摘要:
为解决TCP/IP面临的可扩展性、动态性和安全可控性等问题，研究人员提出了大量的新型网络体系架构、协议和算法等，而这些研究在部署与应用前需要在一定规模的、逼近于真实网络的试验床上进行长时间的测试、评估和优化.近年来,国内外兴起了网络创新试验床的研究和建设.然而,网络创新试验床在设计与实现中面临一系列挑战.以实验所面临的成本、可行性、可信性和可控性4个问题为出发点，总结了网络创新试验床的需求及特性要求,归纳分析了其在虚拟化、网络可编程性、联邦管控、实验控制、测量与监测等方面存在的技术挑战,并对关键技术、架构和研究进展进行介绍与论述.最后分析了国内外网络创新试验床的建设与发展现状,并讨论了网络试验床未来的发展趋势和需要进一步解决的问题.
Abstract:
The current Internet architecture based on TCP/IP is facing with many unprecedented challenges, including scalability, security, mobility and controllability. New clean-slate architecture designs are expected to address these challenges and provide better evolvability. As such, they have been attracting great attention in recent years. Before deployment in production networks, the future Internet architectures, protocols and algorithms should be comprehensively validated, evaluated and optimized in large-scale and realistic testbeds. The testbeds for network innovation should closely simulate the real network, and provide more flexibility. These requirements make it critical to study architectures and key technologies of testbeds. Although there have been long-term interests in network testbeds, a comprehensive survey is still missing. In this paper, we first analyze the requirements of network innovation testbeds based on four common problems of network experiments, namely cost, feasibility, credibility and controllability. We then summarize technical challenges of testbeds design and development from the perspectives of virtualization and programmability of network, federate management/control of resources, as well as monitor/measurement of infrastructures and experiments. In particular, we summarize the state-of-the-art technologies and architectures that aim at addressing the aforementioned challenges. Finally, we present some representative testbeds, and discuss the future trends and open questions.
百度学术
下载:
