GMPLS for Future Applications: Performance Characterization and Measurements

GMPLS for Future Applications: Performance Characterization and Measurements

Weiqiang Sun (Shanghai Jiao Tong University, China), Wei Guo (Shanghai Jiao Tong University, China), Yaohui Jin (Shanghai Jiao Tong University, China), Lilin Yi (Shanghai Jiao Tong University, China) and Weisheng Hu (Shanghai Jiao Tong University, China)
DOI: 10.4018/978-1-4666-3652-1.ch005
OnDemand PDF Download:
No Current Special Offers


Generalized Multiprotocol Label Switching, or GMPLS, is a suite of protocols to enable automated resource discovery, automated service provisioning and automated failure recovery. In recent years, a considerable number of efforts have been seen in the area of putting GMPLS into advanced networking/service environments. This is exemplified by the various research programs in the US, Europe, and Asia. In such programs, GMPLS has not only been used as a way to reduce management complexity and increase reliability, like the industry is doing right now, but also it is used as a new way for service provisioning. In this chapter, the authors first review activities in using GMPLS controlled optical networks in high performance computing environments. They try to identify the benefits, as well as the limitations in such networking practices. Then they introduce the past and on-going standardization work in the Internet Engineering Task Force (IETF) about GMPLS network performance characterization and measurement. Finally, the authors present the performance measurement results from a number of deployed GMPLS networks.
Chapter Preview

Applying Gmpls In High Performance Computing Environments

Distributed storage, high performance computing and next generation e-science applications have long been research interests of the networking and computing communities. Such applications generally require large volumes of data be transferred from one place to another, or a set of steering and control operations from a centralized node be distributed to visualization/computing nodes in a timely manner. Although researches in grid computing has made tremendous advances in connecting widely distributed resources using ubiquitous Internet infrastructure, the fact that Internet is a shared packet switched network and is thus unable to provide the required bandwidth or QoS guarantee has intrigued much interests in building dedicated networks for that purpose. Circuit switched optical networks, because of its huge bandwidth and the guaranteed QoS performance, are regarded as excellent transport infrastructures for such applications.

Given the dynamic and heterogeneous nature of applications, the following requirements make the problem of provisioning circuits to applications even more challenging:

  • Meeting the arbitrary communication needs of applications, while at the same maintaining a high level of efficiency.

  • Providing an user-friendly interface, such that no or little additional complexity incur in application design.

  • Providing optimized performance for a variety of applications with different requirements.

Complete Chapter List

Search this Book: