Neural Best Fit Void Filling Scheduler in fixed time for optical burst switching

Abstract

Optical Burst Switching (OBS), which works only with optical signal processing, is the next generation hopeful technology for Exabyte optical transport networks. Yet, there are still some issues that need to be addressed such as burst assembling, switching, scheduling, contention resolution and quality of service. Indeed, one of the major problems is to schedule efficiently bursts on wavelength channels without buffers, converters, or other additional equipment. In this paper, we propose the Neural Best Fit Void Filling Scheduler (NBFVFS) for optical burst switching, which is easy to implement using Adjustable MAXNET (AMAXNET) and runs in fixed time. This neural scheduler will contribute to this new emerging solution by providing a parallel, fast, flexible, adaptive, and intelligent process. In comparison with the existing schedulers, the proposed NBFVFS is more efficient both in terms of bandwidth usage as well as in terms of processing speed. NBFVFS gives a new algorithm which will exploit efficiently the existing voids in bandwidth, and thus, reduce loss burst, and better manage data contentions.