Connecting Method Research of Distributed Computing for AI Research Based on ScienceDMZ 


Vol. 46,  No. 6, pp. 1006-1022, Jun.  2021
10.7840/kics.2021.46.6.1006


PDF
  Abstract

Recently, the key areas leading ICT technology are by far big data and AI. In the field of science and technology, it is changing from a simulation-oriented research practice to a research environment using AI analysis techniques based on large-scale data. This change faced the problem of data transmission speed according to the change of data size and the problem of lack of computing environment to study AI. In addition, when the existing cloud computing technology is used by a large number of users, the data transmission delay and failure response problems occurred due to the rapid increase in data transmission volume. To this end, we solve the problem using computing technology in distributed processing environment. To this end, the problems are solved by using computing technology in distributed processing environment. Therefore, KREONET aims to improve the fast transmission of big data and the insufficient AI computing environment, centered around the establishment of the ScienceDMZ environment. In this paper, we aim to solve existing problems by building big data dedicated network based on ScienceDMZ and DTNs capable of distributed computing in cooperation with the eight institutions of super facility and develop a platform that provides container-based AI research environment on the established ScienceDMZ infrastructure. Through the case of establishing the ScienceDMZ environment in Korea, the result of performance improvement of more than 80% compared to the bandwidth is shown, and the superiority of distributed computing performance is shown by comparing the performance of single, parallel, and distributed computing using the CPU and GPU computing of the AI platform.

  Statistics
Cumulative Counts from November, 2022
Multiple requests among the same browser session are counted as one view. If you mouse over a chart, the values of data points will be shown.


  Cite this article

[IEEE Style]

K. Kim, J. Moon, W. Kwon, B. Park, W. Seok, W. Hong, S. Lee, J. Jo, T. Yoon, J. Chio, D. Kim, D. Hwang, S. Choi, J. Kim, J. Kim, K. Kim, B. Chong, D. Lee, Y. Yu, E. Park, J. Cheon, "Connecting Method Research of Distributed Computing for AI Research Based on ScienceDMZ," The Journal of Korean Institute of Communications and Information Sciences, vol. 46, no. 6, pp. 1006-1022, 2021. DOI: 10.7840/kics.2021.46.6.1006.

[ACM Style]

Ki-Hyeon Kim, Junghoon Moon, Woochang Kwon, Byungyeon Park, Woojin Seok, Won-taek Hong, Sang-kwon Lee, Jin-Yong Jo, Taejin Yoon, Jaehein Chio, DaeKyeom Kim, Dongah Hwang, SunWoong Choi, Jongho Kim, Junyeop Kim, KiHyoung Kim, Byonghoon Chong, Dosub Lee, Young-Geun Yu, Eun-Sook Park, and Jae-Hong Cheon. 2021. Connecting Method Research of Distributed Computing for AI Research Based on ScienceDMZ. The Journal of Korean Institute of Communications and Information Sciences, 46, 6, (2021), 1006-1022. DOI: 10.7840/kics.2021.46.6.1006.

[KICS Style]

Ki-Hyeon Kim, Junghoon Moon, Woochang Kwon, Byungyeon Park, Woojin Seok, Won-taek Hong, Sang-kwon Lee, Jin-Yong Jo, Taejin Yoon, Jaehein Chio, DaeKyeom Kim, Dongah Hwang, SunWoong Choi, Jongho Kim, Junyeop Kim, KiHyoung Kim, Byonghoon Chong, Dosub Lee, Young-Geun Yu, Eun-Sook Park, Jae-Hong Cheon, "Connecting Method Research of Distributed Computing for AI Research Based on ScienceDMZ," The Journal of Korean Institute of Communications and Information Sciences, vol. 46, no. 6, pp. 1006-1022, 6. 2021. (https://doi.org/10.7840/kics.2021.46.6.1006)

Search
(IN TITLE, AUTHOR, ABSTRACT,KEYWORDS)

Advanced Search ▼

POPULAR KEYWORDS
(TOP 10 KEYWORDS)

Recent Publications
(LAST 3 YEARS)






pISSN : 1226 - 4717
eISSN : 2287 - 3880