X. Zhou et al. (Eds.): EUC Workshops 2006, LNCS 4097, pp. 203 – 212, 2006. © IFIP International Federation for Information Processing 2006 Prediction Efficiency in Predictive p-CSMA/CD Marek Miśkowicz AGH University of Science and Technology, Department of Electronics al. Mickiewicza 30, 30-059 Kraków, Poland miskow@agh.edu.pl Abstract. Predictive p-persistent CSMA protocol is an important MAC solution developed for networked sensor/control applications and used in Local Operat- ing Networks (LonWorks) technology. The protocol uses a built-in network load prediction to support collision avoidance. The paper presents an analytical study of prediction efficiency for a channel with collision detection. The ap- proach based on Markov chains is applied. The procedure of performance analysis includes the definition of transition probabilities of Markov chain for a specified load scenario, calculation of stationary distribution of contention win- dow, and the probabilities of successful/unsuccessful transmission. It is shown that the predictive p-CSMA protocol manages to control the size of a competi- tion window in order to guarantee the sustained probability of a successful transmission. The simulative validation of analytical results is provided. Keywords: MAC protocols, sensor networks, performance analysis, Markov chains. 1 Introduction Predictive p-persistent CSMA protocol is the heart of LonTalk/EIA-709.1 protocol used for communication between smart devices in Local Operating Networks (LON, LonWorks). LON has become a classic solution in building automation, and home networking, but is also used in a wide range of applications including, among others, semiconductor manufacturing, petrochemical industry, and wastewater treatment. In the predictive p-CSMA the collision avoidance is applied for resolving uncoor- dinated access to the network. In order to cope with overload situations, the predictive p-persistent CSMA has been designed as an adaptive version of pure p-CSMA. In p- persistent CSMA, a node transmits with a fixed probability p if the channel is idle, and defers the transmission with the probability (1-p), when the channel is busy [10]. In the predictive p-CSMA, the probability p is variable and dynamically adjusted to expected traffic load using backoff scheme. The advantage of the predictive p-CSMA performance is the throughput optimization, which consists in keeping a high throughput even if a network is overloaded. Consequently, although the predictive p- CSMA is a random access scheme, LonWorks networks offer „guaranteed band- width” regardless of the offered traffic rate [3]. Under light traffic load, the predictive CSMA is reduced to the pure 0.0625- persistent CSMA regardless of the structure of the traffic in a channel [3]. However, the destiny of protocol operation is to cope with high traffic corresponding to the