CLAP: Component-Level Approximate Processing for Low Tail Latency and High Result Accuracy in Cloud Online Services

doi:10.1109/TPDS.2017.2650988

CORC > 计算技术研究所 > 中国科学院计算技术研究所 > 中国科学院计算技术研究所期刊论文 > 英文

	CLAP: Component-Level Approximate Processing for Low Tail Latency and High Result Accuracy in Cloud Online Services
	Han, Rui 2; Huang, Siguang 1; Wang, Zhentao 2; Zhan, Jianfeng 2
刊名	IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
	2017-08-01
卷号	28 期号:8 页码:2190-2203
关键词	Cloud online services tail latency result accuracy component-level approximate processing aggregated data points
ISSN号	1045-9219
DOI	10.1109/TPDS.2017.2650988
英文摘要	Modern latency-critical online services such as search engines often process requests by consulting large input data spanning massive parallel components. Hence the tail latency of these components determines the service latency. To trade off result accuracy for tail latency reduction, existing techniques use the components responding before a specified deadline to produce approximate results. However, they skip a large proportion of components when load gets heavier, thus incurring large accuracy losses. In this paper, we propose CLAP to enable component-level approximate processing of requests for low tail latency and small accuracy losses. CLAP aggregates information of input data to create small aggregated data points. Using these points, CLAP reduces latency variance of parallel components and allows them to produce initial results quickly; CLAP also identifies the parts of input data most related to requests' result accuracies, thus first using these parts to improve the produced results to minimize accuracy losses. We evaluated CLAP using real services and datasets. The results show: (i) CLAP reduces tail latency by 6.46 times with accuracy losses of 2.2 percent compared to existing exact processing techniques; (ii) when using the same latency, CLAP reduces accuracy losses by 31.58 times compared to existing approximate processing techniques.
资助项目	National Natural Science Foundation of China[61502451] ; National Key Research and Development Plan of China[2016YFB1000601]
WOS研究方向	Computer Science ; Engineering
语种	英语
出版者	IEEE COMPUTER SOC
WOS记录号	WOS:000405750200006
内容类型	期刊论文
源URL	[http://119.78.100.204/handle/2XEOYT63/7032]
专题	中国科学院计算技术研究所期刊论文_英文
通讯作者	Han, Rui
作者单位	1.Tsinghua Univ, Sch Software, Beijing 100084, Peoples R China 2.Chinese Acad Sci, Inst Comp Technol, Beijing 100864, Peoples R China
推荐引用方式 GB/T 7714	Han, Rui,Huang, Siguang,Wang, Zhentao,et al. CLAP: Component-Level Approximate Processing for Low Tail Latency and High Result Accuracy in Cloud Online Services[J]. IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS,2017,28(8):2190-2203.
APA	Han, Rui,Huang, Siguang,Wang, Zhentao,&Zhan, Jianfeng.(2017).CLAP: Component-Level Approximate Processing for Low Tail Latency and High Result Accuracy in Cloud Online Services.IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS,28(8),2190-2203.
MLA	Han, Rui,et al."CLAP: Component-Level Approximate Processing for Low Tail Latency and High Result Accuracy in Cloud Online Services".IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS 28.8(2017):2190-2203.