Selectively Fully Polling Scheme For Uplink of WiMAX Network

碩士 === 國立臺灣科技大學 === 電子工程系 === 106 === Worldwide Interoperability for Microwave Access (WiMAX) is a wireless communication schemes. It can provide broadband access to large-scale coverage and mainly used in metropolitan networks. WiMAX is a network using the family of standards of IEEE802.16. To ensu...

Full description

Bibliographic Details
Main Authors: Yu-Sheng Lu, 呂祐昇
Other Authors: Hann-Tzong Chern
Format: Others
Language:zh-TW
Published: 2018
Online Access:http://ndltd.ncl.edu.tw/handle/gadnwk
Description
Summary:碩士 === 國立臺灣科技大學 === 電子工程系 === 106 === Worldwide Interoperability for Microwave Access (WiMAX) is a wireless communication schemes. It can provide broadband access to large-scale coverage and mainly used in metropolitan networks. WiMAX is a network using the family of standards of IEEE802.16. To ensure the Quality of Service (QoS) for different multimedia applications, the service flows are classified into five classes. Nevertheless, the scheduling algorithm was not defined in the standard of IEEE802.16. Therefore, it is a good issue to design a scheduling algorithm to meet the QoS requirement of all classes. With Polling, the base station (BS) can poll subscribe stations (SS) in turn and the polled SS can send bandwidth request to the base station. In “Polling According to Loss-rate Dependent Piggyback Bandwidth Request” (PALDPBR) [1], BS changes the polling interval of extended real-time Polling Service (ertPS) and real-time Polling Service (rtPS) according to the bandwidth requests of these queues. With request upon polled, the polling interval of ertPS will be maintained at the initial value 10ms. Otherwise, polling interval will be increased by 10ms until 40ms. The initial polling interval of rtPS is set as 100ms.With request upon polled, the polling interval of it will be decreased by 10ms until 40ms. Otherwise, the polling interval will be increased 10ms until 150ms. Furthermore, the request of rtPS can be piggybacked on the bandwidth request of ertPS. In other words, the request of ertPS and rtPS of the same SS can be sent in one request if its ertPS is polled. Except this, in PALDPBR, rtPS can also be piggybacked on the request of non-real-time Polling Service (nrtPS) and Best-Effort (BE) according to the packet loss-rate. In this dissertation, we proposed “Selectively Fully Polling Scheme” (SFPS) based on PALDPBR. Our goal is to improve the problem of wasting of frame bandwidth due to no polling. In SFPS, the next polling time of all service flow will be compared if no queue is polled in a frame. If next polling time of one queue is closest to the current time, this queue will be chosen to poll in this kind of frame. In this way, all frames will not be wasted due to no polling. Since the polled queue in this “idle” frame is chosen according to its next pooling time, the order of next polling scheme of all queues will not be changed by this scheme. In this way, the original polling scheme will not be changed greatly. The results show the throughput and packet loss rate of ertPS and rtPS are better than those in the PALDPBR.