Deployment of wireless networks has grown rapidly in the last decade. It is convenient to connect mobile devices to the Internet without wired line in the public areas, such as libraries, train stations, hotel lobbies, cafeterias, airports, and campus. Nowadays, most of mobile devices such as notebooks, tablet PCs, and PDAs are equipped with one or more wireless communication interfaces. So far, IEEE 802.11 is the most mature and popular protocol amongst other wireless protocols such as Bluetooth, WiMAX, and UWB.
The goal of IEEE 802.11 working group established in 1990 is to develop a wireless local area network (WLAN) standard. In 1997, the group approved IEEE 802.11 as the first WLAN standard with data rates of 1 and 2 Mbps only, through three physical medium, infrared (IR), frequency hopping spread spectrum radio (FHSS) and direct sequence spread spectrum radio (DSSS). The IEEE 802.11 MAC sub-layer defines two medium access coordination functions, the base Distributed Coordination Function (DCF) and the optional Point Coordination Function (PCF). Two year later, in 1999, the working group approved two extended WLAN protocol standards. One is 802.11a, it is based on orthogonal frequency division multiplexing (OFDM), operates on U-NII band (Unlicensed National Information Infrastructure) in 5.4GHz, and with a maximum data rate of 54 Mbps. The other is 802.11b, it is based on DSSS, operates in 2.4GHz ISM band, with a maximum data rate of 11 Mbps. The family of IEEE 802.11 is shown in Table 1.1.
Table 1.1 The IEEE 802.11 family.
Standards Description
IEEE 802.11a Standard for 54 Mbit/sec at 5GHz
IEEE 802.11b Standard to support 11 Mbit/sec at 2.4GHz IEEE 802.11c Bridge operation procedures
IEEE 802.11d International (country-to-country) roaming extensions IEEE 802.11e Standard to support QoS
IEEE 802.11f Inter-Access Point Protocol
IEEE 802.11g Standard for 54 Mbit/sec at 2.4 GHz
IEEE 802.11h Spectrum Management for European compatibility IEEE 802.11i Enhancement for security
IEEE 802.11j Extensions only for Japan
IEEE 802.11k Standard for radio resource management
IEEE 802.11m Initiative to perform editorial maintenance, corrections, improvements, clarifications, and interpretations for the IEEE 802.11 family specifications.
IEEE 802.11n Standard for high throughput improvements with MIMO (multiple input, multiple output antennas)
IEEE 802.11p Standard for WAVE ( Wireless Access for the Vehicular Environment ) IEEE 802.11r Standard for fast roaming to specify fast BSS ("Basic Service Set")
transitions
IEEE 802.11s Enhancement for mesh networking
IEEE 802.11t Wireless Performance Prediction (WPP) – test methods and metrics recommendation
IEEE 802.11u Enhancement for interworking with external non-802 networks IEEE 802.11v Wireless network management
Recently, voice over Internet Protocol (VoIP) became one of the most popular applications.
Some products of VoIP, such as Skype, are able to support good quality of voice communication over wired networks. VoIP market grows up quickly due to its low cost and easiness to construct IP network than traditional telecommunication networks. Everyone can use its computer to make cheaper or even free voice call instead of using the expensive cell phone services. According to statistics, it says that the number of residential VoIP users will rise from three million at 2005 to 27 million by the end of 2009 [13].
Therefore, more and more people try to use handheld devices to make VoIP calls through wireless LANs. For this reason, the quality of service (QoS) in WLAN becomes increasingly important. However, voice communication over WLAN features many challenges, such as low bandwidth, large interference, long latency, high loss rates, and jitter. The distributed coordination function (DCF) and the point coordination function (PCF) are unable to guarantee QoS effectively [3], this is because DCF cannot provide QoS trivially, PCF is not efficient for only one frame sent at each polling, point coordinator does not know the QoS requirement of traffic and does not guarantee the delay and jitter bound, so it is harder to provide QoS for real-time audio/video transmission in wireless networks than in wired networks.
IEEE 802.11e standard [1] defined at IEEE 802.11 Task Group E is expected to solve the QoS problem of latency sensitive applications over wireless local area networks. 802.11e standard proposes some MAC mechanisms to support time-sensitive applications. Hybrid Coordination Function (HCF) includes two methods, one is Enhanced Distributed Channel Access (EDCA) that combines DCF, and the other is HCF-Controlled Channel Access (HCCA) that is similar to PCF but with enhancement. Direct Link Protocol (DLP) is the option that allows stations to exchange packets directly with other stations, bypassing the AP. Block Acknowledgement, an optional function in implementation, improves channel efficiency by aggregating several acks into one frame.
However, IEEE 802.11e is still unable to support satisfactory QoS for time sensitive
applications. EDCA provides QoS based on probability, not determinism. This means, in worst case, the quality of delay time or jitter bound is not desirable. Besides, HCCA polling scheme is not specified in the standard, and the study in [2] shows the HCCA polling overhead problem and its negative effect on the QoS of real-time applications in WLAN. Consequently, a good polling scheme can improve the channel performance and thus QoS for time sensitive applications.
Nowadays, the most popular polling scheduler is the round robin (RR) polling scheduler [5]
because it is simple in implementation. Many other polling schemes such as [4], [6], [7], [8] have been proposed.
In this study, we focus on HCCA polling scheme mainly. We amend a new time-based polling scheduler called Adaptive Time-Stamp Polling (ATSP) scheme. ATSP scheme shows that it can improve total channel utilization and decrease delay jitter variation. In our scheme, Hybrid Coordination (HC) operates at the access point (AP), receives traffic specification (TSPEC) from stations which require polling to send out QoS frames, and records the service start time and maximum service interval corresponding to each station. Then, to avoid polling all stations in turn within one contention-free period (CFP), and using the same service interval for all stations to be polled with different duration of interval, we start to poll a station at its start time of registered service, and the interval for polling this station is same as the interval registered in its TSPEC. By this way, we can reduce the number of polling responded with QoS-Null frame because we don’t poll a station with excessive frequency than its frame sampling frequency, and do not poll a station before it starts communication.
In addition, we use a simple mechanism to detect silence mode of a VoIP conversation.
Then, we reduced the frequency of polling a silence station in order to decrease unnecessary waste of time. When the station comes back to talk-spurt mode, we revert to the original frequency to poll.
The rest of this thesis is organized as follows. Chapter 2 introduces the background of 802.11, 802.11e, voice/video transmission characteristic, and a brief survey of current polling
mechanisms. In Chapter 3, we discuss the proposed polling mechanism in detail. In Chapter 4, simulation and numerical results are demonstrated. Finally, the conclusion and future works are presented in Chapter 5.