Science DMZ network architecture deployment and performance evaluation to large scaled data transfer efficiency

 
 
 
  • Abstract
  • Keywords
  • References
  • PDF
  • Abstract


    Background/Objectives: Recently, it has been great issue to transfer large-scale science dada such as scientific field of high energy physic, astronomical space, super-computing simulation. To solve the transfer issue and to increase transfer efficiency, it needs a multi-dimensional approaches.

    Methods/Statistical analysis: To improve the transfer performance, approaches from the perspective of components such as network equipment, transmission protocol, and transmission application have been suggested. Effort to TCP congestion control algorithm and parallelism of data transfer channel are representative example to improve performance. However, the solution through the each component has a limitation in maximizing the transmission efficiency.

    Findings: Science DMZ is a new network architecture that can maximize transfer performance. It maximizes transfer efficiency through approach to all components, such as network equipment, dedicated network path, transfer applications, and local institute firewall policies. With these complicated components, science DMZ network architecture can greatly improve the transfer efficiency. In this paper, we design and construct a science DMZ network architecture between two organizations that utilize supercomputing resources based on KREONET and evaluate the performance.

    Improvements/Applications: After configuring the experiment environment, we measured network performance through iperf and file transfer performance test through SCP. Experiment result showed around 388% Improvement than that of existing method.

     

     


  • Keywords


    Large-Scale Science Data; Transfer Performance; Science DMZ; TCP; UDP Based Transfer; Parallel Data Transfer.

  • References


      [1] C. Jayashri, P. Abitha, S. Subburaj, S.Yamuna Devi, S. Suthir, S. Janakiraman, Big Data Transfers through Dynamic and Load Balanced Flow on Cloud Networks, 2017 Third International Conference on AEEICB.

      [2] T. Kelly, Scalable TCP: improving performance in high speed wide area networks, Computer Communications Review, 2003, 32(2).

      [3] D. Kliazovich, F. Granelli, and D. Miorandi, Logarithmic Window Increase for TCP Westwood+ for Improvement in High Speed, Long Distance Networks, Computer Networks, 2008, 52(12), pp. 2395-2410.

      [4] M. A. Alrshah, M. Othman, Performance Evaluation of Parallel TCP, and Its Impact on Bandwidth Utilization and Fairness in High-BDP Networks based on Test-bed,” in Proc. 2013 malaysia international conference on MICC, pp. 23-28.

      [5] J. Park, D. An, G. Cho, an Adaptive Channel Number Tuning Mechanism on Parallel Transfer with UDT, 2013 International Conference on ICOIN.

      [6] Y. Gu, R. L. Grossman, UDT: UDP-based Data Transfer for High Speed Wide Area Networks, Computer Networks, 2007. 51(7), pp. 1777-1799.

      [7] M. Masirap, M. H. Amaran, Y. M. Yussoff, R. Rahman, H. Hashim, Evaluation of Reliable UDP-based Transport Protocols for Internet of Things (IoT), 2016 IEEE Symposium on ISCAIE.

      [8] E. Dart, L. Rotman, B. Tierney, the Science DMZ: a Network Design Pattern for Data-Intensive Science, Supercomputing Conference, 2013.

      [9] I. Monga, E. Pouyoul, C. Guok, Software-Defined Networking for Big-Data Science-Architectural Models from Campus to the WAN, 2012 SC Companion on SCC.

      [10] H. Neeman, TheOneOklahoma Friction Free Network: Towards a Multi-Institutional Science DMZ in an EPSCoR State, 2014 Annual Conference on ESEDE.


 

View

Download

Article ID: 13884
 
DOI: 10.14419/ijet.v7i2.33.13884




Copyright © 2012-2015 Science Publishing Corporation Inc. All rights reserved.