Performance Studies of a Double-Layered All-Optical Network Architecture

Transmission in complete lightform is now realised with advancements in technology. These include new developments in fabricating the fiber carrier medium, low loss fiber coupling devices, optical switching components for routing lightwave trans mission; laser light sources and sensitive photonic detectors. The in creasing speeds of new generation electronic microprocessors, is capable of resolving the differences in processing and transmission speeds. Access to the medium is regulate d by the medium access control protocol, that permits multiple users to share limited transmission resources of the network. The double-layere dhierarchical all-optical network architecture is proposed, that consists of an upper layer to inter connect sub-networks of the lower layer. The data packets are differentiate d for the two layers . The architecture implements wavelength -space trans mission of wavelength division multiplexed channels. The architecture affords spatial reuse of channels in the lower layer. A non-contentious token-passing medium access protocol is utilised. The token-passing variant that uses one token to provide access to multiple channels is introduced. The performance of the arrayed transmitter of the access node is gauged to determine the suitability of the architecture with the access protocol in supporting multiple accesses. The transmitter can queue a number of data packets awaiting transmission depending on the size of the buffer. Performance indication can be obtained from probabilistic modelling of the changing event states of the transmitter. Performance causal parameters which include the number of nodes, channel allocation and buffer size are defined. The results from the probabilistic models are then analysed and verified with simulation. The architecture provides an inherent feature termed as the bypass that is capitalised to improve performance of the lower layer. Performance indication shows that the architecture is capable of supporting the two types of data packets effectively, and the access protocol is suitable for its purpose. Performance indication of average packet delay improves when the when the bypass feature is implemented. The probabilistic models are found to provide a logical and systematic approach to study and gauge performance of the token-passing access protocol. In conclusion, the double-layered hierarchical AON architecture and the medium access protocol, together serve as a reference for the study of similar scaleable network architectures and their performance.

Text FSKTM_1999_10_A.pdf Download (1MB)

Actions (login required)

View Item