有效提高种群多样性;与其它经典布局算法对比,基于SAGA-PSO的数据布局策略在满足数据安全需求的同时能够大大降低传输时延。To address the issue of data layout in scientific workflows in a hybrid cloud environment,while considering the security requirements of the data,a security-oriented optimization strategy was proposed for scientific workflow layout with the goal of optimizing the cross-data center transmission delay.The security requirements of the dataset and the security services that could be provided by the data center were analyzed,and a security classification rule was proposed.An adaptive particle swarm optimization algorithm based on genetic algorithm and simulated annealing algorithm(SAGA-PSO)was designed and proposed to avoid the algorithm from getting stuck in local optima and effectively improve population diversity.Compared with other classic layout algorithms,the data layout strategy based on SAGA-PSO can significantly reduce transmission delay while meeting the data security requirements.

关 键 词：混合云 科学工作流 数据布局 安全分级 时延优化 遗传粒子群优化算法 模拟退火 

分 类 号：TP338[自动化与计算机技术—计算机系统结构]