## Abstract

Extreme-scale transmissions exhibit schedules dependent upon the equipment used to generate data, availability of the collaborators who will investigate and analyze the data, and the availability of storage and supercomputing resources, which must be dedicated to distribution, backup, computation, and visualization of data. Given such a wide array of dependencies, these applications, which are often scientific in nature, offer the luxury of in-advance reservation of network resources prior to the demand uptimes. Advance reservation (AR) allows for efficient scheduling around or in cooperation with competing reservations from other applications. This work focuses on static inputs to two classifications of AR demands: with and without sliding-window flexibility. We further introduce the ability for a demand to be assigned to more than one dedicated wavelength at different points throughout its lifetime depending on availability. This $\lambda $-switching functionality is incorporated at time-slot granularity, which limits the overhead associated with reconfiguring wavelength resources at runtime and enhances opportunities for scheduling efficiency. A novel $\lambda $-switching integer linear program is presented to evaluate this enhancement under varying degrees of temporal flexibility. Detailed and novel proofs are provided, which demonstrate that AR scheduling with and without $\lambda $-switching is in the complexity class $\mathcal{NP}$-complete. Conservative heuristics are presented to harness the benefits of $\lambda $-switching while lowering the trade-off from transceiver reconfiguration overhead. Lower bounds also are developed as a baseline comparison for these heuristics, and we evaluate these solutions through extensive simulation. We also introduce the notion of considering time-slotted, wavelength-routed networks from the perspective of a finer resource granularity represented by the intersection of the temporal and spectral resource domains.

© 2016 Optical Society of America

Full Article | PDF Article**OSA Recommended Articles**

Jeremy M. Plante and Vinod M. Vokkarane

J. Opt. Commun. Netw. **9**(11) 945-959 (2017)

Pegah Afsharlar, Arash Deylamsalehi, Jeremy M. Plante, Juzi Zhao, and Vinod M. Vokkarane

J. Opt. Commun. Netw. **9**(3) B101-B111 (2017)

Arush Gadkar, Tim Entel, Jeremy M. Plante, and Vinod M. Vokkarane

J. Opt. Commun. Netw. **6**(3) 340-354 (2014)