The virtually unlimited bandwidth of optical fibers has caused tremendous increase in data transmission speed during the past decade, thereby stimulating high demanding Gb/s multimedia services such as distance learning and video conferencing. ATM has been accepted by ITU-T as a promising and efficient technique of transporting Broadband Integrated Services Digital Network (BISDN) applications. The new challenge lies in the deployment of Tb/s multicast ATM switches tomeet the exponential growth of multimedia and internet traffic.
Based on the concept of the researcher's path switching, they propose a Tb/smulticast ATM switch architecture that interconnects electronic switch modules with a quasi-static controlled optical interconnection network (OIN). Routing in the OIN is predetermined to avoid slot-by-slot processing and to provide flexible switching capacity on virtual pathlevel. The surrounding electronic switch modules support multicasting, fast dynamic routing and statistical multiplexing to compensate the the quasi-static routing in the OIN to achieve totally a Tb/s switching capacity. This quasi-static switching architecture simplifies the design of a Tb/s ATM switch to specially featured Gb/s switch modules: the electronic multicast input and output modules and optical central interconnection, which are all feasible with existing technology.
The future research plan includes three aspects. First, the researchers will develop an updating mechanism for readjusting the middle-stage route pattern to adapt the traffic change. The tradeoff between time complexity and performance will be investigated. Secondly, they will identify the requirement of the optical devices to effectively control the crosstalk and the power budget based on our proposed OIN. In addition, they will investigate several approaches of performing output contention resolution within multicast electronic switch modules: work out a cost-effective design for each approach, study the performance in terms of throughput, cell delay and loss rate, and finally identify the best approach that has high performance and reasonable construction complexity.
