摘要
为了提高集中式网络的请求响应效率,提出并设计了边缘云协同数据存储容灾备份方案。通过将云核心网络的数据存储功能推向边缘网络,使其协同完成用户端的请求需要,降低回传链路压力和响应时间。数据存储策略融合了热度数据生命期与流行度两种指标对用户端兴趣数据进行预测,并结合存储命中率与存储溢出作为存储更新的约束条件,对边缘网络中数据进行动态调整。考虑到网络故障对数据安全带来的影响,设计了数据存储容灾备份框架,根据网络数据更新采取动态备份机制,为了数据备份过程中数据恢复成本与恢复时间的联合求解,引入粒子群算法搜索最优备份策略。仿真结果表明,所设计的数据存储容灾备份方案对用户数据倾向具有准确的预测度,能够对边缘网络数据存储进行准确的动态更新,合理的容灾备份,同时有效提高边缘云网络的协同性、高效性与数据安全性。
In order to improve the request-response efficiency of the centralized network, a disaster recovery backup scheme of edge cloud collaborative data storage was proposed and designed. By pushing the data storage function of the cloud core network to the edge network, the cloud core network would cooperate with the client to complete the request and reduce the pressure and response time of the return link. The data storage strategy combined two indexes of heat data life cycle and popularity to predict the interest data of users and combined the hit rate of storage and storage overflow as the constraints of storage update to dynamically adjust the data in the edge network. Considering the impact of network failure on data security, a disaster recovery backup framework for data storage was designed. A dynamic backup mechanism was adopted according to the network data update. In order to solve the cost and time of data recovery in the process of data backup, the particle swarm optimization algorithm was introduced to search for the optimal backup strategy. The simulation results show that the designed data storage and disaster recovery backup scheme can accurately predict the user’s data tendency. At the same time, it can update the edge network data storage accurately and dynamically, and backup the disaster recovery reasonably. It effectively improves the cooperation, efficiency, and data security of edge cloud networks.
作者
沈刚
陈斌
毛明荣
SHEN Gang;CHEN Bin;MAO Ming-rong(Sanjiang University,Nanjing Jiangsu 210012,China;Nanjing Normal University,Nanjing Jiangsu 210023,China)
出处
《计算机仿真》
北大核心
2021年第5期380-383,412,共5页
Computer Simulation
基金
国家自然科学基金项目(61603193)
江苏省高校自然科学研究项目(18KJB520027)
国家重点实验室赛尔网络下一代互联网技术创新项目(NGII20170524)。
关键词
边缘云协同网络
热度生命期
数据流行度
粒子群优化
容灾备份
Edge cloud collaborative network
Heat life cycle
Data popularity
Particle swarm optimization
Disaster recovery backup
