Placement And Migration Algorithm For Big Marine Data In Hybrid Cloud Storage

A variety of high-tech data observation or monitoring equipments are widelyused in marine data collection with the rapidly development of informationtechnology. Satellite remote sensing, aviation, weather balloons, stations, buoys, ships,and underwater sensors are included. The achievement of three-dimensionalmonitoring systems for marine provides the solid foundation for the completerealization of China’s marine information technology. Since the characteristics ofmarine sciences and the various of marine data acquisition means have led to themassive, diversity, real-time, extensity, sensitivity, heterogeneity and other significantcharacteristics of marine data. And the amount of marine data increase rapidly.Marine data will undoubtedly become a model for big data. Not only the massive andcomplex big marine data bring opportunities for information development，newchallenges of the big marine data management. The efficient management of bigmaritime data is the cornerstone and one of bottlenecks for the development of marineinformation technology. Thus, how to management the big marine data efficiently isthe problem to be solved urgently.Cloud storage attracts industry attention which provides new ideas andapproaches to big data management with the rise and development of cloudcomputing. Cloud storage provides an efficient management method for big marinedata which has a large capacity, low cost, high scalability and high reliabilityadvantages. Cloud storage is divided into public cloud, private cloud and hybrid cloudthree types. Cloud storage environment consists of thousands of relativelyinexpensive infrastructure components to provide a large enough space for large-scaledata storage and data calculations. Big marine data has significantly different from thetraditional data characteristics which not only affect big marine data efficientlymanagement, but also affect the big marine data applications. Therefore, the paperputs forward hybrid cloud storage model which suitable for big marine data under theconsideration of big marine data characteristics. The main contents of this paper are as follows.First, the paper summarizes the background and significance of the research.Secondly, it discusses the research status of cloud storage, placement and datamigration based on which analyze the problems in big data management. Due tosignificant features of big marine data there are some problems to be solved in theenvironment of hybrid cloud storage model.1) How to place the big marine data inreason to reduce the transfer and migration in the course of data applications with theconsideration of the significant characteristics of marine data.2) How to determinewhen (When), which data center (Where), which part (Which) data will be migrated.3) How to reduce the expense of data management under the assurance of the datasecurity and data response speed.This section details the significant characteristics of marine data. According tothe spatial characteristics and the characteristics of the big marine data applications,namely in the actual application process the data in close position will be called in thesame time with the high probability. The paper makes correlation analysis of bigmarine data which based on the spatial correlation coefficient. What’s more, we placethe big marine data among the data centers based on correlation between the data. Thehigh correlation data stored on the same data center or close to avoid or reduce thetransfer of data between the data center, thereby reducing the consumption ofresources and time to improve the performance and reduce costs of data storage anddata management.Data migration is the critical issues in the hybrid cloud storage environment. Thepaper comes up with the migration algorithm with the formal description of the issues,precise concepts and definitions of the migration factors. In the migration algorithm,the sensitivity of marine data, data access frequency, data size, the time length of dataand other factors as the migration factors which affecting data migration are fullyweighed. Migration algorithm with consideration of data storage capacity, theproperty features its own marine data and data access processes dynamically.Finally, we verify the validity and effectiveness of the proposed method throughexperiments. The paper verify the correctness of placement in hybrid cloud storageenvironment use certain seas data by analyzing the ArcGIS spatial statistics module.In addition, the paper simulate hybrid cloud storage environment. We compare thedata management cost and data response speed with the traditional approach to datamanagement and proposed approach to data management by using the information is actually stored in a data center status over the years. The experiments show that thedata management cost significantly reduced while migrating through the proposedalgorithm to ensure the speed of data access.The paper is supported by National Natural Science Foundation of China (GrantNo.:61272098). The Ministry of science and technology project973(Grant No.:2012CB316206). The Natural Science foundation of Shanghai (Grant No.:13ZR1455800).