.png){kind=icon}
返回
Evaluating Task-Level CPU Efficiency for Distributed Stream Processing Systems
.png){kind=icon}
OA
AI
DOI：10.3390/bdcc7010049.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Big Data and primarily distributed stream processing systems (DSPSs) are growing in complexity and scale. As a result, effective performance management to ensure that these systems meet the required service level objectives (SLOs) is becoming increasingly difficult. A key factor to consider when evaluating the performance of a DSPS is CPU efficiency, which is the ratio of the workload processed by the system to the CPU resources invested. In this paper, we argue that developing new performance tools for creating DSPSs that can fulfill SLOs while using minimal resources is crucial. This is especially significant in edge computing situations where resources are limited and in large cloud deployments where conserving power and reducing computing expenses are essential. To address this challenge, we present a novel task-level approach for measuring CPU efficiency in DSPSs. Our approach supports various streaming frameworks, is adaptable, and comes with minimal overheads. This enables developers to understand the efficiency of different DSPSs at a granular level and provides insights that were not previously possible.
Keyword:
CPU efficiency
big data
distributed stream processing
performance
task-level measurement
profiling
flink
spark
期刊
B
IF:
4.4
论文数:
1.3K
被引数:
2.4K
机构
引用论文
The impact of wader predation on benthic macrofauna in Merja Zerga lagoon, Morocco: an exclosure experiment.png)
Wader Study
Wader StudyIF0
PREAI
分享收藏Kinetics of the oxidation of Cu(phen)2+ and Cu(bpy)2+ by aminopolycarboxylatocobalt(III) complexes
PREAI
分享收藏Influencing Factors in the Scalability of Distributed Stream Processing Jobs分布式流处理作业可伸缩性的影响因素IEEE ACCESS
OAAI
IEEE ACCESSIF3.6
OAAI分享收藏分享收藏
查看更多内容