一种RAID-5快速扩容技术的研究
发布时间:2018-01-10 07:06
本文关键词:一种RAID-5快速扩容技术的研究 出处:《华中科技大学》2013年硕士论文 论文类型:学位论文
【摘要】:磁盘阵列技术通过数据分条技术来提高存储系统的吞吐率,通过数据冗余技术来提高存储系统的可靠性。当一个磁盘阵列建立好之后,随着用户的使用,可用的空间将会越来越小,这个时候需要进行磁盘阵列的扩容。在扩容的过程中,如果重新建立一个更大的磁盘阵列,将会要数据整体迁移,并且迁移的过程中,原来的磁盘阵列是不能被用户使用的,也就是说存在一个使用的空窗,对于用户来说是很不方便的。 在使用的过程中,同时向磁盘阵列中加入新的磁盘,并且形成一个新的磁盘阵列,也就是在线扩容方式。在线扩容方式,提供了一种不存在用户使用空窗的扩容方式,在扩容的过程中,并不影响用户的正常读写请求,并能保证用户读写的正确性。在研究了现有在线扩容算法的基础上,,提出了一种新的扩容算法,达到了理论上的数据迁移量最小值,并且在扩容之后,打破了标准的磁盘阵列数据分布方式,采用一种新的数据分布方式,保证数据迁移最小值的前提下,并不破坏磁盘阵列的可靠性和高性能。 在数据迁移过程中,引入了一种速度控制机制,保证了在线扩容过程的合理性。当用户读写请求较多的时候,适当降低扩容的速度,能较快响应来自用户的请求;当用户读写请求比较少的时候,加快数据迁移速度,使扩容能快一点完成。同时,在扩容过程中,维护一个扩容窗口,主要是为了保证用户读写的正确性。根据用户读写请求位置与扩容窗口的关系,采用不同方式来处理,保证了扩容之后数据正确性。 对实现的扩容算法进行了离线和在线两种方式的测试。测试结果表明,新的扩容算法在离线扩容情况下,扩容时间比标准方式加快了很多。在线扩容的情况下,速度控制机制较好的发挥了作用,用户的读写响应时间比标准方式下要好,同时扩容时间也比标准的要好。对于扩容之后,也进行了读写测试,测试结果表明扩容之后,新的数据分布方式的读写性能相比标准的并没有降低。
[Abstract]:Disk array technology improves the throughput of storage system by data striping technology, and improves the reliability of storage system by data redundancy technology. When a disk array is set up, with the use of users. The available space will be smaller and smaller, and this time the disk array will need to be expanded. During the expansion process, if a larger array of disks is re-established, the data will have to be migrated as a whole. In the process of migration, the original disk array can not be used by the user, that is to say, there is an empty window, which is very inconvenient for the user. In the process of using, a new disk array is added to the disk array at the same time, and a new disk array is formed, that is, the online expansion mode. The invention provides a capacity expansion mode in which there is no user using empty windows, and the normal read and write request of the user is not affected during the expansion process. Based on the research of the existing online capacity expansion algorithm, a new expansion algorithm is proposed, which can achieve the minimum data migration in theory, and after the expansion. It breaks the standard data distribution mode of disk array and adopts a new data distribution mode to ensure the minimum data migration without destroying the reliability and high performance of the disk array. In the process of data migration, a speed control mechanism is introduced to ensure the rationality of the online expansion process. Can quickly respond to requests from users; When the user read and write requests are few, accelerate the speed of data migration, so that the expansion can be completed a bit faster. At the same time, in the process of expansion, maintain an expansion window. According to the relationship between the location of the user's read-write request and the expansion window, different ways are adopted to process the data, which ensures the correctness of the data after the expansion. The test results show that the new algorithm is in the case of off-line expansion. The expansion time is much faster than the standard mode. In the case of online capacity expansion, the speed control mechanism plays a better role, and the response time of the user read and write is better than that under the standard mode. At the same time, the expansion time is also better than the standard. After the expansion, the read-write test is also carried out. The test results show that the performance of the new data distribution mode is not lower than the standard after the expansion.
【学位授予单位】:华中科技大学
【学位级别】:硕士
【学位授予年份】:2013
【分类号】:TP333
【参考文献】
相关期刊论文 前5条
1 陈凯,白英彩;网络存储技术及发展趋势[J];电子学报;2002年S1期
2 谢长生,王宇德,曹强;进化存储系统的研究[J];计算机工程;2004年22期
3 韩德志,谢长生,李怀阳;存储备份技术探析[J];计算机应用研究;2004年06期
4 金海,张江陵,周欣荣;磁盘阵列技术及其发展趋势[J];微处理机;1995年02期
5 谢长生;董晓明;万继光;谭志虎;刘瑞芳;;磁盘阵列控制器的设计与原型实现[J];小型微型计算机系统;2006年01期
本文编号:1404302
本文链接:https://www.wllwen.com/kejilunwen/jisuanjikexuelunwen/1404302.html