Key Technology Research On Backup System With Twin-engine Continuous Data Protection

Data backup technology is currently a popular data protection technology. It is adata copying technology that used to recover data. The technology always used tocopy production data and it is the basis of the data disaster. Nowadays, with theincreasing of the amount of data, the traditional backup data replication technologyis difficult to meet the needs that people wants, so the research value and researchprospects of data backup systems are very good.A backup system should achieve the goal of mainly lies in three points: the firstis saving storage space of backup, the second is good at backup real-time and thethird is high backup data consistency. In this paper we research for the twin-engineCDP of backup systems, and design and implement a block-level backup systemnamed TE-CDP. And we propose an adaptive scheduling strategy which considersboth the CPU occupation rate and execution time of user tasks. Our works of designand implement the system are as follows:First of all, this paper introduces the research background, the researchsignificance, the research status at home and abroad and the related technologies ofthe TE-CDP backup system.In addition, the paper design and implement the TE-CDP backup system. Wegive the architecture of the TE-CDP backup system. And we propose the singlemirroring method which avoids I/O competition. Besides, the snapshot flow methodthat based on the twin-engine CDP and the consistency agent method of CDP havebeen implemented. These three methods make TE-CDP backup system saving space,good real-time and high data consistency. Based on the three backup methods, weuse the data backup mechanism and the CDP snapshot recovery mechanism. Thesetwo mechanisms increase the efficiency of the TE-CDP backup system. Theexperiment proves that the single mirroring method gives priority to the upperapplication I/O and the TE-CDP backup system can ensure the recovery efficiency inseconds no matter original backup volumes are big or not.Then, this paper presents an adaptive scheduling strategy named the ATS tosolve the problem of high CPU usage because of the multitasking. The strategyconsiders both CPU occupancy rate and the execute time of tasks. When the system is of low CPU occupation rate, the ATS strategy considers executing users’ tasks inshort time. When the system is of high CPU usage, the ATS strategy first considersto reduce the occupation rate of CPU, and then consider ensure more tasks completein a short time. The experiment proves that the ATS strategy can schedule tasksadaptively by considering both CPU occupation rate and the execute time of tasks,and it control the growth of CPU occupation rate of the TE-CDP backup systemeffectively.Last but not the least, we conclude the works and the contributions about thetwin-engine CDP backup system in this paper. We point out the deficiency in theTE-CDP backup system and put forward the direction of future work.