Statistical tools for admission control decisions in wireless networks
Resumen:
The family of IEEE 802.11 protocols has become the most popular wireless access method. In recent years, other technologies start to complement 802.11 networks, for example, for Internet access in rural networks. Specifically, point-to-point links based on WiMAX and on TV White Space Dynamic Spectrum Access technologies are used to connect a wired Internet access with a set of 802.11 end users. These heterogeneous and multi-hop networks have many challenges in order to provide QoS guarantees. One of these challenges is the design of admission control algorithms. In this paper we develop a blackbox approach for designing admission control algorithms suitable for these kind of networks. The methodology is based on a combination of active measurements and statistical learning tools. The results obtained during simulation and testing in a laboratory testbed show that the methodology has a good accuracy.
2013 | |
Telecomunicaciones | |
Inglés | |
Universidad de la República | |
COLIBRI | |
https://hdl.handle.net/20.500.12008/41771 | |
Acceso abierto | |
Licencia Creative Commons Atribución - No Comercial - Sin Derivadas (CC - By-NC-ND 4.0) |
Sumario: | The family of IEEE 802.11 protocols has become the most popular wireless access method. In recent years, other technologies start to complement 802.11 networks, for example, for Internet access in rural networks. Specifically, point-to-point links based on WiMAX and on TV White Space Dynamic Spectrum Access technologies are used to connect a wired Internet access with a set of 802.11 end users. These heterogeneous and multi-hop networks have many challenges in order to provide QoS guarantees. One of these challenges is the design of admission control algorithms. In this paper we develop a blackbox approach for designing admission control algorithms suitable for these kind of networks. The methodology is based on a combination of active measurements and statistical learning tools. The results obtained during simulation and testing in a laboratory testbed show that the methodology has a good accuracy. |
---|