In this paper we discuss some experimental results obtained from a laboratory testing of TCP-friendly rate control (TFRC), when such control is employed in an IEEE 802.11g wireless local area network (WLAN). We show that, despite TFRC exhibiting a good responsive behavior and a considerable TCP-friendliness in a wired environment, occasionally it does not behave in the same manner in the wireless test-bed. We discuss the possible reasons of such unexpected behavior, and propose a tuning of some of its congestion indicators, as well as a simple enhancement of the control, in order to get rid of the undesirable effects we observed. In detail, our proposal consists of an additional mechanism for flow control, similar to the one TCP employs, of a more tightly computed no feedback timer, and of a parameter that rules the way TFRC backs-off when it detects a loss event. We pleasantly notice that our modifications and tuning result quite effective for two reasons: first, they completely succeed at eliminating the aggressive behavior sometimes exhibited by TFRC; second, they do not affect the good performance, nor the TCP-friendly behavior of TFRC in those circumstances where the native control already performs satisfyingly.
Performance and TCP-fairness of TFRC in an 802.11g WLAN: Experiments and tuning / M., Borri; Casoni, Maurizio; Merani, Maria Luisa. - ELETTRONICO. - (2004), pp. 362-366. (Intervento presentato al convegno 1st International Symposium on Wireless Communication Systems, 2004. tenutosi a Mauritius Islands nel settembre 2004).
Performance and TCP-fairness of TFRC in an 802.11g WLAN: Experiments and tuning
CASONI, Maurizio;MERANI, Maria Luisa
2004
Abstract
In this paper we discuss some experimental results obtained from a laboratory testing of TCP-friendly rate control (TFRC), when such control is employed in an IEEE 802.11g wireless local area network (WLAN). We show that, despite TFRC exhibiting a good responsive behavior and a considerable TCP-friendliness in a wired environment, occasionally it does not behave in the same manner in the wireless test-bed. We discuss the possible reasons of such unexpected behavior, and propose a tuning of some of its congestion indicators, as well as a simple enhancement of the control, in order to get rid of the undesirable effects we observed. In detail, our proposal consists of an additional mechanism for flow control, similar to the one TCP employs, of a more tightly computed no feedback timer, and of a parameter that rules the way TFRC backs-off when it detects a loss event. We pleasantly notice that our modifications and tuning result quite effective for two reasons: first, they completely succeed at eliminating the aggressive behavior sometimes exhibited by TFRC; second, they do not affect the good performance, nor the TCP-friendly behavior of TFRC in those circumstances where the native control already performs satisfyingly.Pubblicazioni consigliate
I metadati presenti in IRIS UNIMORE sono rilasciati con licenza Creative Commons CC0 1.0 Universal, mentre i file delle pubblicazioni sono rilasciati con licenza Attribuzione 4.0 Internazionale (CC BY 4.0), salvo diversa indicazione.
In caso di violazione di copyright, contattare Supporto Iris