No abstract

Abstract. Future telephony and multimedia systems will use the Session Initiation Protocol (SIP) for signaling purposes. SIP is a text-based protocol that imposes challenges for an efficient message processing. The ability of SIP entities to process SIP messages quickly is crucial for the performance of these networks, which often have strict timing requirements, e. g., to keep the call setup delays small. This paper studies the performance of SIP message processing in SIP proxies, focusing mainly on the impact of message parsing. We perform a detailed delay analysis for the widely used SIP Express Router (SER). Our measurements show that message parsing actually contributes significantly to a SIP proxy's processing efforts, and therefore confirm other existing studies. However, our results also show that the overall delay in high-performance SIP proxies is stronger affected by other factors, in particular the operating system.

We present a new paging algorithm for wireless networks with ultra-short-range radio access links (picocells). The ubiquitous office (u-office) network is a good example of such a network, and we present some u-office example applications. In addition, we show that conventional paging algorithms are not feasible in such networks. Therefore, we derived a new paging algorithm from the measurement results of an experimental sensor network with short-range wireless links deployed in our office. We equipped persons with sensors and deployed sensor readers at selected places in our office. The sensors transmit messages to the sensor readers at regular intervals. If no sensor reader is in range, the message is lost. Our main observation is that, if a picocell shows an attraction property to a certain person, the residence time of an attached mobile terminal is not gamma distributed (as described in the literature) and the probability of long-lasting residences increases. Thus, if the residence time is larger than a certain threshold, the probability of a long-lasting residence time increases if a sensor reader location has an attraction property to a person. Based on this observation, our proposed paging algorithm registers the location of the mobile terminal only when the residence time in the cell is longer than a predetermined constant. By appropriately setting this constant, we can significantly reduce the registration message frequency while ensuring that the probability of the network successfully connecting to a mobile terminal remains high.