Lambda networking has been emerged to provide network service for data-intensive e-science by making use of high bandwidth offered by optical communication technology. Recent technological and cost breakthroughs in optical communication technology have made it possible to transmit a score of wavelengths on a single strand of optical fiber. A lambda, in networking terminology, is a fully dedicated wavelength of light supporting greater than 10 Gbps bandwidth. It is possible to guarantee ultimate quality of service (QoS) by assigning dedicated lambdas to each of e-science communities on a single fiber infrastructure. This lambda is desirable units of networking especially for e-science applications requiring high bandwidth and certain quality of service.
It, however, is hardly expected to provide a end-to-end lightpath interconnecting only optical lambdas to every e-science community. Therefore, the concept of lambda networking should not be confined to optical lambda (layer 1). Instead, it is required to be extended to include layer 2 and/or layer 3 in order to offer transparent end-to-end network service. Regarding this, a lightpath is defined as any channel or link where the end points and topology can be controlled. In addition, a lightpath might require allocating bandwidth according to service requirement.
In this paper, we study applicable technologies on layer 1, layer 2, and layer 3 to lambda networking. Since technologies for lambda networking on layer 1 and layer 2 are relatively well known, we particularly focus on the way how to extend lambda networking to include layer 3. A major challenge of establishing a lightpath on layer 3 is to uniquely identify a flow at the ingress and egress points to a routed network as well as how it is routed internally through a network. Regarding this challenge, we study the way how to identify flow and establish a specific path on layer 3 using two technologies, MPLS and virtual router. In addition, we stu...