Dynamic bandwidth allocation in WDM passive star networks with asymmetric traffic

When asymmetric traffic is offered to a WDM passive star network, the offered bandwidth must be allocated is such a way that each station takes a portion of the available bandwidth proportional to its needs. When the traffic characteristics are fixed and a priori known, then, the bandwidth allocation scheme can be based on these characteristics. Unfortunately, the traffic characteristics are often unknown and time-variable. In this paper, a dynamic bandwidth allocation scheme is presented, which is based on the network feedback information in order to be capable of adapting to the changing traffic characteristics. According to the proposed scheme, a set of learning automata processes the network feedback information and dynamically allocates the available bandwidth to the stations according to their needs.