Research And Implementation Of A Private Cloud And Mobile Storage Based Continuous Data Protection Technology

In the ear of big data, data security has become increasingly important. In order to meet the higher demand for data protection, continuous data protection (GDP) came into being. CDP can capture and track the changes in real-time, and place these changes outside the source data to ensue recovery to any point in the history. For each operation, CDP will calculate the differential, and then store it in another area. So CDP can restore data to any point in the history theoretically.This article will firstly describe the traditional data protection technology and CDP. Then describe our own CDP architecture from the point of scalability and mobility. This article will conclude:(1) Traditional CDP server is based on single disk, its scalability and security is poor. In this paper, we will use private cloud as the CDP server data storage. Since private cloud implementation, mechanism, it can ensure the data security. When the storage space is not enough, private cloud can expand the storage space without stopping CDP server. This article describes the mainstream of private cloud, and highlights MooseFS distributed file system as the CDP server’s private cloud.(2) CDP service objects are typically LAN users, which limits the use of CDP. This paper proposes the use of U disk as a storage medium, expands the CDP service to the mobile storage device. This article will describe the CDP architecture from point view of data security and consistency.(3) CDP server only within the LAN is not security, enough, there is a need of remote mirror to back up CDP server. This paper will use inotify and rsync to build remote mirror.