Research On Data Dynamic Storage Layout Strategy Based On Load Prediction In Cloud Environment

With the advent of the data information era,traditional centralized data storage methods can no longer meet the storage needs of users,and are gradually replaced by distributed data storage methods in the cloud environment.As the mainstream service platform for the storage and management of big data resources,cloud storage should continuously improve its storage performance to meet the user's efficient access requirements.In cloud storage,data layout strategy is a key factor affecting its storage performance.When users upload a large number of data files to the cloud,a reasonable data layout strategy can quickly locate the appropriate target storage node in the cluster to perform it.Distributed storage,and can enable the system to achieve better load balancing,improve its overall storage performance,and promote the rapid development of the big data industry.This paper takes the data storage layout in the cloud environment as the research object.Aiming at the problems of traditional data layout strategy,such as severe static,information lag and uneven load,a data dynamic storage layout strategy based on load forecasting is proposed.First,an in-depth analysis of the important indicators affecting the storage load to establish an accurate storage node load evaluation method.Secondly,the dynamic optimization of the second exponential smoothing prediction model is used to accurately predict the load information of the storage node.Then,based on the load prediction result,the overall state of the cluster is evaluated,and a dynamic threshold considering the overall state of the cluster is set,and the storage node is divided into three sets of different load states.Finally,the data dynamic storage layout algorithm based on time interval equalization is used to quickly locate the appropriate target storage node among the storage nodes in the low load state set,and the data files to be stored are stored reasonably,which not only improves the load balancing effect of the system,but also improves the load balancing effect of the system.At the same time,the storage efficiency of the system is taken into account.The experimental analysis shows that this strategy not only effectively avoids the information lag problem by applying load prediction to the data layout work,but also dynamically and fully exerts its performance according to the storage state of each storage node during the layout process.The storage performance of the overall system.Compared with the Random layout strategy and the SFLS static layout strategy,the strategy has good effects in terms of storage node data volume consistency,storage node load balancing,and data file upload completion time.