Performance evaluation of power management for best effort applications in IEEE 802.16 networks

Abstract

Reduced energy consumption is a crucial aspect of mobile Broadband Wireless Access (BWA) networks, which are expected to be populated by battery-operated devices, like mobile phones and palmtops. For this reason the IEEE 802.16e specifies a set of power saving mechanisms to be employed by the Base Station (BS) and Mobile Stations (MSs) to concurrently reduce the amount of time the latter have the wireless interface on. The mechanisms are classified into three classes, which are designed for different types of applications. In this paper we focus on class I, which fits the typical requirements of best-effort traffic. With class I, an MS with power saving enabled alternates between sleep and listening periods. The duration of the sleep periods increases by a factor 2 each time a listening period ends, up to a maximum sleep window size. Since the standard does not provide guidelines for setting the above parameters, which are negotiated between the BS and MSs when setting up a power saving class, we evaluate via simulation their impact on the performance, in terms of both application-specific metric, i.e. delay or throughput depending on the type of traffic, and the amount of energy saved.

Power-saving mechanisms for energy efficient IEEE 802.16e/m

“…For FTP traffic with large RTT value or high transmission rate on wireless link, power saving could be achieved by setting the minimum and the maximum sleep window as queue empty time without sacrificing the total file transfer delay much. Similar result is reported in Anastasi et al (2008) from evaluating the performance of PSM with interactive and non-interactive traffic. Performance model for PSC I and PSC II is presented in Kong and Tsang (2006) for showing the distinguishing feature of them for different requirements of service quality.…”

“…PSC I faithfully follows the goal of PSM by keeping the sleep state as long as possible sacrificing response delay. Since PSC I has relatively complicated operation and there is many possibility to improve its performance, study on PSC I attracted many researchers' interests (Xiao, 2005(Xiao, , 2006Zhang and Fujise, 2006;Seo et al, 2004;Anastasi et al, 2008). PSC II is defined for Unsolicited Grant Service (UGS) supporting real-time applications generating fixed-size packets on a periodic basis and Real-Time Variable Rate (RT-VR) service.…”

Reducing Power Consumption with QoS Constraints in IEEE 802.16e Wireless Networks

“…16 [3], which includes the Orthogonal Frequency Division Multiplexing (OFDMA) air interface and supports mobility with users since its "e" amendment in 2005 [4]. QoS is supported since its early versions [5], while the power saving mechanisms have been added recently to cope with the power conservation issue of mobile users.…”

“…In this work, an arrival process modeled through a Markov chain with short-range traffic correlation is used instead of the classic Poisson arrival process. To assess the performance with more realistic traffic models, packet-level simulation has been used in [15], [16], where the authors use the Network Simulator 2. In particular, in [16], we presented a preliminary performance evaluation with the same simulator used to obtain the results in the current work.…”

“…To assess the performance with more realistic traffic models, packet-level simulation has been used in [15], [16], where the authors use the Network Simulator 2. In particular, in [16], we presented a preliminary performance evaluation with the same simulator used to obtain the results in the current work. In our previous work, only power saving class I was involved, i.e., no class II power saving for multimedia applications, and with two traffic models, i.e., telnet and asymptotic TCP, which are very simple compared to the realistic HTTP traffic that we use in the present work.…”