Real-time POV for Cloud Storage Without Caching Hash Values of Files

碩士 === 國立臺灣師範大學 === 資訊工程學系 === 104 === Cloud storage are getting popular and important now. We use them to conserve the files that are significant to us. But there still does not exist an efficient solution to detect and prove the consistency on those files, such as integrity, write in serial and re...

Full description

Bibliographic Details
Main Authors: Hung-Fu Chen, 陳虹甫
Other Authors: Hwang Gwan-Hwan
Format: Others
Language:en_US
Published: 2016
Online Access:http://ndltd.ncl.edu.tw/handle/61686001035749410903
Description
Summary:碩士 === 國立臺灣師範大學 === 資訊工程學系 === 104 === Cloud storage are getting popular and important now. We use them to conserve the files that are significant to us. But there still does not exist an efficient solution to detect and prove the consistency on those files, such as integrity, write in serial and read freshness. We do not want to read the files that out of latest version or got lost. It needs a procedure to clarify the responsibility between us and cloud storage when the disputes happen to us. And no one can repudiate the result. To fix the issue, we propose Real-time POV, which is light duty and fast way designed for client devices and cloud storage. Client devices have no need to cache any hash values of files or other information, but only two short values shared by synchronization server. Cloud storage also only need a small size of data structure, FBHTree, to maintain the attestation besides the files. With combining the protocol proposed in this paper, the system makes client devices have almost the same transaction time comparing to pure file transaction on reading a file from cloud storage for all the size on them. And 2 times higher when writing small files, for example 100 Kb. But also having the same transaction time on bigger files, such as higher than 10Mb.