Transmission priority scheme with adaptive backoff technique in fiber-wireless networks
- Wan Hafiza Wan Hassan^{1, 2}Email author,
- Horace King^{1},
- Shabbir Ahmed^{1} and
- Mike Faulkner^{1}
https://doi.org/10.1186/s13638-015-0323-4
© Wan Hassan et al.; licensee Springer. 2015
Received: 4 November 2014
Accepted: 8 March 2015
Published: 3 April 2015
Abstract
This paper considers the realization of fiber-wireless (Fi-Wi) networks using the gigabits passive optical network (GPON) and the infrastructure-based wireless local area network (WLAN). The bottleneck of such hybrid system is the WLAN where interference limits the performance. As such, we focus on enhancing the WLAN performance by analytically deriving the optimum contention window (CW) sizes of access points (APs) and wireless users (WUs), respectively. An adjustable transmission priority factor is introduced to allow uplink-downlink transmission fairness. Further, an adaptive backoff technique using information from monitoring the GPON and WLAN networks is proposed. Simulations show the CW sizes of all WUs are maintained within a standard deviation of 1.5% at a cost of a 3% loss in throughput due to the effect of convergence and other estimates.
Keywords
1 Introduction
Fiber-wireless (Fi-Wi) networks combine the capacity of optical fiber networks with the ubiquity and mobility of wireless networks [1-7]. The use of energy-efficient and stable passive optical splitters makes them the favored choice among other fiber-to-the-home (FTTH) technologies. Gigabits passive optical network (GPON) and Ethernet passive optical network (EPON) are the major standards in PONs. In comparison to EPON, the current GPON offers higher capacity per user and allows more users per PON [8]. This makes GPON more attractive, and hence, it is adopted by major telecommunications service providers (e.g., National Broadband Network (NBN) in Australia and Ultra-Fast Broadband (UFB) network in New Zealand [9]).
However, the increasing number of wireless terminals within households, especially in dense urban and sub-urban areas, results in dominant interference or congestion issues in WLANs. Such congestions occur when a WLAN channel is shared by many users. Hence, the scope of this paper is fully focused on the WLAN performance where multiple APs are connected through GPON at close proximity to each other and share one single channel to serve their respective wireless users (WUs).
WLAN adopts the IEEE 802.11 standard which specifies the distributed coordination function (DCF) as its main medium access control protocol. DCF is a random access scheme, based on the carrier sense multiple access with collision avoidance (CSMA/CA) [12]. This protocol only allows a station to initiate a transmission after it senses the channel is unoccupied for a period defined as DCF interframe space (DIFS). If the channel is sensed busy either immediately or within the DIFS period, the station keeps on listening on the channel until it is sensed idle for a DIFS period. Consequently, the station generates a backoff interval which is randomly chosen from the backoff window (known as contention window (CW) size) [13]. The window size begins with a minimum CW size, and it is doubled at each retransmission up to a maximum CW size. Retransmission takes place whenever there is a packet collision, indicated by the absence of acknowledgment frame (ACK) from the receiver. This backoff algorithm is known as binary exponential backoff (BEB) technique [12].
It is extensively agreed that the BEB algorithm adopted in the MAC protocol for the IEEE 802.11 standard is the key factor to the WLAN performance degradation [13-22]. There are two major drawbacks found in BEB that cause the network degradation as discussed in [17]. First, the contention window is increased upon transmission failure regardless of the cause of failure. Second, after a successful packet transmission, the contention window is reset to the minimum size, thus forgetting its knowledge of the current congestion level in the network and increasing its collision chances.
Numerous modifications have been proposed in literature to improve the legacy BEB algorithm. In [14], the authors proposed to exponentially increase the CW size when there is a collision and exponentially decrease the CW size when there is a successful transmission, known as the EIED (exponentially increases exponentially decreases) algorithm. The exponential factors are then optimized to get maximum throughput. The proponents of [15] use a more conservative approach; they linearly decrease the CW size when there is a successful transmission. Similarly, [20] proposes to halve the CW size (unlike BEB where the CW size is reset to minimum) to increase the overall throughput. The IEEE 802.11e standard changes the minimum and maximum CW limits in the BEB algorithm to define different quality of service access categories [23,24]. In common, all of the above techniques steadily decrease the CW after a successful transmission to retain the overall network state information. Though this approach may have a positive correlation, it usually falls short of predicting the overall network status.
Previous works in [18] and [19] have shown that if exact knowledge (e.g., number of active stations and packet size) of the network is known, the CW size can be tuned to achieve a protocol capacity very close to its theoretical bound. In reality, it is almost impossible for stations to obtain an exact knowledge of the network, only estimates are likely. Listening to the channel to obtain the average packet length and slot transmission probability (alternatively described as slot utility or number of free slots) is the usual starting point [25].
The adaptive window algorithm (AWA) proposed by Bianchi et al. [22] used the number of active stations in the network to control the optimum CW size. A similar approach is employed in [18] and [19] to estimate the number of active stations which is then used in a complex p-persistent MAC protocol to select the optimum backoff interval. In some cases, it is possible to avoid the intermediate step of estimating the number of users [21,25,26].
In common, the above studies focused on the enhancement of the IEEE 802.11 protocol with all stations having an equal chance to transmit. In infrastructure networks, the AP requires more transmission opportunities to give fairness to the uplink/downlink performance [27]. Priority access for an individual user can be obtained by scaling the CW [26], changing the C W _{min} [28-32], utilizing reduced inter-frame spacings (PIFS and SIFS instead of DIFS) [33,34], and adjusting transmission opportunity (TXOP) limits [31,35]. Such techniques have been proposed for QoS enhancements and to provide uplink/downlink fairness in a single BSS infrastructure network. They are less suitable for an infrastructure WLAN network where there is more than one AP sharing the same spectrum and each serving its own basic service set (BSS) of associated WUs.
In this paper, we employ a novel approach known as transmission priority (TxPriority) scheme by introducing a priority factor that can be adjusted to allow AP transmission priorities as required within an infrastructure WLAN network. Mathematical analysis is carried out, and expressions are derived for optimum AP and WU contention window sizes. The derived formulations show that the optimum CW sizes depend on the number of active APs and WUs; unknown quantities. We capitalize on information from both GPON and WLAN networks to provide appropriate estimates and make the system adaptive. A new convergence function is introduced to improve the reliability of the adaptive system. Our mathematical formulations are verified using network simulator software OPNET 16.1 [36].
The remaining sections of this paper are organized as follows: ‘Transmission priority scheme’ section describes the transmission priority scheme and derives optimum contention window sizes. ‘GPON network indicator’ section outlines the GPON architecture and defines an indicator to estimate the number of active APs. ‘Adaptive backoff technique’ section estimates the number of active WUs and proposes an adaptive system. ‘The convergence function’ section presents a new convergence function. Finally, ‘Conclusions’ section concludes the paper and provides directions for future work.
2 Transmission priority scheme
This section derives optimum contention window sizes for the two types of contending stations, an AP and a WU in the WLAN network. The analysis considers the worst case scenario in terms of frequency availability in densely populated urban areas. All BSSs are within close proximity, i.e., all wireless stations, APs and WUs, share one single channel and can hear each other, albeit being potential interferers. The spectrum is considered ‘closed’, that is a single 20-MHz channel entirely dedicated to this Fi-Wi network. When a station wants to contend the channel, the backoff counter is randomly selected within the range (0,CW−1) and then decremented at each subsequent idle slot.
List of notations
Notation | Description |
---|---|
p _{ap} | Probability of an AP transmitting |
p _{wu} | Probability of a WU transmitting |
m | Total number of APs (or BSS’s) |
n | Total number of WUs |
C W _{ap} | Current contention window size for an AP |
C W _{wu} | Current contention window size for a WU |
\(P^{\text {ap}}_{s}\) | Probability that a transmission is a successful AP |
transmission | |
\(P^{\text {wu}}_{s}\) | Probability that a transmission is a successful WU |
transmission |
A value of k<1 implies higher downlink throughput; while a value of k>1 represents higher uplink throughput.
The derived formulation is verified in Appendix A : validation of the formulated optimum CW size.
Our proposed scheme aims to keep the CW sizes as close as possible to the optimum values at all the time. Hence, after a successful transmission, the CW sizes of the stations remain constant; they are not reset to a minimum value as in the BEB scheme.
OPNET simulation parameters
Parameter | Value |
---|---|
Packet size | 8,184 bits |
MAC header | 224 bits |
PHY header | 20 μs |
ACK length | 134 bits/control rate + PHY header |
Data rate | 54 Mbps |
Control rate | 6 Mbps |
Channel bandwidth | 20 MHz |
Slot time | 9 μs |
SIFS time | 16 μs |
DIFS time | 34 μs |
ACK timeout | 70 μs |
Packet transmission time (T) | 30 slots |
Transmission priority factor (k) | 1 |
As expected, lower DL throughputs are observed in comparison to UL throughputs for the BEB and AWA schemes since they are both designed for ad hoc networks. The plots show that at a high traffic load (i.e., 30 BSS), the DL throughputs for BEB and AWA schemes are 0.06 and 0.09 with the corresponding UL throughputs of 0.25 and 0.35, respectively. These resulted in the ratio between DL and UL throughputs for both schemes to be approximately 0.25, which reflects the simulation scenario of having one AP serving every four WUs. On the other hand, our TxPriority scheme with k=1 has an equal DL and UL throughputs of 0.22. An imbalance between DL and UL throughputs is noted for low BSS numbers due to the binomial approximations used in Equation 17 to derive optimum CW size (Equations 21 and 22). Apart from this imbalance, the TxPriority scheme attains throughputs close to the theoretical values derived in Equations 12 and 13 for uplink and downlink throughputs, respectively. Although the TxPriority scheme brings a balance between the uplink and downlink throughputs, it does not compromise the overall throughput. The throughput matches the theoretical results in Equation 9 and that of the AWA scheme. In both schemes, the throughput remained constant over increasing number of BSSs since the CW sizes are assigned on the number of active stations. The BEB scheme does not consider the number of contending stations; it simply doubles its CW size at every unsuccessful transmission and results in deteriorating throughput performance as the number of BSSs increases.
3 GPON network indicator
GPON employs a time division multiplexing (TDM) for downstream (1,480 to 1,500 nm) and time division multiplexing access (TDMA) for upstream (1,260 to 1,360 nm) transmissions. The downstream traffic is broadcast to all ONTs from OLT, and each frame is labeled with the address of its target ONT.
The next section proposes a technique for estimating the number of active WUs n, thus, allowing each station to set optimum CW size.
4 Adaptive backoff technique
where B is the number of times the channel is busy within an observation period (i.e., starting from the time a station first contends the channel until it completes transmission) and I is the total number of idle slots.
where α is a smoothing factor. A value of α=0.8 is chosen since it has been shown in literature [18,22] and [19] as a good compromise between accuracy and precision.
5 The convergence function
In this section, a new convergence function is added into the computation of CW (t). The convergence function acts as a correcting factor in estimating the \(\hat {n}\). In what follows, an analysis is carried out to show that the convergence function for an ad hoc network described in [22] is not appropriate for an infrastructure multiple AP scenario.
6 Conclusions
The paper proposes a transmission priority scheme with an adaptive backoff technique to enhance the WLAN performance of a Fi-Wi (GPON-WLAN) hybrid.The scheme introduces a transmission priority factor k to control the UL/DL fairness. The optimum CW sizes for AP and WU are derived. Performance evaluations show that the scheme is comparable to the adaptive window algorithm (AWA) scheme in terms of overall throughput and delay. Nonetheless, it outperforms the legacy BEB scheme with 40% overall throughput improvement and a maximum 80% of delay reduction. The adaptive backoff technique is introduced. This requires estimates of the number of APs (m) and the number of WUs (n). The first is directly estimated from GPON frame information, and the second is estimated by measuring the activity on the WLAN channel. However, the studied behavior of \(\bar {n}\) (estimate of n) indicates that it does not converge to the correct value. Thus, a new convergence function, \(c(m,\bar {n})\), is added. Simulations show its robustness for practical values of k, m, and n; but there is a small 3% reduction in overall throughput and a slight offset in k. The latter can be corrected by pre-compensation. The scheme shows uniform convergence among all WUs with a measured standard deviation of their CW sizes being less than 1.5%. It is worth noting that the proposed algorithm is compatible with the whole range of WLAN standards because it is aimed at improving the legacy BEB technique in the DCF protocol adopted by all 802.11 family members [24,38]. The scheme improves the integration of fiber-wireless networks as it ensures each station in the BSS is given a fair transmission opportunity so that the huge bandwidth capacity provided by the GPON (backhaul) can be fully utilized. Our future work will further improve the scheme by utilizing additional information from GPON and removing some of the constraints, for example, allowing neighboring GPON clusters and non-closed spectrum.
7 Endnote
^{a} OPNET software was chosen over other simulation packages because of their vast library of built-in functions.
8 Appendix A : validation of the formulated optimum CW size
The above derived equation is identical to the optimum contention window size formulated in [22].
Declarations
Acknowledgements
We acknowledge the support from OPNET Technologies Inc. for providing us the educational version of OPNET 16.1 software. We also acknowledge Prof. Giuseppe Bianchi, for his kindness in providing us further information on proving the stability of an estimate.
Authors’ Affiliations
References
- M Milosavljevic, P Kourtessis, A Gliwan, JM Senior, in Proc. ICTON 2009, paper Tu.A5.4. Advanced PON topologies with wireless connectivity (Azores, Portugal, 2009).Google Scholar
- BO Obele, MH Kang, in Proc. IEEE NGMAST. Fixed mobile convergence: a self-aware QoS architecture for converging WiMAX and GEPON access networks (Wales, UK, 2008), pp. 411–418.Google Scholar
- G Shen, RS Tucker, CJ Chae, Fixed mobile convergence architectures for broadband access: integration of EPON and WIMAX. IEEE Commun. Mag. 45(8), 44–50 (2007).View ArticleGoogle Scholar
- N Ghazisaidi, M Martin, M Reisslein, VMP: A MAC protocol for EPON-based video-dominated FiWi access networks. IEEE Trans. Broadcasting. 58.3(2012), 440–453 (2012).View ArticleGoogle Scholar
- J Wang, K Wu, S Li, C Qiao, Performance modeling and analysis of multi-path routing in integrated fiber-wireless networks, (San Diego, USA, 2010).Google Scholar
- M Maier, M Herzog, STARGATE: The next evolutionary step toward unleashing the potential of WDM EPONs. IEEE Commun. Mag. 45(5), 50–56 (2007).View ArticleGoogle Scholar
- F Aurzada, M Lèvesque, M Maier, FiWi access networks based on next-generation PON and gigabit-class WLAN technologies: a capacity and delay analysis. IEEE Trans. Netw. Mag. 22(4), 1176–1189 (2014).View ArticleGoogle Scholar
- S Bindhaiq, ASM Supa, N Zulkifli, AB Mohammad, RQ Shaddad, MA Elmagzoub, A Faisal, Recent development on time and wavelength-division multiplexed passive optical network (TWDM-PON) for next-generation passive optical network stage 2 (NG-PON2). Optical Switching Netw. 15(2015), 53–66 (2014).Google Scholar
- F Beltràn, Fibre-to-the-home, high-speed and national broadband plans: tales from down under. Telecommun. Policy. 38(8), 715–729 (2014).View ArticleGoogle Scholar
- M Maier, N Ghazisaidi, M Reisslein, in Proc. AccessNets. The audacity of fiber-wireless (FiWi) networks (Hong Kong, China, 2009), pp. 16–35.Google Scholar
- HT Win, AK Pathan, On the issues and challenges of fiber-wireless (Fi-Wi) networks. J. Eng. (2013). [Online] Available: http://dx.doi.org/10.1155/2013/645745.
- IEEE 802.11. standard for wireless LAN medium access control (MAC) and physical layer (PHY) specifications (2007).Google Scholar
- G Bianchi, Performance analysis of the IEEE 802.11 distributed coordination function. IEEE J. Selected Areas Commun. 18(3), 535–547 (2000).View ArticleGoogle Scholar
- N-O Song, B-J Kwak, J Song, LE Miller, in Proc. IEEE VTC-Spring, 4. Enhancement of IEEE 802.11 distributed coordination function with exponential increase exponential decrease backoff algorithm (Jeju, South Korea, 2003), pp. 2775–2778.Google Scholar
- V Bharghavan, A Demers, S Shenker, L Zhang, in Proc. SIGCOMM94. MAWAW: A media access protocol for wireless LANs (London, England, 1994), pp. 212–225.Google Scholar
- WHW Hassan, H King, M Faulkner, in Proc. 23rd IEEE International Symposium on Personal, Indoor and Mobile Radio Communications (PIMRC 2012). Modified backoff technique in fiber-wireless networks (Sydney, Australia, 2012), pp. 431–435.Google Scholar
- I Mouhamad, S Alouf, in NETWORKING 2006 (Networking Technologies, Services, and Protocols; Performance of Computer and Communication Networks; Mobile and Wireless Communications Systems). Design and analysis of an adaptive backoff algorithm for IEEE 802.11 DCF mechanism (SpringerBerlin, Heidelberg, 2006), pp. 184–196.Google Scholar
- F Cali, M Conti, E Gregori, IEEE 802.11 protocol: design and performance evaluation of an adaptive backoff mechanism. IEEE J. Selected Areas Commun. 18(9), 1774–1786 (2000).View ArticleGoogle Scholar
- F Cali, M Conti, E Gregori, Dynamic tuning of the IEEE 802.11 protocol to achieve a theoretical throughput limit. IEEE/ACM Trans. Netw. 8(6), 785–790 (2000).View ArticleGoogle Scholar
- P Chatzimisions, V Vitasas, A Boucouvalas, M Tsoulfa, Achieving performance enhancement in IEEE 802.11 WLANs by using DIDD backoff mechanism. Int. J. Commun. Syst. 20, 23–41 (2007).View ArticleGoogle Scholar
- K Hong, S Lee, K Kim, Y Kim, Channel condition based contention window adaptation in IEEE 802.11 WLANs. IEEE Trans. Commun. 60(2), 469–478 (2012).View ArticleMathSciNetGoogle Scholar
- G Bianchi, L Fratta, M Oliveri, in Proc. IEEE Personal, Indoor and Mobile Radio Communications (PIMRC’96). Performance evaluation and enhancement of the CSMA/CA MAC protocol for 802.11 wireless LANs (Taipei, Taiwan, 1996), pp. 392–396.Google Scholar
- G Cena, L Seno, A Valenzano, C Zunino, On the performance of IEEE 802.11e wireless infrastructures for soft-real-time industrial applications. IEEE Trans. Industrial Inform. 6(3), 425–437 (2010).View ArticleGoogle Scholar
- S Mangold, S Choi, P May, O Klein, G Hiertz, L Stibor, in Proc. European Wireless, 2. IEEE 802.11e wireless LAN for quality of service (Florence, Italy, 2002), pp. 32–39.Google Scholar
- B Luciano, M Conti, E Gregori, Runtime optimization of IEEE 802.11 wireless LANs performance. IEEE Trans. Parallel Distributed Syst. 15(1), 66–80 (2004).View ArticleGoogle Scholar
- E Lopez-Aguilera, M Heusse, Y Grunenberger, F Rousseau, A Duda, J Casademont, An asymmetric access point for solving the unfairness problem in WLANs. IEEE Trans. Mobile Comput. 7(10), 1213–1227 (2008).View ArticleGoogle Scholar
- WHW Hassan, H King, S Ahmed, M Faulkner, in Proc. IEEE African Continent Conference (AFRICON 2013). Modified backoff technique with priority transmission scheme in fiber-wireless networks (Mauritius, 2013).Google Scholar
- M Al-Hubaishi, T Alahda, R Alsaqour, A Berqia, M Abdelhaq, O Alsaqour, Enhanced binary exponential backoff algorithm for fair channel access in the IEEE 802.11 medium access control protocol. Int. J. Commun. Syst. 27(12), 4166–4184 (2013).View ArticleGoogle Scholar
- BAHS Abeysekera, T Matsuda, T Takine, Dynamic contention window control mechanism to achieve fairness between uplink and downlink flows in IEEE 802.11 wireless LANs. IEEE Trans. Wireless Commun. 7(9), 3517–3525 (2008).View ArticleGoogle Scholar
- Y Gao, L Dai, Optimal downlink/uplink throughput allocation for IEEE 802.11 DCF networks. IEEE Wireless Commun. Lett. 2(6), 627–630 (2013).View ArticleGoogle Scholar
- WS Lim, DW Kim, YJ Suh, Achieving fairness between uplink and downlink flows in error-prone WLANs. IEEE Commun. Lett. 15(8), 822–824 (2011).View ArticleGoogle Scholar
- C Wang, Achieving per-flow and weighted fairness for uplink and downlink in IEEE 802.11 WLANs. EURASIP J. Wireless Commun. Netw. 1(2012), 1–9 (2012).Google Scholar
- WS Kim, BS Kim, Y Fang, Downlink and uplink resource allocation in IEEE 802.11 wireless LANs. IEEE Trans. Vehic. Technol. 54(1), 320–327 (2005).View ArticleGoogle Scholar
- F Micò, J Villalón, A Valverde, in Wireless and Mobile Networking Conference (WMNC). Guaranteed access mode for downlink traffic over IEEE 802.11 WLANs (Toulouse, France, 2011).Google Scholar
- G Min, J Hu, ME Woodward, Modeling and analysis of TXOP differentiation in infrastructure-based WLANs. ELSEVIER Comput. Netw. 55(11), 2545–2557 (2011).View ArticleGoogle Scholar
- OPNET Modeler 16.1. http:www.opnet.com.
- ITU-T Recommendation G.984.3. Gigabit-capable Passive Optical Networks (GPON): transmission convergence layer specification (March 2008).Google Scholar
- D Skordoulis, Q Ni, H-H Chen, AP Stephens, C Liu, A Jamalipour, IEEE 802.11n MAC frame aggregation mechanisms for next-generation high-throughput WLANs. IEEE Wireless Commun. 15(1), 40–47 (2008).View ArticleGoogle Scholar
- J Johny, S Shashidharan, in Proc. 4th IEEE International Congress on Ultra Modern Telecommunications and Control Systems and Workshops (ICUMT). Design and simulation of a Radio Over Fiber system and its performance analysis (St. Petersburg, Russia, 2012), pp. 636–639.Google Scholar
- Huawei Technologies, GPON fundamentals [Online] Available: http://www.huawei.com.
Copyright
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.