CC*DNI Integration: Enhancing Science Through Custom Paths For Trusted Users

CC*DNI 集成：通过受信任用户的自定义路径增强科学

• 批准号: 1541426
• 负责人: James Griffioen
• 金额: $ 99.93万
• 依托单位: University of Kentucky Research Foundation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2015
• 资助国家: 美国
• 起止时间: 2015-09-01 至 2019-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1541426&HistoricalAwards=false
• 关键词: CC; DNI; Integration; Enhancing; Science

Progress in medicine, science, and engineering -- as in essentially all areas of human endeavor -- increasingly depends on network access to very large data sets (billions to trillions of bytes) containing information such as scientific measurements, monitoring information, images, videos, simulation results, etc. The number of researchers working with, transferring, and sharing very large data sets is growing at an alarming rate. Although network transmission speeds are rapidly increasing, devices deployed in the middle of the network for security and management purposes -- e.g., firewalls, intrusion detectors, and the like -- create major bottlenecks that prevent researchers from being able to quickly move these large files between laboratories, high-performance computing sites, and storage facilities.Historically, network designers have addressed this problem by creating and maintaining a special "Science DMZ" network that provided a "fast lane", bypassing performance-limiting security devices thanks to the trusted nature of the users and applications connected to the Science DMZ. Today, however, the growing number and variety of users who need high-performance data transfer makes it increasingly difficult to maintain such a separate infrastructure. This project is developing ways to provide trusted users on the normal (campus) network with dynamic access to such fast lanes by authenticating them on-the-fly, determining the nature of their data transfer, and whether it should be allowed or not.This project leverages emerging Software-Defined Networking (SDN) capabilities to single out individual data flows for specialized treatment in the network. Using the SDN infrastructure already deployed on campus, the project is developing new control software and protocols that enable trusted users to authenticate themselves to the network and set up network connections free from middlebox interference while still enforcing middlebox security functionality for normal campus traffic. The ability of trusted users to create such fast lanes -- as opposed to (only) network administrators -- will enable future networks to handle the rapidly growing number of big data users.

医学、科学和工程的进步--基本上与人类奋进的所有领域一样--越来越依赖于对包含科学测量、监测信息、图像、视频、模拟结果等信息的超大型数据集（数十亿至数万亿字节）的网络访问。 虽然网络传输速度正在迅速提高，但出于安全和管理目的而部署在网络中间的设备--例如，防火墙、入侵检测器等--造成了主要的瓶颈，使研究人员无法在实验室、高性能计算站点和存储设施之间快速移动这些大文件。历史上，网络设计人员通过创建和维护一个特殊的“科学DMZ”网络来解决这个问题，该网络提供了一条“快车道”，由于连接到科学DMZ的用户和应用程序的可信性，可以绕过性能限制安全设备。 然而，如今，需要高性能数据传输的用户数量和种类不断增加，使得维护这样一个独立的基础设施变得越来越困难。 该项目正在开发各种方法，为普通（园区）网络上的可信用户提供动态访问此类快速通道的能力，方法是在运行中对其进行身份验证，确定其数据传输的性质以及是否应该允许。该项目利用新兴的软件定义网络（SDN）功能，挑选出单个数据流，以便在网络中进行专门处理。 该项目正在利用已经部署在校园内的SDN基础设施开发新的控制软件和协议，使受信任的用户能够在网络上进行身份验证，并建立不受中间盒干扰的网络连接，同时仍然为正常的校园流量实施中间盒安全功能。 可信用户创建这种快速通道的能力-而不仅仅是网络管理员-将使未来的网络能够处理快速增长的大数据用户数量。

项目成果

Debugging SDN in HPC Environments

• DOI: 10.1145/3219104.3229277
• 发表时间: 2018-07
• 期刊: Proceedings of the Practice and Experience on Advanced Research Computing
• 作者: M. Hayashida;Sergio Rivera;James N. Griffioen;Zongming Fei;Yongwook Song

Dynamically Creating Custom SDN High-Speed Network Paths for Big Data Science Flows

为大数据科学流程动态创建自定义 SDN 高速网络路径

• DOI: 10.1145/3093338.3104155
• 发表时间: 2017
• 期刊: Practice & Experience in Advanced Research Computing Conference (PEARC 2017
• 作者: Rivera, Sergio;Hayashida, Mami;Griffioen, James;Fei, Zongming
• 通讯作者: Fei, Zongming

Navigating the Unexpected Realities of Big Data Transfers in a Cloud-based World

在基于云的世界中应对大数据传输的意外现实

• DOI: 10.1145/3219104.3229276
• 发表时间: 2018
• 期刊: PEARC '18 Proceedings of the Practice and Experience on Advanced Research Computing
• 作者: Rivera, Sergio;Griffioen, James;Fei, Zongming;Hayashida, Mami;Shi, Pinyi;Chitre, Bhushan;Chappell, Jacob;Song, Yongwook;Pike, Lowell;Carpenter, Charles
• 通讯作者: Carpenter, Charles