Predictive traffic identifier-to-link updates in wireless networks
Systems and techniques for performing traffic management in a wireless network using predictive traffic identifier (TID)-to-link mapping are described. An example technique includes obtaining one or more metrics associated with communication between a client station (STA) and an access point (AP) in a wireless network. The communication between the client STA and the AP is based on a first TID-to-link map. A second TID-to-link map is determined, based at least in part on evaluating the one or more metrics with a machine learning model. Communications between the client STA and AP are performed, based on the second TID-to-link map.
1 . A computer-implemented method comprising:
obtaining one or more metrics associated with communication between a client station (STA) and an access point (AP) in a wireless network, wherein the communication between the client STA and the AP is based on a first traffic identifier (TID)-to-link map that allocates traffic from an application running on the client STA to a first communication link established between the client STA and the AP;
determining a second TID-to-link map, different from the first TID-to-link map, based at least in part on evaluating the one or more metrics with a machine learning (ML) model, wherein:
the ML model is configured to output an indication of a second communication link, different from the first communication link, established between the client STA and the AP that satisfies target performance criteria for the application;
the ML model is further configured to output an indication of a time instance when the first communication link established between the client STA and the AP will not satisfy the target performance criteria for the application; and
the second TID-to-link map allocates the traffic from the application to the second communication link established between the client STA and the AP; and
performing communications between the client STA and the AP on the second communication link, based on the second TID-to-link map, wherein performing the communications comprises moving the traffic from the first communication link to the second communication link prior to the time instance.
2 . The computer-implemented method of claim 1 , wherein:
the first communication link is associated with a first radio of the AP that is configured to operate on a first band; and
the second communication link is associated with a second radio of the AP that is configured to operate on a second band.
3 . The computer-implemented method of claim 1 , wherein:
the first communication link is associated with a first radio of the AP that is configured to operate on a first band using a first transmission power scheme; and
the second communication link is associated with a second radio of the AP that is configured to operate on the first band using a second transmission power scheme.
4 . The computer-implemented method of claim 1 , wherein the application is a quality-of-service (QoS)-sensitive application.
5 . The computer-implemented method of claim 1 , wherein the ML model is trained using a dataset comprising at least one of (i) a set of application metrics, (ii) a set of communication link metrics, or (iii) a set of wireless sensing feedback.
6 . The computer-implemented method of claim 1 , wherein the one or more metrics comprise at least one of (i) one or more first metrics associated with one or more applications running on the client STA, (ii) one or more second metrics associated with one or more communication links established between the client STA and the AP, or (iii) one or more third metrics associated with wireless sensing feedback from the client STA.
7 . The computer-implemented method of claim 1 , wherein the ML model is configured to use channel state information (CSI) metrics associated with mobility of the client STA to determine that the client STA is moving towards an edge of a cell of the AP.
8 . The computer-implemented method of claim 1 , further comprising updating, based on the ML model, a TID-to-link map of another client STA associated with the AP to move at least some traffic from the second communication link, such that the second communication link is available for the traffic from the application.
9 . A system comprising:
a memory; and
a processor communicatively coupled to the memory, the processor being configured to perform an operation comprising:
obtaining one or more metrics associated with communication between a client station (STA) and an access point (AP) in a wireless network, wherein the communication between the client STA and the AP is based on a first traffic identifier (TID)-to-link map that allocates traffic from an application running on the client STA to a first communication link established between the client STA and the AP;
determining a second TID-to-link map, different from the first TID-to-link map, based at least in part on evaluating the one or more metrics with a machine learning (ML) model, wherein:
the ML model is configured to output an indication of a second communication link, different from the first communication link, established between the client STA and the AP that satisfies target performance criteria for the application;
the ML model is further configured to output an indication of a time instance when the first communication link established between the client STA and the AP will not satisfy the target performance criteria for the application; and
the second TID-to-link map allocates the traffic from the application to the second communication link established between the client STA and the AP; and
performing communications between the client STA and the AP on the second communication link, based on the second TID-to-link map, wherein performing the communications comprises moving the traffic from the first communication link to the second communication link prior to the time instance.
10 . The system of claim 9 , wherein:
the first communication link is associated with a first radio of the AP that is configured to operate on a first band; and
the second communication link is associated with a second radio of the AP that is configured to operate on a second band.
11 . The system of claim 9 , wherein:
the first communication link is associated with a first radio of the AP that is configured to operate on a first band using a first transmission power scheme; and
the second communication link is associated with a second radio of the AP that is configured to operate on the first band using a second transmission power scheme.
12 . The system of claim 9 , wherein the application is a quality-of-service (QoS)-sensitive application.
13 . The system of claim 9 , wherein the ML model is trained using a dataset comprising at least one of (i) a set of application metrics, (ii) a set of communication link metrics, or (iii) a set of wireless sensing feedback.
14 . The system of claim 9 , wherein the one or more metrics comprise at least one of (i) one or more first metrics associated with one or more applications running on the client STA, (ii) one or more second metrics associated with one or more communication links established between the client STA and the AP, or (iii) one or more third metrics associated with wireless sensing feedback from the client STA.
15 . A non-transitory computer-readable storage medium comprising computer executable code, which when executed by one or more processors, performs an operation comprising:
obtaining one or more metrics associated with communication between a client station (STA) and an access point (AP) in a wireless network, wherein the communication between the client STA and the AP is based on a first traffic identifier (TID)-to-link map that allocates traffic from an application running on the client STA to a first communication link established between the client STA and the AP;
determining a second TID-to-link map, different from the first TID-to-link map, based at least in part on evaluating the one or more metrics with a machine learning (ML) model, wherein:
the ML model is configured to output an indication of a second communication link, different from the first communication link, established between the client STA and the AP that satisfies target performance criteria for the application;
the ML model is further configured to output an indication of a time instance when the first communication link established between the client STA and the AP will not satisfy the target performance criteria for the application; and
the second TID-to-link map allocates the traffic from the application to the second communication link established between the client STA and the AP; and
performing communications between the client STA and the AP on the second communication link, based on the second TID-to-link map, wherein performing the communications comprises moving the traffic from the first communication link to the second communication link prior to the time instance.
16 . The non-transitory computer-readable storage medium of claim 15 , wherein:
the first communication link is associated with a first radio of the AP that is configured to operate on a first band using a first transmission power scheme; and
the second communication link is associated with a second radio of the AP that is configured to operate on the first band using a second transmission power scheme.
17 . The non-transitory computer-readable storage medium of claim 15 , wherein:
the first communication link is associated with a first radio of the AP that is configured to operate on a first band; and
the second communication link is associated with a second radio of the AP that is configured to operate on a second band.
18 . The non-transitory computer-readable storage medium of claim 15 , wherein the application is a quality-of-service (QoS)-sensitive application.
19 . The non-transitory computer-readable storage medium of claim 15 , wherein the ML model is trained using a dataset comprising at least one of (i) a set of application metrics, (ii) a set of communication link metrics, or (iii) a set of wireless sensing feedback.
20 . The non-transitory computer-readable storage medium of claim 15 , wherein the one or more metrics comprise at least one of (i) one or more first metrics associated with one or more applications running on the client STA, (ii) one or more second metrics associated with one or more communication links established between the client STA and the AP, or (iii) one or more third metrics associated with wireless sensing feedback from the client STA.