Design and analysis of grouping-based DCF (GB-DCF) scheme for the MAC layer enhancement of 802.11 and 802.11n

Abstract

The 802.11 has emerged as the prominent wireless LAN technology as the mobile computing devices such as notebooks and PDA have replaced the desktop computers to be the main trend products. However, if the number of active stations is large, that is high-loading condition for the legacy DCF of 802.11, the capacity will be very low due to high collision costs. In this paper, we introduce the TDMA concept to partition all numerous active stations into several groups to avoid all stations transmitting the frames simultaneously. When Point Coordinator (PC, generally referring to AP) finds that the number of active stations (M) is large i.e. bigger than 8, it broadcasts number of groups (Ng) and group head (Nh) bits (such as 00000100 00000000) information in the TIM field of the beacon frame. Once all stations receive this instruction, the stations which last two LSB bits of the MAC address (IEEE EUI-48 or EUI-64) are 00 belonging to group 0 will transfer their frame first. On the contrary, all stations belonging to other groups will set their waiting time, that is, Network Allocation Vector (NAV) much more precisely. Analysis shows that the capacity of our GB-DCF will be near to the theoretical capacity limit of 802.11 WLAN even if the distributions of all active stations among all groups are not so uniform. This capacity could be independent of the number of active stations and CWMax (Contention window maximum). In this article, we also introduce the grouping cycle concept to our scheme, called GB-DCF+, to reduce the heavy overhead of the legacy DCF and to increase the MAC layer throughput of the upcoming 802.11n protocol. The key idea of GB-DCF+ is that DIFS, SIFS, and ACK are added to the grouping cycle which consists of the transmissions of all groups' slot instead of a single frame. Simulations show that the capacity of our scheme could approach to 45.1% when the PHY data rate, frame size, number of stations, and number of groups are 216 Mbps, 2160 bytes, 128, and 32 respectively. On the contrary, the capacity of DCF will be low to 19.3% with the same scenario. This is due to the tremendous collision costs as well as fixed and heavy overhead of the legacy DCF.