Finding WDM Network Topographies that are Nonblocking without Wavelength Interchange

University Of Pittsburgh, Pittsburgh PA

Investigators

Abstract

Modern long-haul optical networks are capable of carrying a wavelength of light (a lambda or an optical circuit) and its associated bit stream thousands of kilometers without regeneration, conversion from an optical signal to an electrical one and back again (OEO). This long-haul technology has substantially reduced the cost of optical systems since OEO regeneration has traditionally been expensive. It has also offered the (as yet unrealized) promise of running multiple data formats. However, when two lambdas of the same wavelength need to be inserted onto the same optical fiber at least one of them must undergo OEO in order to change its wavelength. This is called wavelength interchange. The problem of routing and wavelength assignment (RWA) to minimize the number of wavelength interchanges has been deeply studied. This project will study a different question: what types of topologies minimize the need for wavelength interchange? This question is expected to lead to graph properties that minimize the need for wavelength interchange in physical networks. By taking a different approach to understanding the minimization of wavelength interchange the project couples the RWA problem to issues of network design. In turn this allows questions of network resilience to be coupled to questions of optical-layer design. Long-haul optical networks are a foundational element of the modern economy. Both network cost minimization and network flexibility and resilience are concerns affecting our national interests.