Research On Cloud Storage Strategy Of Eelectricity Consumption Data Based On HDFS

With the development of smart grid and the scale of electirc energy acquire systems constantly expanding,the electricity consumption data by electric energy data acquire systems showing the characteristics of large amounts of data and many types of data,and how to collect and access these data rapidly is an urgent problem for power enterprises.Cloud storage system can solve the problem of electricity consumption data storage with easy expansion and storage capacity effectively.Cloud storage system implements distributed storage of data by replica technology,which can improve the data redundancy and processing capability effectively,however,it can easily lead to a series of problems such as unbalanced distribution of replicas and prolonged replica reading.A good data storage strategy can avoid the occurrence of unbalanced copy distribution effectively,and the dynamic replication management strategy can reduce the read time delay effectively.Therefore,focusing on replication placement strategy and dynamic replication management strategy of the cloud storage,combined with the storage characteristics and requirements of electricity consumption data,the cloud storage strategy of electricity consumption data based on HDFS is proposed,the special studies are as follows:1.According to the analyze the business requirement and storage performance of electricity consumption data,explicate the cloud storage related technologies,and study the read and write process of HDFS,the cloud storage architecture solution of electricity consumption data based on HDFS is put forward.2.Analyzing the storage characteristics and characteristics of electricity consumption data,basically HDFS and combined with the characteristics and application characteristics of BP neural network,the electricity consumption data storage strategy of HDFS based on BP neural network is proposed.The strategy takes the remaining space of disk,memory usage rate,network distance,bandwidth utilization rate and CPU usage rate as evaluation indicators,evaluating the response time of each node by BP neural network and sorting,and then selects the nodes with excellent performance for store data.Simulation results show that the strategy can balance the distribution of replicas,and increase the storage rate of electricity consumption data.3.Aiming at the shortage of HDFS original replication management strategy,the HDFS dynamic replication management strategy based on the access rules of electricity consumption data is proposed.The electricity consumption data files with different access volumes are counted by analysis of the access rules of electricity consumption data,and the quantity of replications can be appropriately adjusted according to the amount of access to electricity consumption data.Simulation results show that the strategy can avoid access to hot issues,reduce data access latency,and improve user access rate.This thesis embarks form the replication of the cloud storage for electricity consumption data,studies the storage and management of electricity consumption data’s replication,and conducts experiments on the proposed strategy by building a simulation platform.The simulation results show that the proposed strategy can solve the storage and access issues of large amounts of electricity consumption data effectively.