计算机科学 ›› 2012, Vol. 39 ›› Issue (Z6): 93-95.

• • 上一篇    下一篇

高能物理实验中数据传输系统的研究与实现

曾珊,齐法制,王萌   

  1. (中国科学院高能物理研究所计算中心 北京100049);(山东大学物理学院 济南250100)
  • 出版日期:2018-11-16 发布日期:2018-11-16

Study and Implementation of Data Transfer System in Experiment of High Energy Physics

  • Online:2018-11-16 Published:2018-11-16

摘要: 高能物理实验每天会产生大量的实验数据,由于高能物理实验本身的跨地域的建设特性,这些实验数据需要 传输到远程的数据和计算中心进行离线分析。如何将这些数据实时、可靠、高效地传输到远程的数据和计算中心则是 目前高能物理实验中需要解决的一个重要问题。介绍了一种高能物理环境下支持大批量数据传输的实时系统。系统 提供了与数据产生系统和数据管理系统的接口,并实现实验数据从实验现场到数据中心和计算中心的可靠传输。目 前该系统具有的功能包括:多路径源数据扫描、数据传输、数据缓冲区自动释放、传输过程管理、数据传输过程和性能 监视、传输过程日志记录等。为了解决广域网上高延迟的特性,提高网络传输效率,系统应支持多流并发传输,同时支 持数据中继服务,从而解决传输过程中由于网络或者某一传输节点失效造成的单点故障问题,提高系统自身的健壮性 和可靠性。实验表明,该系统在高能物理实验数据的传输过程中具有良好的效果。

关键词: 大批量数据,实时传输,可靠传输,多流并发,数据中继

Abstract: In the circumstance of high energy physics experiments, usually a large amount of experimental data is pro- duced every day. Also the high energy physics experiments are regional-crossing. Because of this feature, the experimen- tal data needs to be transferred to the remote data center and the computing center for data analysis. So how to transfer the data to the remote data center and computing center reliably and efficiently in real time is an important problem which needs to be solved for carrying on a high energy physics experiment successfully. This paper presented a mass da- to real-time transfer system under the environment of high energy physics. I}his system provides the interfaces with the data generation system and the data management system, and can transmit the experiment data from onsite to the re- mote data center and computing center reliably. I}he functionalities of the system include; multipath source data scan- ning, data transmitting, data buffer automatic releasing, data management interface, data transmitting performance moni- toring and logging and etc. I}he system also provides multi-flow concurrent transmission in order to solve the high delay in WAN and improve the transmission efficiency. In addition, to solve the single point problems caused by the network status or failure of the transmission point, the system also supports data relaying to make the system work robustly and reliably. Experiments show that the system performs well in data transmission process under the environment of high energy physics.

Key words: Mass data, Real-time transmission, Reliable transmission, Multi-flow concurrency, Data relay

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!