UPnP服务发现算法性能分析与改进

针对目前可用的U PnP服务发现算法在普适计算环境中存在大量服务时响应消息严重丢失的问题,通过仿真试验分析认为原因在于根设备瞬间发送响应消息和相互独立地随机选择发送延时而产生的响应消息拥塞。进一步地给出了U PnP改进算法,采用根设备断续发送响应消息或(和)综合考虑整个U PnP网络来选择发送延时以减少响应消息拥塞的发生。实验表明改进算法可以明显地降低响应消息的丢失率。同时,改进算法还可以减少控制点对响应消息缓冲区大小的要求。

Evaluation and improvement of UPnP service discovery algorithm

The current UPnP service discovery algorithm can cause severe drops with response message rate when applied in pervasive computing environments with large scales of service.The deficiency lies in the instantaneous transmission of all response messages and the independent random selection of the transmission delay so that the response messages collide and congest.To reduce the congestion,a refined algorithm was developed which transmits the response messages with a certain interval between adjacent messages or(and) selects a planned delay according to the all service-providing nodes.Simulations verify that the refined algorithm gives better performance in the response message's drop ratio and reduces the response message buffer size.