| With the coming of big data,personal,business,government and scientific research are producing and using data all the time.Data is not only a combination of ‘0' and ‘1',but also become the cornerstone of the development of the whole society.This development is bound to put forward higher requirements for the capacity,transmission rate and reliability of the storage equipment.Continuous data protection(CDP)provides unlimited granular recovery point objective(RPO)and nearly instant recovery time objective(RTO).CDP has been widely used in most data centers.But it has a requirement for great fluctuations in the performance of storage system.When access is active(usually during the day),the system requires higher storage bandwidth;as it is inactive(during the night),the system requires less storage bandwidth.Faced with such dynamic loads,traditional RAID layout may cause a performance bottleneck or higher power consumption.Aim at this problem,MS-RAID,an energy-saving data layout with multi-level storage space is proposed on the basis of S-RAID.MS-RAID has different levels of grouping strategies.Storage system's performance can meet the demand for a realtime dynamic load by adjusting the number of parallel disks.The specific research contents are as follows:Firstly,this paper introduces the storage technology,analyzes the principles,advantages and disadvantages of backup,snapshot and CDP,further analyzes the principle and reliability of the common RAID scheme,puts forward the theoretical basis for adopting the appropriate RAID plan according to the specific needs.And in the actual observation and analysis of the CDP storage system,there are high time related load fluctuations,such as securities companies,which have high requirements for data security.During the business period(daytime),a real time backup is needed,and the load produced is much higher than the business space window(evening).It needs to be adapted to this time.Data layout of inter related loadSecondly,in the application scenario of CDP,the multilevel grouping strategy is proposed,and the data layout of multilevel mapping is designed and implemented: MS-RAID.The data blocks on the same strip in the storage system are divided into multiple groups.The number of data blocks within the different levels of the disk group is different,which can provide multiple level access performance.The number of data blocks in the low performance disk group is small,and the energy consumption is low.When the system access is active,the performance requirements are higher,the high-performance disk group is started,the low performance disk group is in standby state in order to save energy.On the other hand,when the system access is not active,the storage system has low performance requirements and the low performance disk group is started to the standby state,which ensures the performance of the system.It also reduces energy consumption.And the DIP algorithm is used to optimize the writing performance,and the energy efficiency ratio of the storage system is realized with the energy saving scheduling of disk;Finally,experiments show that,when the 40%,60%,80% and 100% workload is sequential,6-disk two-level MS-RAID5 performance has been improved then 2-disk group S-RAID5 by 15.9%,29.0%,31.4% and 33.6%,saved 34.6% energy consumption than 3-disk S-RAID5.Proved that MS-RAID is a more energy-efficient data layout. |
PDF Full Text Request