在802.11之存取層下以群為基之可擴充高效能分散協同函數(DCF)與低電耗DCF之設計與分析

Abstract

Today’s 802.11 Wireless Local Area Network (WLAN) technology has prevailed over other technologies and has become a popular protocol for Internet access. However, this technology is not scalable at all because its capacity will deteriorate with an increase in the number of active stations, due to the huge collision costs involved. In this thesis, we propose a high performance Grouping Distributed Coordination Function (HG-DCF) based on [29, 30] which introduces the TDMA concept to partition all active stations into several groups to prevent all stations from transmitting frames simultaneously and to reduce the heavy overhead of legacy DCF and to increase the MAC layer efficiency of the 802.11 protocol. The key idea behind HG-DCF is that the Distributed Inter-Frame Space time (DIFS), Short Inter Frame Space time (SIFS), and Acknowledge (ACK) frames are added to the grouping cycle, which consists of the transmissions of all groups’ slots instead of a single frame. Block-ACK performed by the AP is based on this grouping cycle. Our analysis shows that the capacity of our HG-DCF could reach 93.8%, which is 2.4% larger than the theoretical capacity limit of 802.11 WLAN of 91.4%, even if the distribution of active stations among all groups is not completely uniform. This improvement will increase as the data rate increases or the frame size decreases due to the shorter data time and this capacity can be independent of the number of active stations and the contention window maximum (CWMax). Our research shows that If the data rate is up to 600 Mbps, the capacity of HG-DCF can be up to 27.17% if the new partition scheme shown in this thesis is applied. On the contrary, the capacity limit of DCF with the same scenario is only 14.4%, and thus this improvement can be up to (22.9-14.4)/14.4%∼60%. This grouping technique can also be applied to any DCF-based protocol such as EDCA (Enhanced Distributed Coordination Access) to get the high performance grouping EDCA, denoted as HG-EDCA. Simulations show that HG-EDCA solves the delay jittering problem. A new group partition technique is also proposed, analyzed and simulated in this article to avoid the scenarios with empty group slots and to keep the distribution of the group sizes of HG-DCF more uniform. On the other hand, the energy efficiency of 802.11 is very poor is also due to these heavy overheads in idle listening to the idle channel during the DIFS and back-off and the energy consumed for idle listening is similar to the energy consumed while receiving data [41]. In this thesis, an intelligent scheme for reducing the energy consumed in idle listening is proposed. Our analysis and simulation programs show that our scheme can lengthen the battery endurance up to 3 times due to the shortening in idle-listening time effectively especially when the number of active stations is large. An important characteristic of our scheme is that it is fully compatible with legacy Distributed Coordinated Function (DCF), and there will be no throughput reduction if this power saving scheme is applied to the DCF of 802.11. We also propose an accurate power consumption model in the MAC layer which to the best of our knowledge has not been presented in any earlier research except [46].