蓝光光盘库缓存管理机制的研究与实现

发布时间：2018-01-21 22:51

本文关键词： 蓝光光盘库缓存替换算法命中率 DB-LRU算法　出处：《中山大学》2013年硕士论文　论文类型：学位论文

【摘要】：数据的存储与检索是现在热门的话题，许多社会生产活动都是围绕着数据的存储和检索展开的。光盘库作为主流的存储设备，随着蓝光技术的，由于其容量大、价格低、安全性高等优点，受到社会上越来越多人的关注，在医院、电子政务等许多领域都有应用。本文立足于蓝光光盘库管理系统这一项目，选取其缓存管理模块进行针对性的研究与实现。缓存系统是光盘库中的重要模块，它缓解了光盘库由于自身机械结构上带来响应时间上的瓶颈问题，让整个系统的效率得到了很大提升。本文分析了蓝光光盘库缓存的特点，从缓存命中率上做相关的研究与分析。在缓存命中率与缓存算法的开销上做了折衷，提出了基于光盘分组的最近最少访问（Disk-Based Least Recently Used）算法。DB-LRU算法是对传统的LRU算法的改进，它继承了传统LRU算法的快速与高效，同时契合了光盘库数据存储的特点，也在一定程度上解决了LRU算法存在的缓存污染的问题。DB-LRU算法利用在同一个光盘中的数据具有的关联性，，对缓存中的数据按其所在光盘分组后，按照LRU规则来淘汰数据。本文通过实验，比较FIFO算法、LRU算法与DB-LRU算法，在随机访问、有热点数据和存在突发性数据的三种访问模式下的命中率的情况，发现DB-LRU算法在命中率比其它二者更高，且在缓存污染问题方面都比LRU算法有所改进。 DB-LRU算法符合光盘库数据存储的特性，也符合它对蓝光光盘库大容量，这也为缓存管理模块的后续的改进打下了基础。
[Abstract]:Data storage and retrieval is a hot topic, many social production activities are around the data storage and retrieval. As the mainstream storage device, with the Blu-ray technology, due to its large capacity. Low price, high security and other advantages have attracted more and more people's attention, and have been applied in many fields, such as hospitals, e-government and so on. Based on the project of Blu-ray optical disk library management system, this paper selects its cache management module to carry out targeted research and implementation. The cache system is an important module in the optical disk library. It alleviates the bottleneck of the response time caused by the mechanical structure of the optical disk library and greatly improves the efficiency of the whole system. This paper analyzes the characteristics of Blu-ray optical disk library cache, makes relevant research and analysis from cache hit ratio, and makes a tradeoff between cache hit rate and cache algorithm overhead. In this paper, Disk-Based Least Recently used based on optical disk packet is proposed. DB-LRU algorithm is an improvement on the traditional LRU algorithm. It inherits the fast and high efficiency of the traditional LRU algorithm and accords with the characteristics of the data storage of the optical disk library. To some extent, it solves the problem of cache contamination in LRU algorithm. DB-LRU algorithm takes advantage of the correlation of data in the same disc, and groups the data in the cache according to its optical disk. According to the LRU rule to eliminate the data. This paper compares the FIFO algorithm with the DB-LRU algorithm in random access. In the case of hit rate under three access modes with hot data and unexpected data, it is found that the hit rate of DB-LRU algorithm is higher than that of the other two. And the problem of cache pollution is improved compared with LRU algorithm. The DB-LRU algorithm accords with the characteristics of data storage in optical disk library and the large capacity of Blu-ray optical disk library, which lays the foundation for the further improvement of cache management module.
【学位授予单位】：中山大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP333

【参考文献】