Elsevier

Computer Communications

A comparison of stop-and-wait and go-back-N ARQ schemes for IEEE 802.11e wireless infrared networks

Abstract

The upcoming IEEE 802.11e standard adds a new optional acknowledgment scheme, which is called Burst Acknowledgment (BurstAck) in order to support Quality of Service (QoS) and better utilization of the wireless medium (WM). In this paper the efficiency of the well-known Stop-and-Wait (SW) mechanism and the enhanced Burst Acknowledgment (BurstAck) behavior, utilized as a Go-Back-N (GBN) Automatic Repeat Request (ARQ) scheme with Sliding Window is studied. Link parameters such as, the window size of the transmitted MAC protocol data units (MPDUs), the number of stations, (STAs) the frame error rate (FER) and the signal to noise ratio (SNR) are considered. In our analysis, the specific characteristics of the infrared physical layer as well as the 802.11 Management Information Base (MIB) parameters for infrared wireless LANs and the complex behavior of 802.11 MAC protocol are taken into account. The results obtained indicate that BurstAck utilized as GBN performs better for medium sized networks with large window size and not very high FER. However, for small window size, bad channel quality and large networks the GBN scheme is not suggested.

Introduction

The stations (STAs) in infrared (IR) Wireless Local Area Networks (WLANs) according to the IEEE standard transmit in a fixed wavelength from 850 to 950   nm [1]. Infrared radiation is reflected by indoor environment surfaces, which are nor dark or transparent [2], [3]. IR radiation propagates through multiple reflections and as result a system similar to radio in terms or coverage area is established. As a consequence full mobility of STAs is provided [4], [5]. IR WLANs are preferable in places where the interference produced must be avoided (e.g. airplanes, airports, ships, conference halls, etc.). IR WLANs provide wireless connectivity and support purely cellular architecture [4], [5], which makes them advantageous in covering large indoor spaces. Moreover, optical wireless communication systems can be candidates for Wireless Home Link (WHL) since they can provide high-speed communication between home devices and are unlicensed.

However, IR has several drawbacks. Multi-path dispersion, which is related to the time dispersion of the received pulse, is observed as inter symbol interference (ISI) to the receiver for transmission rates higher than 10   Mbps [3], [5], [6], [7]. In the case of IEEE 802.11 the IR links have transmission rate 1 and 2   Mbps and those phenomena are avoided. Another drawback of the IR link is that ambient light provokes shot noise, due to the random nature of the photo-detection process, while artificial light provokes interference due to light intensity periodic variations [8], [9]. For low and moderate rates as in the case of IEEE 802.11, the ambient noise is the major factor degrading the wireless infrared link performance [8], [9].

It is common to use for the transmission over infrared medium intensity modulation (IM) for the transmitter and direct detection (DD) for the receiver. The signal-to-noise ratio (SNR) is proportional to the square of the received optical power when a DD receiver is used, while in radio transmission it is proportional to the received power. Thus, high levels of optical power must be emitted due to shadowing and ambient noise, which is not permitted by international safety regulations and by power consumption constraints of STAs. Therefore, the transmitted signal must be processed to allow its detection with the lowest possible signal-to-noise ratio. Pulse position modulation (PPM) is adopted by the IEEE 802.11 standards [1] as the transmission technique that offers the best transmission characteristics for this type of transmission channel [9]. Recently, due to the increased demand for multimedia applications on mobile/portable devices, an enhanced version of the IEEE 802.11 standard has been developed [10] to support differential services and quality of service (QoS). One of the features of the new version is the optional acknowledgment scheme, known as Burst acknowledgement (BurstAck). The ARQ scheme performance has been previously addressed for radio transmission [11] and IR transmission [12], [13]. The performance of BurstAck utilized as a selective repeat (SRP) ARQ in the simple case of one transmitter and one receiver is studied in [14]. The more complicated case of arbitrary number of transceivers has been addressed in [15]. In [14], [15] the channel is released by an STA when the successful transmission of N packets, using sub sequenced bursts (with the same or decreasing window size), is completed.

In this work we assume a fixed MPDU size window, the wireless medium (WM) is released immediately after the end of a single burst, regardless if the packets have been transmitted successfully or not and the easier in implementation GBN ARQ scheme with sliding window is adopted. It is the aim of this work to analyze the performance of BurstAck, utilized as a Go-Back-N (GBN) Automatic Repeat Request (ARQ) with sliding window. More specifically, we compare the GBN ARQ scheme with the well-known Stop-and-Wait (SW) ARQ utilized up to now in the IEEE 802.11 standards. In our analysis, we have taken into account the IR physical layer, but it is easy to switch to other physical layers taking into account their specific features. We assume saturation conditions (i.e. the maximum load that can be handled without loosing stability) and a finite number of STAs, in the network, which are contenting to access an IR erroneous channel to transmit a specific number (N) of MAC Protocol Data Units (MPDUs). The network uses the EDCF [16] access method and the request-to-send/clear-to-send (RTS/CTS) scheme being more effective (compared to the basic access scheme) when saturation conditions apply [17], [18]. Numerical results are presented for two cases: (a) all STAs use the SW acknowledgment scheme and (b) all STAs use the BurstAck mechanism. The efficiency of the acknowledgement scheme and the operating conditions are investigated.

The paper is organized as follows: In Section 2 the SW and BurstAck mechanisms are shortly described. The analysis and derivation of the infrared frame error rate (FER) is presented in Section 3. In Section 4, the access delay of a MPDU is derived using the typical MIB parameters of the 802.11 MAC protocol for IR. In 5 Analysis of the SW ARQ scheme, 6 Analysis of GBN ARQ scheme with sliding window, we analyze the average transmission delay for SW and GBN with sliding window. Numerical results and the comparison of SW and GBN performance are presented in Section 7. In Section 8, the advantages and limitations of our analysis are presented.

Section snippets

IEEE 802.11e acknowledgment schemes

The 802.11e MAC protocol [16] supports three types of acknowledgment (Table 1), through a modification of the MAC frame. A new field called Quality of Service (QoS) is used with a subfield called Acknowledgment (ACK) Policy field (Fig. 1). This field can be defined by the bit values presented in Table 1. There is the possibility to use either SW or BurstAck or even no acknowledgment mechanism. The latter is used when the channel quality is extremely high and the non-delivery of some MPDUs such

Analysis of infrared frame error rate

Considering an Additive White Gaussian Noise (AWGN) channel without optical interference and ignoring thermal noise, as specified in the IEEE 802.11 IR PHY section [1], the FER of the IR frame, can be derived [8], [9] for different channel conditions. FER can be written as: FER = 1 − P SYNC P SDF P DR P LENGTH P CRC P MPDU , where P SYNC, P SDF, P DR, P LENGTH, P CRC and P MPDU are the probabilities of the SYNC, SDF, DR, LENGTH, CRC and MPDU fields (Fig. 5) to be correctly detected. The first three fields of the IR

Analysis of 802.11 MAC access delay

Denote as D, the delay between the time an STA has a frame ready to transmit and the time that the STA has captured the medium and is ready to transmit successfully (without collisions) over the WM. E[D] is the mean value of D which is calculated following [18]. E[D] is expressed as: E [ D ] = E [ N c ] ( E [ BD ] + T c + T o ) + E [ BD ] , where E[N c] is the average number of collisions that an STA experiences until the successful reservation of the WM and T 0 represents the time that an STA has to wait, when its frame

Analysis of the SW ARQ scheme

The probabilityP i for the transmitter to send i MPDUs in order to N of them to be received correctly by the receiver is given as: P i = C i − 1 , N − 1 ( 1 − P E ) N ( P E ) i − N , where C i − 1 , N − 1 = ( i − 1 N − 1 ) = ( i − 1 ) ! [ ( i − 1 ) − ( N − 1 ) ] ! ( N − 1 ) ! . and P E is the frame error rate (FER) of the MPDU.

Assuming that the small frames RTS, CTS and ACK are error free, the average time, which is needed to transmit N correct MPDUs, is given as: T SW = ∑ i = N ∞ ( i − 1 N − 1 ) ( 1 − P E ) N ( P E ) i − N i ( T D + T 1 ) , where T 1 = T RTS + T δ + T SIFS + T CTS + T δ + T SIFS + T MPDU + T δ + T SIFS + T ACK + T δ + T AIFS ,

Analysis of GBN ARQ scheme with sliding window

We define P n,i as the probability for the transmitter to transmit successfully N distinguished MPDUS which consist the first created window, using n bursts (windows with N MPDUs) with i MPDUs out of the first N to be transmitted during the last (nth) window. The erroneous MPDUs in each window are discarded by the receiver and retransmitted at the next window, and they do not affect the analysis and as a result simple combinatorics can be used to compute the probability P n,i when BurstAck is

Results

We assume MPDU size 512 bytes, OOK transmitted bit rate 4   Mbps and PPM transmitted bit rate for the MPDU 1   Mbps. The latter is low enough to ignore ISI effects due to multi-path effects [3]. We use a simple PPM threshold receiver with negligible front-end noise. Thus, the noise produced at the receiver is dominated mainly by the ambient light [8]. Note also, that the presented values of SNR of the transmitted IR pulse refer to the electrical SNR. All the parameters of the system under discussion

Discussion

In this paper, we carried out a comparison between the well-known SW and the GBN ARQ scheme over IR 802.11 wireless networks. The presented analysis combined the features of the IR physical layer and the complicated behavior of the CSAMA/CA MAC layer in order to examine in detail the presented ARQ schemes. These two layers affect the performance of the IR system. In particular, it is affected by three main factors: The number of STAs (n) in the network, the FER or SNR of the IR channel and the

Conclusions

We compared the well-known SW acknowledgement mechanism of 802.11 MAC protocol with the enhanced BurstAck scheme of the 802.11e MAC protocol, utilized as a GBN ARQ scheme with sliding window. The network uses the EDCF access method of 802.11e with the same parameters as the standard DCF. A finite number of STAs is assumed to transmit a number of MPDUs with fixed size for two cases. In the first case, all the STAs use the SW acknowledgment scheme, while in the second case they use the BurstAck

References (27)

  • et al.

    CSMA/CA Performance under High Traffic Conditions: Throughput and delay analysis

    Comparative Communications

    (2002)
  • ISO/IEC and IEEE Draft International Standards Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer...
  • J.R. Barry

    Wireless Infrared Communications

    (1994)
  • J.M. Kahn et al.

    Wireless infrared communications

    IEEE Proceedings

    (1997)
  • P. Theodorou

    Analysis of prioritised handover MAC protocol for speech traffic on ATM infrared wireless LANs

    International Journal of Communications Systems

    (2000)
  • G.W. Marsh et al.

    Channel reuse strategies for indoor infrared wireless communications

    IEEE Transactions on Communications

    (1997)
  • J.B. Carruthers et al.

    Modeling of nondirected wireless infrared channels

    IEEE Transactions on Communications

    (1997)
  • S.T. Jivkova et al.

    Multispot diffusing configuration for wireless infrared access

    IEEE Transactions on Communications

    (2000)
  • R.T. Valadas et al.

    The infrared physical layer of the IEEE 802.11 standard for wireless local area networks

    IEEE Communication Magazine

    (1998)
  • A.C. Moreira, R.T Valadas, A.M. Duarte, Performance Evaluation of the IEEE 802.11 Infrared Physical Layer, Proceedings...
  • IEEE Std 802.11e/D3.0. Draft Supplement to STANDARD FOR Telecommunications and Information Exchange Between Systems-...
  • R. Cam et al.

    Throughput analysis of some ARQ protocols in the presence of feedback errors

    IEEE Transactions on Communications

    (1997)
  • V. Vitsas et al.

    Automatic repeat request schemes for infrared wireless communications

    IEE Electronics Papers

    (2002)

    • Cited by (12)

    View full text

    Copyright © 2005 Elsevier B.V. All rights reserved.