Method of congestion management in a wireless mesh network
View Patent ↗A method of congestion management in a wireless mesh network in which the CSMA/CA algorithm is used to access wireless medium, said network comprising a first station and at least one neighboring station able to communicate directly with the first station, comprises: —broadcasting ( 22 ) an announcement message by the first station to the at least one neighboring station when the first station experiences a congestion in its environment, said announcement message comprising a congestion level parameter, —activating ( 26 ) by the neighboring station, at reception of the announcement message, a congestion state in which the minimum size of the time window, called contention window, to access the medium before transmission of a message is strictly greater than the minimum size (CWmin) defined in absence of congestion, said minimum size in congestion state being defined as a function of the congestion level parameter.
1. A method of congestion management in a wireless mesh network in which CSMA/CA algorithm is used to access a wireless medium, said network comprising a first station and at least one neighbouring station able to communicate directly with the first station, comprising:
broadcasting an announcement message by the first station to the at least one neighbouring station when the first station experiences a congestion in its environment, said announcement message comprising a congestion level parameter,
activating by the at least one neighbouring station at reception of the announcement message a congestion state, in which a minimum size of a time window is called a contention window, to access the wireless medium before transmission of a message is strictly greater than the minimum size defined in absence of congestion, said minimum size in the congestion state being defined as a function of the congestion level parameter,
wherein the congestion level parameter contains the size of the contention window to be used by the at least one neighbouring station in the congestion state, and
wherein the first station keeps track of the number of active neighbouring stations and sets the size of the contention window to approximate a transmission probability of the inverse of the number of active neighbouring stations.
2. The method according to claim 1 , wherein, after having activated the congestion state, if the at least one neighbouring station fails to transmit a packet, the at least one neighbouring station doubles the size of the contention window before retransmitting the packet.
3. The method according to claim 1 , wherein the congestion level parameter is a congestion flag set in case of congestion, and the size of the contention window is doubled in the congestion state compared to the minimum size defined in absence of congestion.
4. The method according to claim 3 , wherein the congestion flag is unset in case of absence of congestion and the size of the contention window is reduced by two when the at least one neighbouring station deactivates the congestion state.
5. The method according to claim 3 , wherein the first station broadcasts the announcement message when the number of collisions during transmission of packets is above a predetermined threshold.
6. The method according to claim 1 , wherein the congestion level parameter is a strictly positive integer, and the at least one neighbouring station sets the size of the contention window in the congestion state to two power the congestion level minus 1.
7. The method according to claim 6 , wherein the congestion level is based on the number of announcements sent by the first station before being able to send a packet.