Collision Resolution Scheme with Offset for Improved Performance of Heterogeneous WLAN

Abstract

CSMA/CA based DCF of 802.11 MAC layer employs best effort delivery model, in which all stations compete for channel access with same priority. Heterogeneous conditions result in unfairness among stations and degradation in throughput, therefore, providing different priorities to different applications for required quality of service in heterogeneous networks is challenging task. This paper proposes a collision resolution scheme with a novel concept of introducing offset, which is suitable for heterogeneous networks. Selection of random value by a station for its contention with offset results in reduced probability of collision. Expression for the optimum value of the offset is also derived. Results show that proposed scheme, when applied to heterogeneous networks, has improved throughput and fairness than conventional scheme. Results show that proposed scheme also exhibits higher throughput and fairness with reduced delay in homogeneous networks.

Appendix

Homogeneous Network

To evaluate system throughput for homogeneous WLAN using (7) following equations are used.

$${\text{E}}\left[ {\text{U}} \right] = {\text{n}}\tau (1 - \tau )^{{{\text{n}} - 1}} {\text{T}}_{\text{s}}$$

$$T_{s} = T + SIFS + DIFS + ACK + 2\alpha$$

$${\text{E}}\left[ {\text{IDLE}} \right] = (1 - \tau )^{n} \sigma$$

$${\text{E}}\left[ {\text{UTC}} \right] = \left[ {1 - (1 - \tau )^{n} - n\tau (1 - \tau )^{{{\text{n}} - 1}} } \right]{\text{T}}_{\text{c}}$$

$$Tc = T + DIFS + \alpha$$

Heterogeneous Network

If P _1,tr and P _2,tr be the probabilities of only one transmission from class one and class two respectively in heterogeneous network then

$$P_{1,tr} = n_{1} \tau_{1} (1 - \tau_{1} )^{{n_{1} - 1}} (1 - \tau_{2} )^{{n{}_{2}}}$$

$$P_{2,tr} = n_{2} \tau_{2} (1 - \tau_{1} )^{{n_{1} }} (1 - \tau_{2} )^{{n{}_{2} - 1}}$$

The E[U] mean value of useful period and E[NU] mean value of non useful period are evaluated as

$$E[U] = [P_{1,tr} + P_{2,tr} ]T_{s}$$

$$E[NU] = E[IDLE] + E[UTC]$$

$$E[IDLE] = [(1 - P_{tr} )\sigma ]$$

where (1 − P _tr )is the probability of no transmission from any class and is given by $(1 - P_{tr} ) = (1 - \tau_{1} )^{{n_{1} }} (1 - \tau_{2} )^{{n_{2} }}$

$$E[UTC] = [1 - (P_{1,tr} + P_{2,tr} ) + (1 - P_{tr} )]T_{c}$$