文章基本信息

标题：Simulating and Evaluating a Real-World ElasticSearch System Using the RECAP DES Simulator
本地全文：下载
作者：Malika Bendechache ; Sergej Svorobej ; Patricia Takako Endo 等
期刊名称：Future Internet
电子版ISSN：1999-5903
出版年度：2021
卷号：13
期号：4
页码：83
DOI：10.3390/fi13040083
出版社：MDPI Publishing
摘要：Simulation has become an indispensable technique for modelling and evaluating the performance of large-scale systems efficiently and at a relatively low cost. ElasticSearch (ES) is one of the most popular open source large-scale distributed data indexing systems worldwide. In this paper, we use the RECAP Discrete Event Simulator (DES) simulator, an extension of CloudSimPlus, to model and evaluate the performance of a real-world cloud-based ES deployment by an Irish small and medium-sized enterprise (SME), Opening.io. Following simulation experiments that explored how much query traffic the existing Opening.io architecture could cater for before performance degradation, a revised architecture was proposed, adding a new virtual machine in order to dissolve the bottleneck. The simulation results suggest that the proposed improved architecture can handle significantly larger query traffic (about 71% more) than the current architecture used by Opening.io. The results also suggest that the RECAP DES simulator is suitable for simulating ES systems and can help companies to understand their infrastructure bottlenecks under various traffic scenarios and inform optimisation and scalability decisions.
关键词：simulation; modelling; ElasticSearch; DES; CloudSim; CloudSimPlus; query; workload; search engines simulation ; modelling ; ElasticSearch ; DES ; CloudSim ; CloudSimPlus ; query ; workload ; search engines