| LAN |  |
 |
 |
| |
| |
| |
| |
| |
|
Next Generation WLANs and WMANs Researchers: Yaser P. Fallah , Prof. Hussein Alnuweiri
|
|
Supporting Multimedia Applications over 802.11e Wireless Local Area NetworksMultimedia applications have specific delay and bandwidth requirements that cannot be fulfilled by the current IEEE 802.11-based wireless LANs. To address this issue, new enhancements have been introduced to the Medium Access Control (MAC) layer of the 802.11 standard under the framework of the IEEE 802.11e standard. Although the 802.11e standard offers new features for supporting QoS in the MAC layer, it does not provide or mandate a final solution and intentionally leaves it to developers to devise their own solutions using the available features. Our work aims at filling this gap by providing a practical and standard compliant Quality of Service solution for 802.11e WLANs. We present a solution that employs the controlled access mechanisms of the 802.11e to provide per-session guaranteed quality-of-service. Voice and video applications will specially benefit from such services. Our solution, called Controlled Access Phase Scheduling or CAPS, easily outperforms methods based on contention and priority (the simple WMM based methods). CAPS provides a QoS framework consisting of an architecture that centralizes the scheduling task in the naturally distributed CSMA/CA environment of a WLAN. CAPS uses the new feature of 802.11e that allows controlled access phases (CAP) to be generated even during a contention period. CAPS describes when and for how long controlled access phases should be generated. It provides a hybrid HCCA/EDCA solution which guarantees access for sessions that make reservations with the Access Point and allows the remaining capacity to be shared in a fair contention manner. This novel solution is based on the concepts of virtual packets, combined uplink/downlink scheduling and a generalized processor sharing based scheduler. More details on this algorithm can be found in the following articles:
To evaluate our solution, we implemented CAPS and integrated it into an IEEE 802.11e OPNET model. The following figures illustrate some of the results gathered from simulation experiments using OPNET. We considered H.264 video sources in presence of background data traffic as well as a voice only WLAN:
Average and Max Delay for a Voice only WLAN Average Delay for a Video flow as background traffic increases OTHER ALGORITHMS: Other than CAPS, which is the main focus of our work, we have devised several simpler algorithms that may be used in less demanding cases. DACE (Distributed- Adaptively Configured EDCA) and FAHPS (FAir HCCA Priority Scheduling), are examples of such algorithms. DACE adaptively configures EDCA parameters and achieves higher throughput than the standard EDCA. DACE is a contention based mechanism. FAHPS, on the other hand, is a controlled access based mechanism and provides better fairness and even higher throughput in the heavily loaded environments, the following figure is a glimpse of how they perform in saturation (compare to standard EDCA). More information can be found in the following article: Pourmohammadi Fallah Y, Alnuweiri H.M. "Enhanced Controlled-Access and Contention-Based Algorithms for IEEE 802.11e Wireless LANs", Wireless Networks, Communications and Mobile Computing, 2005 International Conference on Volume 1, 13-16 June 2005 Page(s):409 - 414
|
| © Copyright UBC |