Research On Storage Efficiency Of I/O-Intensive Application Based On Cloud Computing Environment

Development of Cloud computing technology provides excellent solution of computing resource and storage resource for big data processing. In recent years, with the rapid development of electronic commerce with the Internet thinking method, emerge of internet of vehicles, large user increase of social network and selfie medias, as a result large amounts of user browser history files, goods properties files, voice files, video files and pictures were produced taking the momentum of high increase. The data center of cloud computing has scaled up as well as data amounts have increased. Researchers of cloud computing and big data have paid more and more attention on the problem of transmitting, processing and results dealing for huge amounts of data sets gradually. For one thing, the object of cloud computing processing is huge amounts of data sets, transmitting to cloud server has an effect on processing rate of I/O-intensive application task. For another thing, storage and reading performance for result files has an effect on response rate of I/O-intensive application task. Therefore, in this paper how to design transmitting and storage methods for huge amounts of data files are researched, which makes a big difference for response rate and storage efficiency improvements of I/O-intensive, reducing energy consumption in cloud data center, enhancing green cloud computing.This paper is based on theory of huge amounts of data sets management and open source cloud computing platform is the research platform, researching the storage efficiency of I/O-intensive application based on cloud computing environment. And the tactics testing and modifying has been done with the help of CloudStack platform. Main research jobs include the following.l)During the transmitting of huge amounts data sets to cloud server by I/O-intensive application task, the files are packed before transmitting with a special size according to the characteristics of files to reduce the intensive of I/O use to release the overload of I/O resources at both sending side and receiving side. In order to improve the transmitting efficiency, file size threshold value is calculated via getting the physical computer parameters and networks bandwidth of cloud computing platform. Before transmitting the files to server, files are packed becoming one file around threshold value. Comparing with other file size, while packing files with threshold the transmitting rate is more higher.2)Parallel processing is introduced to transmitting processing as parallel transmitting. With the overlapping property of time, on the sending side large amounts of files are packed according to the threshold value in different batches, then the package files are transmitted to the server. Meanwhile, on the receiving side the former batches of files are being verified and unpacked according to the transmitting log. Parallel processing method saves the extra unpacking time of serialization transmitting, while is considerable for transmitting and storage of large amounts of files.3)A lot of experiments have been carried out comparing strategy in this paper with traditional methods. And the results show that the strategy got promotion in rate of file transmitting, performance of storage and response performance of I/O-intensive application.