Journal of Computer Science and Technology

   

SOCA-DOM: A Mobile System-on-Chip Array System for Analyzing Big Data on the Move

Le-Le Li (李乐乐)1,2, Jiang-Yi Liu (刘江佾)1, Jian-Ping Fan (樊建平)3, Member, IEEE, Xue-Hai Qian (钱学海)4, Member, IEEE, Kai Hwang (黄铠)5, Fellow, IEEE, Yeh-Ching Chung (钟叶青)5, Senior Member, IEEE, and Zhi-Bin Yu (喻之斌)1, Member, IEEE   

  1. 1Center for Heterogeneous and Intelligent Computing, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzen 518055, China
    2School of Science and Engineering, The Chinese University of Hong Kong, Shenzhen, Shenzhen 518172, China
    3Center for High Performance Computing, Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences, Shenzen 518055, China
    4The Ming Hsieh Department of Electrical Engineering, The Department of Computer Science, University of Southern California, Los Angeles, CA 90089-0001, USA
    5School of Data and Science, The Chinese University of Hong Kong, Shenzhen, Shenzhen 518172, China
  • Received:2020-10-20 Revised:2022-02-21 Accepted:2022-04-25
  • Contact: Zhi-Bin Yu E-mail:zb.yu@siat.ac.cn
  • About author:Zhi-Bin Yu received his Ph.D. degree in computer science from the Huazhong University of Science and Technology, Wuhan, in 2008. He is a professor at the Shenzhen Institute of Advanced Technology, Chinese Academy of Science. His research interests include microarchitecture simulation, computer architecture, performance evaluation, and big data processing. He is a member of IEEE.

Recently, analyzing big data on the move is booming. It requires that the hardware resource should be low volume, low power, light in weight, high performance, and highly scalable whereas the management software should be flexible and consume little hardware resource. To meet these requirements, we present a system named SOCA-DOM that encompasses a mobile system-on-chip array architecture and a two-tier ``software-defined'' resource manager named Chameleon. First, we design an Ethernet communication board to support an array of mobile system-on-chips. Second, we propose a two-tier software architecture for Chameleon to make it flexible. Third, we devise data, configuration, and control planes for Chameleon to make it "software-defined'' and in turn consume hardware resource on demand. Fourth, we design an accurate synthetic metric that represents the computational power of a computing node. We employ 12 Apache Spark benchmarks to evaluate SOCA-DOM. Surprisingly, SOCA-DOM consumes up to 9.4x less CPU resource and 13.5x less memory than Mesos which is an existing resource manger. In addition, we show that a 16-node SOCA-DOM consumes up to 4x less energy than two standard Xeon servers. Based on the results, we conclude that an array architecture with fine-grained hardware resources and a software-defined resource manager works well for analyzing big data on the move.


中文摘要

1、研究背景(context):
近年来,各类边缘计算的应用场景使得对移动式大数据分析计算平台的需求激增,譬如在航天飞行和无人机实时视频处理等领域。但现有的边缘计算平台无法满足它们提出的6大新需求,包括硬件架构的体积小、功耗低、重量轻、性能高和扩展性强以及软件部分的高灵活性和低资源消耗。
2、目的(Objective):
本文研究旨在研究一种新的移动式大数据处理系统,包括硬件架构和资源管理软件两部分。在硬件上,我们的新架构相比现有的服务器系体积更小,功耗更少和扩展更灵活;在软件上,该资源管理系统能够运行大数据分析应用的同时,消耗更少的硬件资源。
3、方法(Method):
我们提出了一种针对大数据分析的移动式SoC集群系统SOCA-DOM,它由可移动的SoC集群(硬件)和基于软件定义的资源管理系统Chameleon(软件)两部分组成。移动式SoC集群主要由若干个轻巧的SoC节点集成在一块通信板子上构成。若干个通信板子可以灵活搭建成具备高性能和低功耗的便携式集群;软件层面上,Chameleon通过采用两层的控制架构增强ADOM系统的扩展性,并设计了三个平面(控制、配置和数据)的技术使得系统按需使用资源。
4、结果(Result & Findings):
12个Spark大数据标准测试程序能够在SOCA-DOM上运行成功。同时,与具有代表性的工业界资源管理器Mesos对比,SOCA-DOM使用的CPU资源高达Mesos的9.4倍,内存多达Mesos的13.5倍。此外,在能耗方面,16个节点的SOCA-DOM的能耗高达两台标准Xeon服务器的4倍。
5、结论(Conclusions):
为满足移动式大数据分析计算平台的六大需求,我们设计了一种新的移动式SoC集群系统SOCA-DOM。它由移动式SoC集群(硬件)和于软件定义的资源管理系统Chameleon(软件)组成。实验表明,SOCA-DOM能够运行典型的Spark大数据分析应用,并达到了体积小、功耗低、重量轻、性能高、扩展性强和低资源消耗的效果。


Key words: edge computing; mobile architecture; resource management; big data analytics; software-defined systems;

[1] Yue-Wen Wu, Yuan-Jia Xu, Heng Wu, Lin-Gang Su, Wen-Bo Zhang, Hua Zhong. Apollo: Rapidly Picking the Optimal Cloud Configurations for Big Data Analytics Using a Data-Driven Approach [J]. Journal of Computer Science and Technology, 2021, 36(5): 1184-1199.
[2] Sa Wang, Yan-Hai Zhu, Shan-Pei Chen, Tian-Ze Wu, Wen-Jie Li, Xu-Sheng Zhan, Hai-Yang Ding, Wei-Song Shi, Yun-Gang Bao. A Case for Adaptive Resource Management in Alibaba Datacenter Using Neural Networks [J]. Journal of Computer Science and Technology, 2020, 35(1): 209-220.
[3] Leo Mendiboure, Mohamed-Aymen Chalouf, Francine Krief. Edge Computing Based Applications in Vehicular Environments: Comparative Study and Main Issues [J]. Journal of Computer Science and Technology, 2019, 34(4): 869-886.
[4] Yan Li (李 研), Feng-Hong Chen (陈峰宏), Xi Sun (孙 熙), Ming-Hui Zhou (周明辉), Member, CCF| Wen-Pin Jiao (焦文品), Senior Member, CCF, Dong-Gang Cao (曹东刚) and Hong Mei (梅 宏), Senior Member, CCF. Self-Adaptive Resource Management for Large-Scale Shared Clusters [J]. , 2010, 25(5): 945-957.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!

ISSN 1000-9000(Print)

         1860-4749(Online)
CN 11-2296/TP

Home
Editorial Board
Author Guidelines
Subscription
Journal of Computer Science and Technology
Institute of Computing Technology, Chinese Academy of Sciences
P.O. Box 2704, Beijing 100190 P.R. China
Tel.:86-10-62610746
E-mail: jcst@ict.ac.cn
 
  Copyright ©2015 JCST, All Rights Reserved