Service discovery protocols are used to facilitate dynamic cooperation among devices / services with minimal user intervention. These protocols work by exchanging messages to identify and locate the relevant services. When new devices and services are added the flow of messages often appears as a burst. These bursts act as a critical factor that may cause the loss messages which dramatically influences the performance of the service discovery process. Thus, the loss of messages results in uncompleted discovery process which will directly affect the sharing of resources in home networks. This paper proposes an algorithm that computes the minimum period of time required between a consecutive burst of messages and determines the minimum required queue sizes for the routers to manage the traffic and avoid dropped messages. The algorithm has been applied to the Universal Plug and Play (UPnP) protocol and considered the used of Active Queue Management (AQM). It was tested when the routers were connected in two configurations; decentralised and centralised. The message length and bandwidth of the links among the routers were taken into consideration. The result shows better improvement in number of dropped messages `among the routers.
展开▼
