Recently, IP satellite networks are gaining a considerable interest to deliver high bandwidth services to nation-wide areas. Satellite bandwidth is a very valuable resource, hence it must be carefully shared among users. Moreover, IP traffic is inherently a bursty traffic, hence a dynamic allocation is needed. However, some difficulties still exist to implement such a method on geostationary satellite networks, since allocation delays have a negative impact on Quality of Service provisioning. The main problem arises from the high delay between the traffic start and the actual bandwidth allocation. To overcome this problem, a prediction of the resource needs has to be considered. The proposed technique takes advantage of a proper statistical traffic modeling to make a prediction of the needed bandwidth. Moreover, requests take into account the DiffServ traffic class in order to maintain a QoS priority among different users.
An Improved Dynamic Bandwidth Allocation Scheme Based on Traffic Prediction for DVB-RCS Systems / F. Chiti;R. Fantacci;T. Pecorella;L. Giacomelli;M. Poggesi;F. Poggianti. - ELETTRONICO. - (2003), pp. 1-10. (Intervento presentato al convegno First International Conference on Advanced Satellite Mobile Systems ASMS 2003 tenutosi a Frascati (Italia) nel 2003-Jul.).
An Improved Dynamic Bandwidth Allocation Scheme Based on Traffic Prediction for DVB-RCS Systems
CHITI, FRANCESCO;FANTACCI, ROMANO;PECORELLA, TOMMASO;
2003
Abstract
Recently, IP satellite networks are gaining a considerable interest to deliver high bandwidth services to nation-wide areas. Satellite bandwidth is a very valuable resource, hence it must be carefully shared among users. Moreover, IP traffic is inherently a bursty traffic, hence a dynamic allocation is needed. However, some difficulties still exist to implement such a method on geostationary satellite networks, since allocation delays have a negative impact on Quality of Service provisioning. The main problem arises from the high delay between the traffic start and the actual bandwidth allocation. To overcome this problem, a prediction of the resource needs has to be considered. The proposed technique takes advantage of a proper statistical traffic modeling to make a prediction of the needed bandwidth. Moreover, requests take into account the DiffServ traffic class in order to maintain a QoS priority among different users.I documenti in FLORE sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.