Cloud Storage System Based On MooseFS Used In Industrial Acquisition Data

With the development of Internet, the rise of Internet of Things and cloud computing,building up the industry network become the inevitable trend of the development of industrialmodernization.Facing the difficult of mass data’s management,highly cost of data storage, lowscalability,difficult to guarantee safety and reliability,wise enterprise will build up his ownindustrial private cloud storage system.They adopt distributed collection,management systemand use those characteristics of cloud storage: high universality, high scalability, high reliabilityand Mass storage.The traditional industrial data acquisition system combined with structure ofCloud storage would accelerate the construction of the industrial networking.Industrial data acquisition cloud storage system was architected based on the MooseFS(MFS)which detailed code was analyzed,design philosophy of storage nodes as compute nodes isproposed on this paper. Special distributed storage algorithm is designed for the problem of massindustrial acquisition data is difficult to storage, the dedicated interface industry is provided todata storage.Alot of time and resources should be spended for quering large-scale industrial datain cloud storage system, parallel-distributed query algorithm for the timing data can greatlyshorten the query time and improve the resources utilization ratio. Query efficiency becomes akey problem for query large-scale random data, distributed index algorithm and random dataparallel-distributed algorithm is designed to improve the query efficiency.The main contributions of this thesis are listed as following:(1) Solutions that industrial acquisition data storage to cloud storage is proposed,cloud storageresources pool is also industrial data computing resources pool, compute node which is alsostorage node complete tasks such as create distributed index and parallel computing;(2) Data Mining can be executed on industrial data cloud system:random data index is created inorder to quickly inquires the random industrial data, at the same time the timing data would bealso quickly inquired;(3) This system inherited characteristic of high scalability in cloud storage system, it can bedeployed in different scale cluster environment ranging from3nodes to10000nodes and alsocan be deployed to low-cost distributed embedded system.