大规模协同计算平台下缓存层的研究与优化

发布时间：2017-12-27 20:41

本文关键词：大规模协同计算平台下缓存层的研究与优化　出处：《西安电子科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：Redis在云计算领域内得到了广泛的应用,可以用来构建一个高性能、大规模分布式缓存系统。Redis缓存系统将所有数据以键值对的形式存储于内存中,以减少访问关系型数据库的次数,提高读写效率。然而,在大规模协同云计算平台的实际应用中,Redis大规模分布式缓存层有会出现吞吐量急剧降低的问题,影响了平台服务的性能与稳定性,成为目前亟待解决的问题。首先,论文对Redis勺内部内存、持久化、事件处理、集群模式的机制与工作原理进行详细介绍,并将Redis与Memcached进行全方位的对比分析,说明使用Redis构造大规模缓存层的原因。然后,论文对云平台实际应用中出现的Redis缓存层吞吐量急剧降低问题进行分析。针对吞吐量急剧降低阶段发生的必要条件,缩小问题的分析范围,定位到导致吞吐量急剧增大运行流程中,所涉及的数据存储细节与算法流程。之后对其进行详细分析,以定位缓存层吞吐量下降的原因。针对导致吞吐量急剧增大的算法细节,论文提出改进方案。以使得Redis服务器在大规模协同云计算平台下构建的缓存层,可以提供良好稳定的服务。最后,论文对本文提出的改进方案进行了对比实验。实验表明,本文提出的针对Redis的改进,很好地解决了Redis在本文应用场景下的吞吐量急剧降低问题。
[Abstract]:Redis has been widely used in the field of cloud computing, which can be used to build a high-performance, large-scale distributed caching system. The Redis caching system stores all data in memory in the form of key value pairs to reduce the number of access relational databases and improve the efficiency of reading and writing. However, in the practical application of large-scale collaborative cloud computing platform, the large scale distributed cache layer of Redis will cause a sharp decline in throughput, which will affect the performance and stability of platform services, which has become an urgent problem at present. First of all, the paper introduces the Redis spoon's internal memory, persistence, event handling, clustering mode and the working principle, and makes a comprehensive comparative analysis between Redis and Memcached, and explains the reason of using Redis to build massive cache layer. Then, the paper analyzes the problem of the sharp reduction in the throughput of Redis cache layer in the actual application of the cloud platform. Aiming at the necessary conditions for the rapid reduction of throughput, we narrow the scope of the problem analysis and locate the data storage details and algorithm flow in the operation process, which results in a sharp increase in throughput. It is then analyzed in detail to locate the cause of the decline in the throughput of the cache layer. In view of the algorithm details that lead to a sharp increase in throughput, an improved scheme is proposed in this paper. The caching layer that enables the Redis server to build under a large-scale collaborative cloud computing platform can provide a good and stable service. Finally, the paper makes a comparative experiment on the improved scheme proposed in this paper. The experiment shows that the improvement proposed in this paper is a good solution to the problem of the sharp decrease in throughput of Redis in this application scenario.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TP333

【相似文献】