Design And Implementation Of Cloud Storage Management System For Heterogeneous Ocean Data

With the deepening of ocean exploration and research,ocean big data has attracted more and more attention.In particular,the construction of marine ecological environment is an indispensable part of the national strategy of strengthening the country.However,in recent years,due to the increasingly serious marine pollution,many marine ecological disasters have occurred one after another,which not only seriously damaged the ecological balance,but also caused serious economic losses.Therefore,the importance of an efficient and accurate marine monitoring system is becoming more prominent.However,due to the large scale and strong heterogeneity of ocean monitoring data,how to store massive amounts of heterogeneous data quickly and efficiently becomes a maj or challenge.With the rise of cloud storage concepts,it will replace traditional local storage as a better solution for such scenarios.However,cloud storage also has certain disadvantages,such as high access latency and insufficient network bandwidth.How to combine the low latency and high bandwidth of LAN storage with the high performance and high scalability of cloud storage is an important issue worth studying.At the same time,in terms of storage media,SSDs have gradually replaced HDDs in storage structures because SSDs have better shock resistance,lower power consumption,shorter seek times,and higher performance than traditional HDDs(mechanical hard disks).The roles in the public cloud storage,large data centers and other scenarios have been widely adopted.In the field of ocean monitoring,the immediacy and high value of data determine its need for highly reliable storage.The reliability of SSD is not as good as that of HDD in the whole life cycle.How to improve the reliability of SSD is also an important issue to be studied.This article aims to solve the above problems.Firstly,the paper focuses on the problem of high data access delay and insufficient network bandwidth,and combines it with LAN storage to propose a three-tier data storage management architecture on the cloud-edge-local side.By buffering the hotspot data on the cloud platform side to the edge side,the delay when the local side acquires data is reduced.At the same time,the storage format of the database is designed to meet the requirements of easy query and retrieval on the cloud platform side and the high-speed and low-latency requirements of the edge side,so as to maximize the performance of the local side to obtain data.Secondly,the paper proposes a multi-layer hybrid RAID-5 array using SSD+HDD to achieve high-performance HDD-aware data backup in the RAID controller layer for the shortcomings of SSD+HDD due to the limited reliability of SSD.The strategy and the data recovery strategy enable the disk array to keep the data from being lost or damaged when the multiple SSDs fail.This solves the problem of continuous failure of multiple disks faced by the all-flash disk array,and effectively ensures the reliability of the cloud platform data.Finally,this paper develops a cloud storage management system for ocean multivariate data,and uses a responsive web design language to write a multi-platform adaptive front-end page for displaying the multi-heterogeneous data of the ocean and the reliability of the three-tier architecture.The real-time performance of data acquisition was tested accordingly.The test results show that compared with the traditional architecture,the work done in this paper effectively improves the real-time data access and the reliability of data storage,and has a good application prospect and social significance.