Optimization for resource management in heterogeneous spark clusters

Huanhuan Fang; Huanle Xu; Jingjing Xu

doi:10.1117/12.2660979

16 December 2022 Optimization for resource management in heterogeneous spark clusters

Huanhuan Fang, Huanle Xu, Jingjing Xu

Proceedings Volume 12500, Fifth International Conference on Mechatronics and Computer Technology Engineering (MCTE 2022); 125005O (2022) https://doi.org/10.1117/12.2660979
Event: 5th International Conference on Mechatronics and Computer Technology Engineering (MCTE 2022), 2022, Chongqing, China

Abstract

Spark, as a large-scale data processing framework based on in-memory computing, is widely used in various big data processing scenarios. However, its existing resource scheduling strategy does not consider different types of applications and real-time utilization of cluster resources, resulting in low efficiency of both cluster resource usage and big data applications. To solve above problems, this paper proposes a strategy named Resource Scheduling for Heterogeneous Spark clusters (RSFHS). Firstly, the application is classified into CPU-intensive and IO-intensive according to the difference of resources consumed by the application; secondly, by monitoring the utilization of various resources of cluster nodes in real time, the remaining available resources of nodes are calculated, and then the nodes with strong processing capability and consistence with the type of applications are scheduled and allocated first. Experiments show that compared with Spark's existing resource scheduling strategy, RSFHS strategy can decrease the average execution time of applications by 6.5% ~12.26% and optimize the scheduling of different types of applications.

Citation Download Citation

Huanhuan Fang, Huanle Xu, and Jingjing Xu "Optimization for resource management in heterogeneous spark clusters", Proc. SPIE 12500, Fifth International Conference on Mechatronics and Computer Technology Engineering (MCTE 2022), 125005O (16 December 2022); https://doi.org/10.1117/12.2660979

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

PERSONAL
Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available