Switchover times in optical communication systems — Can they always be ignored?

Cyclic service queueing systems have a broad range of applications in communication systems. From legacy systems like the slotted ring networks and switching systems, to more recent systems like optical burst assembly, Ethernet over SDH/SONET mapping and traffic aggregation at the edge nodes, all may employ cyclic service as a means of providing fairness to incoming traffic. This would require the server to switch to the next traffic stream after serving one. Usually in optical networks, this switching is done very fast as compared to the service and hence, the switchover time - which is the time taken by the server to switch from one stream to the next - is usually ignored. This however, can cause large differences in the results, especially if the switchover time is not very small as compared to the service time. We present an analytical study of an optical edge node that employs non-exhaustive cyclic service to serve various incoming streams as a two stage process (serving and switchover), and finite size queues to model real systems as closely as possible. We then study the effect of switchover time on the performance of such systems. We compare systems with various ratios of switchover times to the service times and discuss the scenarios under which switchover times cannot be ignored.