超百亿亿级系统面临的网络和通信挑战

Dhabaleswar PANDA; 陆晓义; Hari SUBRAMONI

doi:10.1631/FITEE.1800631

Your Location：

Home >

Browse articles >

超百亿亿级系统面临的网络和通信挑战

后E级超算专辑 | Updated：2022-05-19

- 超百亿亿级系统面临的网络和通信挑战
- Networking and communication challenges for post-exascale systems
- 信息与电子工程前沿（英文） 2018年19卷第10期页码：1230-1235
- Affiliations：
  
  Department of Computer Science and Engineering, The Ohio State University, Ohio 43210, USA
- Author bio：
  
  Dhabaleswar PANDA, E-mail: panda@cse.ohio-state.edu
  [ "Xiao-yi LU, E-mail: luxi@cse.ohio-state.edu" ]
  [ "Hari SUBRAMONI, E-mail: subramon@cse.ohio-state.edu" ]
- Funds：
  
  Project supported by the National Science Foundation of the USA (Nos. IIS-1447804 and CNS-1513120)
- DOI：10.1631/FITEE.1800631
  中图分类号： TP311
- 收稿：2018-10-09，
  
  修回：2018-10-15，
  
  纸质出版：2018-10
- Accepted：
Scan QR Code
Dhabaleswar PANDA, 陆晓义, Hari SUBRAMONI. 超百亿亿级系统面临的网络和通信挑战[J]. 信息与电子工程前沿（英文）, 2018,19(10):1230-1235.

Dhabaleswar PANDA, Xiao-yi LU, Hari SUBRAMONI. Networking and communication challenges for post-exascale systems[J]. Frontiers of Information Technology & Electronic Engineering, 2018, 19(10): 1230-1235.
Dhabaleswar PANDA, 陆晓义, Hari SUBRAMONI. 超百亿亿级系统面临的网络和通信挑战[J]. 信息与电子工程前沿（英文）, 2018,19(10):1230-1235. DOI： 10.1631/FITEE.1800631.

Dhabaleswar PANDA, Xiao-yi LU, Hari SUBRAMONI. Networking and communication challenges for post-exascale systems[J]. Frontiers of Information Technology & Electronic Engineering, 2018, 19(10): 1230-1235. DOI： 10.1631/FITEE.1800631.

摘要

由于新兴处理器、内存和网络技术的显著进步，百亿亿级系统将在未来几年（2020-2022）推出。随着百亿亿级系统被配置和使用，具有更细粒度、更短时间步长和更大数据量的下一代应用程序将被持续需求。从发展趋势看，2025-2035年间，下一代应用程序将需要超百亿亿级系统。本文关注超百亿亿级系统在网络和通信方面面临的挑战。首先，提出超百亿亿级系统的设想架构。其次，从不同方面阐述面临的挑战，包括多种网络技术、高性能通信和同步协议、加速器和现场可编程门户阵列的集成支持、容错和服务质量支持、能量感知通信方案和协议、软件定义网络以及多种内存和存储器的可扩展通信协议。再次，指出在这些系统上进行支持高性能计算、大数据和深度学习的高效编程模型设计面临的挑战。最后，强调了这些系统的上层共同设计运行时间的关键需求，以实现最优性能和可扩展性。

Abstract

With the significant advancement in emerging processor

memory

and networking technologies

exascale systems will become available in the next few years (2020-2022). As the exascale systems begin to be deployed and used

there will be a continuous demand to run next-generation applications with finer granularity

finer time-steps

and increased data sizes. Based on historical trends

next-generation applications will require post-exascale systems during 2025-2035. In this study

we focus on the networking and communication challenges for post-exascale systems. Firstly

we present an envisioned architecture for post-exascale systems. Secondly

the challenges are summarized from different perspectives: heterogeneous networking technologies

high-performance communication and synchronization protocols

integrated support with accelerators and field-programmable gate arrays

fault-tolerance and quality-of-service support

energy-aware communication schemes and protocols

software-defined networking

and scalable communication protocols with heterogeneous memory and storage. Thirdly

we present the challenges in designing efficient programming model support for high-performance computing

big data

and deep learning on these systems. Finally

we emphasize the critical need for co-designing runtime with upper layers on these systems to achieve the maximum performance and scalability.

关键词

Keywords

references

ASCAC Subcommittee on Exascale Computing . . The Opportunities and Challenges of Exascale Computing . . 2010 . . https://science.energy.gov/media/ascr/ascac/pdf/reports/Exascale_subcommittee_report.pdf https://science.energy.gov/media/ascr/ascac/pdf/reports/Exascale_subcommittee_report.pdf , , . .

R Biswas , , , XY Lu , , , DK Panda . . Accelerating tensorFlow with adaptive RDMA-based gRPC . . 25 th IEEE Int Conf on High Performance Computing, Data, and Analytic , , 2018 . . .

YF Cui , , , R Moore , , , K Olsen , , , 等 . . Enabling verylarge scale earthquake simulations on parallel machines . . In: Shi Y, van Albada GD, Dongarra J, et al. (Eds.), Computational Science. Springer Berlin Heidelberg , , 2007 . . p.46 - - 53 . . . .

Energy Government . . Workshop on Terabits Networks for Extreme Scale Science . . 2011 . . https://science.energy.gov//media/ascr/pdf/programdocuments/docs/Terabit_networks_workshop_report.pdf https://science.energy.gov//media/ascr/pdf/programdocuments/docs/Terabit_networks_workshop_report.pdf , , . .

RL Graham , , , D Bureddy , , , P Lui , , , 等 . . Scalable Hierarchical Aggregation Protocol (SHArP): a hardware architecture for efficient data reduction . . Proc 1 st Workshop on Optimization of Communication in HPC , , 2016 . . p.1 - - 10 . . DOI: 10.1109/COMHPC.2016.006 http://doi.org/10.1109/COMHPC.2016.006 . .

Intel . . Intel Omni-Path Architecture Driving Exascale Computing and HPC . . 2016 . . https://www.intel.com/content/www/us/en/highperformance-computing-fabrics/omni-path-drivingexascale-computing.html https://www.intel.com/content/www/us/en/highperformance-computing-fabrics/omni-path-drivingexascale-computing.html , , . .

RZ Li , , , C DeTar , , , S Gottlieb , , , 等 . . MILC code performance on high end CPU and GPU supercomputer clusters . . 2017 . . http://cn.arxiv.org/abs/1712.00143 http://cn.arxiv.org/abs/1712.00143 , , . .

XY Lu , , , D Shankar , , , S Gugnani , , , 等 . . Highperformance design of Apache Spark with RDMA and its benefits on various workloads . . Proc IEEE Int Conference on Big Data , , 2016 . . p.253 - - 262 . . DOI: 10.1109/BigData.2016.7840611 http://doi.org/10.1109/BigData.2016.7840611 . .

BlueField Mellanox . . Multicore System on Chip . . 2017 . . http://www.mellanox.com/related-docs/npu-multicoreprocessors/PB_Bluefield_SoC.pdf http://www.mellanox.com/related-docs/npu-multicoreprocessors/PB_Bluefield_SoC.pdf , , . .

NVMe Express . . NVMe over Fabrics . . 2016 . . http://www.nvmexpress.org/wp-content/uploads/NVMe_Over_Fabrics.pdf http://www.nvmexpress.org/wp-content/uploads/NVMe_Over_Fabrics.pdf , , . .

ORNL . . Summit: America’s Newest and Smartest Supercomputer . . 2018 . . https://www.olcf.ornl.gov/summit/ https://www.olcf.ornl.gov/summit/ , , . .

MWU Rahman , , , XY Lu , , , NS Islam , , , 等 . . HOMR: a hybrid approach to exploit maximum overlapping in MapReduce over high performance interconnects . . Proc 28 th ACM Int Conf on Supercomputing , , 2014 . . p.33 - - 42 . . DOI: 10.1145/2597652.2597684 http://doi.org/10.1145/2597652.2597684 . .

R Rajachandrasekar , , , J Jaswani , , , H Subramoni , , , 等 . . Minimizing network contention in InfiniBand clusters with a QoS-aware data-staging framework . . IEEE Int Conf on Cluster Computing , , 2012 . . p.329 - - 336 . . DOI: 10.1109/CLUSTER.2012.90 http://doi.org/10.1109/CLUSTER.2012.90 . .

AMK Sarvestani , , , C Bailey , , , J Austin . . Performance analysis of a 3D wireless massively parallel computer . . J Sens Actuat Netw , , 2018 . . 7 ( ( 2 ): ): 18 DOI: 10.3390/jsan7020018 http://doi.org/10.3390/jsan7020018 . .

D Shankar , , , X Lu , , , N Islam , , , 等 . . High-performance hybrid key-value store on modern clusters with RDMA interconnects and SSDs: non-blocking extensions, designs, and benefits . . IEEE Int Parallel and Distributed Processing Symp , , 2016 . . p.393 - - 402 . . DOI: 10.1109/IPDPS.2016.112 http://doi.org/10.1109/IPDPS.2016.112 . .

H Subramoni , , , P Lai , , , S Sur , , , 等 . . Improving application performance and predictability using multiple virtual lanes in modern multi-core InfiniBand clusters . . Int Conf on Parallel Processing , , 2010 . . .

浏览量

Downloads

CSCD

文章被引用时，请邮件提醒。

Submit

工具集

关联资源

MltAuxTSPP: a unified benchmark for deep learning-based traffic state prediction with multi-source auxiliary data

Q-space-coordinate-guided neural networks for high-fidelity diffusion tensor estimation from minimal diffusion-weighted images

CRGT-SA: an interlaced and spatiotemporal deep learning model for network intrusion detection

Federated model with contrastive learning and adaptive control variates for human activity recognition

A comprehensive survey of physical adversarial vulnerabilities in autonomous driving systems