高能物理实验中的数据与计算技术

Data and computing for high energy physics experiments

高能物理也称作粒子物理,是研究物质微观结构与相互作用、宇宙起源的重要基础科学前沿分支.高能物理实验装置非常复杂,产生的实验数据规模巨大,从20世纪后期的实验每年产生TB级发展到现在最新的实验每年产生近百PB.过去数十年来高能物理一直对数据与计算技术提出挑战,同时也充当数据与计算技术发展的推动者.高能物理实验一方面在利用最新计算机技术实现对数据及计算的支撑,同时也在推动计算机技术发展与突破.本文首先介绍高能物理实验数据与计算的基本模式,列举典型的高能物理实验说明数据规模及数据处理的挑战.在此基础上介绍高能物理计算技术的发展过程,特别就当前主流的网格计算、云计算技术及数据共享进行了讨论.

High energy physics(HEP),which is also known as particle physics,is the study of the fundamental building blocks of nature,the interactions of elementary particles,and the origin of the Universe.Large-scale HEP experiments can produce hundreds of petabytes of real and simulated data yearly.This capability has made HEP a major explorer of scientific computing in the last few decades.In this work,three computing tasks,i.e.,event reconstruction,physics simulation,and data analysis,are introduced.Subsequently,data and computing challenges are demonstrated with different examples of HEP experiments.To address the challenges,the HEP community has been focusing on developing computing models in the last thirty years.The HEP-computing model has evolved from local centers with mainframes and workstations to distributed clusters and worldwide computing grids.Some technologies generated through HEP computing,such as WWW,have made a considerable impact on the world.In the concluding part of this work,the perspective of computing technologies in HEP is discussed.