Abstract: Learning-based service migration in mobile edge computing may be provided. First, a service migration policy may be created for a network that includes a plurality of edge clouds configured to provide a service to users. Next, a movement of a user receiving the service from a source edge cloud may be detected. The source edge cloud may be associated with a first area and the detected movement may be from the first area to a second area. Then, the service migration policy may be applied to determine whether to migrate the service for the user from the source edge cloud. In response to determining to migrate the service, a target edge cloud may be identified and the service for the user may be migrated from the source edge cloud to the target edge cloud. The service migration policy may then be updated based on a success of the migration.

Abstract: Pattern recognition and classification of packet error characteristics for Multi-user Multiple Input Multiple Output (MU-MIMO) optimization may be provided. First, a packet error characteristic data of a channel for a Downlink (DL) Multi-user (MU) group may be received. Next, the received packet error characteristic data may be provided as input data to a classifier model. Then, in response to providing the received packet error characteristic data as the input data to the classifier model, output data may be received from the classifier model. The output data may indicate at least one probability value corresponding to at least one channel effect. An optimization for improving performance of DL MU-MIMO in the presence of the at least one channel effect may then be performed when the at least one probability value is above or below a predetermined level.

Abstract: Learning-based service migration in mobile edge computing may be provided. First, a service migration policy may be created for a network that includes a plurality of edge clouds configured to provide a service to users. Next, a movement of a user receiving the service from a source edge cloud may be detected. The source edge cloud may be associated with a first area and the detected movement may be from the first area to a second area. Then, the service migration policy may be applied to determine whether to migrate the service for the user from the source edge cloud. In response to determining to migrate the service, a target edge cloud may be identified and the service for the user may be migrated from the source edge cloud to the target edge cloud. The service migration policy may then be updated based on a success of the migration.

Abstract: Inter-protocol interference reduction for hidden nodes may be provided. A first service end point may determine that an inter-protocol interference is present on a channel. Next, an initial packet failure count value on the channel may be determined. A transmit (Tx) power for selected packets may then be increased until a subsequent packet failure count value on the channel is less than the initial packet failure count value.

Abstract: Learning-based service migration in mobile edge computing may be provided. First, a service migration policy may be created for a network that includes a plurality of edge clouds configured to provide a service to users. Next, a movement of a user receiving the service from a source edge cloud may be detected. The source edge cloud may be associated with a first area and the detected movement may be from the first area to a second area. Then, the service migration policy may be applied to determine whether to migrate the service for the user from the source edge cloud. In response to determining to migrate the service, a target edge cloud may be identified and the service for the user may be migrated from the source edge cloud to the target edge cloud. The service migration policy may then be updated based on a success of the migration.

Abstract: Inter-protocol interference reduction for hidden nodes may be provided. A first service end point may determine that an inter-protocol interference is present on a channel. Next, an initial packet failure count value on the channel may be determined. A transmit (Tx) power for selected packets may then be increased until a subsequent packet failure count value on the channel is less than the initial packet failure count value.

Abstract: In one embodiment, a device in a wireless network receives a target wake time (TWT) request from a wireless client. The device computes TWT parameters based on the received request. The device predicts, using the computed TWT parameters as input to a machine learning model, whether the computed TWT parameters will be accepted by the wireless client. The device provides the computed TWT parameters to the wireless client, based on a prediction by the machine learning model that the client will accept the computed TWT parameters.

Abstract: Presented herein are methodologies for setting a target wake time for a wireless client that is being served by an access point, wherein the access point also performs scanning for locating the wireless client. The methodology includes determining a start time of a time window for scanning, e.g., for locations of respective wireless clients, setting a target wake time of a given wireless client, among the wireless clients, based on the start time of the time window for scanning respective wireless clients, and wirelessly communicating the target wake time to the given wireless client.

Abstract: In one embodiment, a method is performed. An access point (AP) device may communicate signaling relating to a client device allocation for a location-related measurement. The AP device may transmit a trigger frame. The AP device may receive a response to the trigger frame from a client device. The response may include an uplink orthogonal frequency division multiple access (OFDMA) transmission. A location-related measurement of the client device may be determined based on the response to the trigger frame.

Abstract: In one embodiment, an apparatus comprises an enhanced distributed channel access (EDCA) selection agent configured to receive a plurality of measurements pertaining to a plurality of clients, compute a set of optimal EDCA parameters using the plurality of measurements, and provide an EDCA configuration for the plurality of clients, and a client behavior predictor configured to receive the plurality of measurements pertaining to the plurality of clients, to receive the set of optimal EDCA parameters, and to compute a plurality of client mode predictions. Client mode predictions may be evaluated and potentially used for additional EDCA parameter optimization by the EDCA selection agent.

Abstract: In one embodiment, a device in a wireless network receives a target wake time (TWT) request from a wireless client. The device computes TWT parameters based on the received request. The device predicts, using the computed TWT parameters as input to a machine learning model, whether the computed TWT parameters will be accepted by the wireless client. The device provides the computed TWT parameters to the wireless client, based on a prediction by the machine learning model that the client will accept the computed TWT parameters.

Abstract: Pattern recognition and classification of packet error characteristics for Multi-user Multiple Input Multiple Output (MU-MIMO) optimization may be provided. First, a packet error characteristic data of a channel for a Downlink (DL) Multi-user (MU) group may be received. Next, the received packet error characteristic data may be provided as input data to a classifier model. Then, in response to providing the received packet error characteristic data as the input data to the classifier model, output data may be received from the classifier model. The output data may indicate at least one probability value corresponding to at least one channel effect. An optimization for improving performance of DL MU-MIMO in the presence of the at least one channel effect may then be performed when the at least one probability value is above or below a predetermined level.

Abstract: Learning-based service migration in mobile edge computing may be provided. First, a service migration policy may be created for a network that includes a plurality of edge clouds configured to provide a service to users. Next, a movement of a user receiving the service from a source edge cloud may be detected. The source edge cloud may be associated with a first area and the detected movement may be from the first area to a second area. Then, the service migration policy may be applied to determine whether to migrate the service for the user from the source edge cloud. In response to determining to migrate the service, a target edge cloud may be identified and the service for the user may be migrated from the source edge cloud to the target edge cloud. The service migration policy may then be updated based on a success of the migration.

Abstract: In one embodiment, a process tracks measured carrier frequency offsets (CFOs) of identified transmitters over a period of activity of the identified transmitters, and determines predicted CFOs for the identified transmitters and predicted transmitter behavior as a probability of specific transmitters of the identified transmitters being active at given times based on the activity of the identified transmitters. The process may then determine, based on the predicted CFOs and predicted transmitter behavior, CFO ranges that a receiver should expect for upcoming packets, and instructs the receiver to use the CFO ranges as a prioritized list of dynamically selected CFOs to use to extract single or colliding packets from among potential interferences using frequency demodulation.

Abstract: In one embodiment, an apparatus comprises an enhanced distributed channel access (EDCA) selection agent configured to receive a plurality of measurements pertaining to a plurality of clients, compute a set of optimal EDCA parameters using the plurality of measurements, and provide an EDCA configuration for the plurality of clients, and a client behavior predictor configured to receive the plurality of measurements pertaining to the plurality of clients, to receive the set of optimal EDCA parameters, and to compute a plurality of client mode predictions. Client mode predictions may be evaluated and potentially used for additional EDCA parameter optimization by the EDCA selection agent.

Abstract: Presented herein are methodologies for setting a target wake time for a wireless client that is being served by an access point, wherein the access point also performs scanning for locating the wireless client. The methodology includes determining a start time of a time window for scanning, e.g., for locations of respective wireless clients, setting a target wake time of a given wireless client, among the wireless clients, based on the start time of the time window for scanning respective wireless clients, and wirelessly communicating the target wake time to the given wireless client.

Abstract: An access point determines the highest signal strength at which it can detect transmissions from other access points or overlapping-basic-service-set (OBSS) client devices associated with the other access points. The access point sets an OBSS packet-detect threshold to the lesser of a minimum cell-size threshold or the highest signal strength and sets a receiver start-of-packet (RxSOP) threshold to the OBSS packet-detect threshold minus an offset. The access point avoids decoding incoming packets when respective signal strengths of the incoming packets do not satisfy the RxSOP threshold. The access point also avoids initiating transmission when incoming packets from the other access points or from the OBSS client devices have signal strengths that satisfy the OBSS packet-detect threshold.

Abstract: In one embodiment, a method is performed. An access point (AP) device may communicate signaling relating to a client device allocation for a location-related measurement. The AP device may transmit a trigger frame. The AP device may receive a response to the trigger frame from a client device. The response may include an uplink orthogonal frequency division multiple access (OFDMA) transmission. A location-related measurement of the client device may be determined based on the response to the trigger frame.

Abstract: An access point determines the highest signal strength at which it can detect transmissions from other access points or overlapping-basic-service-set (OBSS) client devices associated with the other access points. The access point sets an OBSS packet-detect threshold to the lesser of a minimum cell-size threshold or the highest signal strength and sets a receiver start-of-packet (RxSOP) threshold to the OBSS packet-detect threshold minus an offset. The access point avoids decoding incoming packets when respective signal strengths of the incoming packets do not satisfy the RxSOP threshold. The access point also avoids initiating transmission when incoming packets from the other access points or from the OBSS client devices have signal strengths that satisfy the OBSS packet-detect threshold.

Abstract: The present disclosure discloses a network device and/or method for computing coverage set and resource allocations in wireless networks. The disclosed network device selects a radio frequency subdomain in a wireless network, and further determines a coverage set for the selected radio frequency subdomain. The coverage set includes a subset of access nodes in the selected radio frequency domain. Moreover, a respective access node in the radio frequency subdomain satisfies one of (a) the respective access node is a member of the coverage set, and (b) the respective access node is covered by at least one member of the coverage set with a signal strength stronger than a predetermined threshold.