作　　者： (）;

机构地区： 广东科学技术职业学院计算机工程技术学院

出　　处： 《电子设计工程》 2017年第24期109-113,共5页

摘　　要： 传统网格数据负载均衡方法由于在多租户环境下存在数据分布不均衡、节点数据动态增长、节点不同其性能不同等问题,造成了服务器平均响应时间长、最佳适应度小、不稳定、收敛性差等现象,为此,基于数据负载均衡模型和网格数据表提出了矢量数据网格化负载均衡算法。根据数据的负载范围,采用四叉树网格化方法,将外包矩阵范围进行矫正,获取数据负载网格化信息;利用请求时间最近的范围跨度,计算大量数据负载权值以及用户请求范围内网格集合;计算网格集合中数据总量、扫描网格集合,得出分配后数据范围并均衡分配到矩阵范围内,从而完成多机群网格数据负载均衡方法的研究。通过实验对比可知,该算法能够完善传统算法中存在的弊端,缩短服务器平均响应时间,提高最佳适应度,具有良好的稳定性、收敛性和抗过重负载能力,不仅能够针对多租户环境下的多机群网格数据进行负载均衡的研究,还能有效的支持海量数据查询,具有良好的动态性和扩展性。 Traditional grid data load balance method due to exist in a multi-tenant environment data distribution is not balanced , the node data and dynamic growth , different performance problems , caused the server to the average response time is long , the best fitness small , instability , poor convergencephenomenon , therefore , based on the data load balance model and the grid data table using the vector data grid load balancing algorithm are put forward. According to the load range of data , a four-tree mesh method is adopted to correct the scope of outsourcing matrix and obtain the data load grid information. Using the recent range of the request time, computes a large number of data load weights and grid sets within the user request range; Compute grid total collection of data collection , scanning grid , it is concluded that after the distribution within the scope of data range and equilibrium to the matrix, thus a fleet of grid data load balance method , complete research. Through the contrast experiment this algorithm can improve the disadvantages that exist in the traditional algorithm , shorteaverage response time , improve the best fitness , has a good stability , convergence and resistance , can not only for multi-tenant environment more group grid data load balance of the study ,also can effectively support the huge amounts of data query , has good dynamic and extensibility.

关 键 词： 多租户环境下 多机群 网格数据 负载均衡

领　　域： []