论文部分内容阅读
结合虚拟节点技术和均分存储区域技术,提出了嵌套循环式数据一致性哈希优化分布式集群存储的多副本放置策略.按照此优化策略,能够有序选择数据副本机架,确定数据节点存储位置,保证数据存储的均衡性分布,可以针对集群的实际要求开展扩展,并按照扩展情况制定使数据存储完成自适应优化调整,加快数据处理的速度.有效实验表明存储优化后算例的执行速度得到很大提升,能够保证解决负载均衡问题;而针对实际情况中可能出现的扩展与删减问题进行测试后表明,使用优化存储策略处理此类问题时,振荡对整体负载均衡影响不大,且执行时间与负载占比变化趋势一致.
Combined with the technology of virtual nodes and the technology of homogeneous storage area, a multi-copy placement strategy of hash-optimized distributed cluster storage with nested cyclic data consistency is proposed.According to this optimization strategy, the data copy rack can be selected in order and the data nodes Storage location, to ensure a balanced distribution of data storage, you can expand for the actual requirements of the cluster, and in accordance with the expansion of the data stored to complete the adaptive optimization and adjustment, speed up the data processing speed.Experimental results show the implementation of the storage optimization Speed can be greatly improved to ensure that the problem of load balancing can be solved. However, testing for the possible expansion and deletion problems in actual situations shows that the use of optimized storage strategy to deal with such problems has little effect on the overall load balancing, And the execution time is consistent with the change trend of load proportion.