1 Abstract— The paper presents a preliminary study of a revised analysis of IEEE 802.11e performance. One of many possible topologies is analyzed in order to emphasize the severe problem of the incapability to prioritize traffic in networks with hidden nodes. The article also provides some innovative conclusions. Index Terms— 802.11e, hidden nodes, QoS, RTS/CTS I. INTRODUCTION IRELESS ad-hoc networks are currently one of the most evolving and popular technologies. Their easy configuration and fast deployment makes them ideal, not only for the average customer and ISPs, but also for emergency situations. Ad-hoc networks, themselves, are not able to satisfy the QoS requirements of different services, such as real-time video streaming or VoIP. Therefore, in 2005, the IEEE 802.11e standard [1] was developed in order to face this serious problem. For ad-hoc networks, it provides QoS guarantees through EDCA (Enhanced Distributed Channel Access). Unfortunately the 802.11 protocols are not resistant to the hidden node (HN) problem. As a remedy, several different solutions have been proposed (cf., [4]), however, the RTS/CTS mechanism is most commonly chosen to combat the hidden node problem. This paper presents a thorough analysis of a four-node 802.11b star topology network with the 802.11e extension. It highlights the problem of the incapability to prioritize traffic in networks with HNs. The main stress is put on the throughput levels of high priority traffic (voice and video) and unfairness in granting medium access. The validation of the given conclusions is done by eliminating HNs. Additionally, the usefulness of the RTS/CTS mechanism is addressed. The rest of the paper is organized as follows. Section II gives a brief description of the analyzed testbed and discusses the obtained results. The article concludes with Section III. II. SIMULATION STUDY The simulation study was performed with the use of an improved version of the TKN EDCA extension [3] to the ns2 simulator. The modifications mostly affect the RTS/CTS mechanism. Important parameters are given in Tables 1 and 2. The work has been realized under NoE CONTENT project, no. 038423. K. Kosek, M. Natkaniec and A. R. Pach are with the AGH - Department of Telecommunications (Poland). L. Vollero is with the CINI - University of Napoli (Italy). {kkosek, natkaniec, vollero}@ieee.org, pach@kt.agh.edu.pl TABLE 1 EDCA PARAMETER SET Priority AC CWmin[AC] CWmax[AC] AIFSN[AC] P0 Vo 7 15 2 P1 Vi 15 31 2 P2 BE 31 1023 3 P3 BK 31 1023 7 TABLE 2 SIMULATION PARAMETERS SIFS 10 μs DIFS 50 μs PIFS 30 μs Slot Time 20 μs Tx Range 250 m Tx Power 0.282 W Frame Size 1000 B Traffic Type CBR/UDP Carrier Sensing (CS) Range 263 m (network w/ hidden nodes) 550 m (network w/o hidden nodes) Node Distance 200 m Wireless Standard IEEE 802.11b with 802.11e The analyzed scenario (Fig. 1) consisted of four nodes (N) sending traffic with priorities (P), with varying sending rates (form 10 kb/s to 7 Mb/s). For the clarity of presentation, only rates up to 4.8 Mb/s per node are presented. Three nodes were hidden. Fig. 1 Star topology network Two main tests were conducted. The aim of the first one was to check the impact of the HNs on the 802.11e performance. The results are shown in Figures 2a-5a. The second experiment was performed with the CS range increased so as to make HNs not hidden any more (see Table 2). The obtained results are presented in Figures 2b-5b, where the error of each simulation point for a 95% confidence interval does not exceed 2%. With enabled RTS/CTS (Fig. 2a) the order of the achieved throughput levels for HNs is in line with 802.11e guidelines. However, from all nodes, the highest throughput is achieved by the unhidden N1 which is sending the lowest priority traffic (P3). This odd performance can be explained by the total frame loss. Due to the fact that nodes experience no DATA Performance Analysis of 802.11e Networks with Hidden Nodes in a Star Topology Katarzyna Kosek, Marek Natkaniec, Luca Vollero, and Andrzej R. Pach, Member, IEEE W