Abstract: This disclosure describes techniques for improving allocation of computing resources to computation of machine learning tasks, including on massive computing systems hosting machine learning models. A method includes a computing system, based on a computational metric trend and/or a predicted computational metric of a past task model, allocating a computing resource for computing of a machine learning task by a current task model prior to runtime of the current task model; computing the machine learning task by executing a copy of the current task model; quantifying a computational metric of the copy of the current task model; determining a computational metric trend based on the computational metric; deriving a predicted computational metric of the copy of the current task model based on the computational metric; and, based on the computational metric trend, changing allocation of a computing resource for computing of the machine learning task by the current task model.

Abstract: In one embodiment, a supervisory service for a wireless network obtains frequency-time Doppler profile information for an endpoint node attached to a first access point in the wireless network. The supervisory service uses the frequency-time Doppler profile information for the endpoint node as input to a machine learning model. The machine learning model is trained to output an action for the endpoint node with respect to the wireless network. The supervisory service causes the action for the endpoint node with respect to the wireless network to be performed.

Abstract: Systems, methods, and computer-readable media for the secure creation of application containers for 5G slices. A MEC application in a MEC layer of a 5G network can be associated with a specific network slice of the 5G network. A backhaul routing policy for the MEC application can be defined based on the association of the MEC application with the specific network slice of the 5G network. Further, a SID for the MEC application that associates the MEC application with a segment routing tunnel through a backhaul of the 5G network can be generated. A MEC layer access policy for the MEC application can be defined based on the SID for the MEC application. As follows, access to the MEC application through the 5G network can be controlled based on both the backhaul routing policy for the MEC application and the MEC layer access policy for the application.

Abstract: A computer-implemented method includes first collecting, from wireless devices at known locations in a venue, first ultra wideband (UWB) location measurements obtained using a first location technique based on first UWB transmissions made by a mobile device at a first rate. The method also includes second collecting, from the wireless devices, second UWB location measurements obtained using a second location technique based on second UWB transmissions made by the mobile device at a second rate. The method further includes detecting that the mobile device is in a first UWB coverage hole for the venue with respect to the first UWB location measurements based on a first UWB coverage hole criterion. The method also includes, based on detecting, increasing the second rate relative to the first rate to obtain additional second UWB location measurements using the second location technique to compensate for the first UWB coverage hole.

Abstract: In one embodiment, a device in a wireless network receives telemetry data from a plurality of autonomous vehicles. The telemetry data is indicative of radio signal quality metrics experienced by the vehicles at a particular location over time. The device forms an array of wireless roaming thresholds by applying regression to the telemetry data. The device computes an optimum roaming threshold from the array of wireless roaming thresholds to be used by the vehicles when approaching the location. The device triggers, based on the computed optimum threshold, one or more of the autonomous vehicles to initiate access point roaming when approaching the particular location.

Abstract: Automatic onboarding of a device onto a cellular network may be provided through a Wireless Local Area Network (WLAN). Subsequent to a device connecting to a first network (e.g., the WLAN), information associated with the device and the first network may be received. One or more tags may be generated and an intent profile may be defined for the device based on the received information, where the intent profile may indicate at least a second network (e.g., the cellular network) that the device is enabled to connect with and one or more policies associated with the connection. The tags and intent profile may be transmitted to a service provider platform, and an onboarding profile template identified using the tags and the intent profile may be received from the service provider platform. The onboarding profile template may be provided to the device to enable connection to the second network.

Abstract: Presented herein are techniques for assigning Ultra-Wideband (UWB) anchors for client ranging. A control device can monitor UWB ranging between a mobile device and a primary anchor. In response to determining that a signal strength between the mobile device and the primary anchor is below a threshold, the control device can identify anchors for which the mobile device has had a signal strength above the threshold during a period of time, and select one of the anchors as a new primary anchor for the mobile device. For example, the control device can select the new primary anchor based on a relative collision tolerance mapping for the new primary anchor and at least one other anchor within a UWB range of the new primary anchor. The control device can send a command causing UWB ranging to be performed between the mobile device and the new primary anchor.

Abstract: Presented herein are techniques for assigning Ultra-Wideband (UWB) anchors for client ranging. A location server can estimate a coarse location of a mobile device using a localization technique other than a UWB localization technique. The localization technique can involve multiple wireless access points or other radio devices. The location server can define an area around the coarse location to identify a set of candidate anchors for UWB ranging. The set of candidate anchors can be disposed within the area and include at least a subset of the radio devices. The location server can modify the set of candidate anchors to create a modified set of candidate anchors that includes only UWB-enabled devices. The location server can select a primary anchor from the modified set of candidate anchors and send a command to cause a UWB ranging procedure to be initiated between the primary anchor and the mobile device.

Abstract: A method performed by a system which manages station transition from a Wi-Fi access point includes sending, based on the station roaming to an edge-of-domain access point, a request to neighboring access points to the edge-of-domain access point for neighbor reports. The method includes receiving a cellular signal value from the station, comparing a first signal grade associated with the edge-of-domain access point to a second signal grade associated with the cellular signal value to yield a comparison and generating a link usability rating associated with a link between the station and the edge-of-domain access point. Based at least in part on the comparison and the link usability rating, the method includes transmitting a message to the station suggesting that the station roam from the edge-of-domain access point to a cellular network. An edge of the domain can also be dynamic and vary based on station characteristics.

Abstract: Techniques for associating manufacturer usage description (MUD) security profiles for Internet-of-Things (IoT) device(s) with secure access service edge (SASE) solutions, providing for automated and scalable integration of IoT devices with SASE frameworks. A MUD controller may utilize a MUD uniform resource identifier (URI) emitted by an IoT device to fetch an associated MUD file from a MUD file server associated with a manufacturer of the IoT device. The MUD controller may determine that a security recommendation included in the MUD file is to be implemented by a cloud-based security service provided by the SASE service and cause the IoT device to establish a connection with a secure internet gateway associated with the cloud-based security service. Additionally, or alternatively, the MUD file may include SASE extensions indicating manufacturer recommended cloud-based security services. Further, cloud-based security services may be implemented if local services are unavailable.

Abstract: Methods are provided in which a user device connects a participant to a collaboration session in which the participant communicates with at least one other participant using audio and/or video, which is distributed in a media stream to the at least one other participant via a respective user device. In these methods, the user device detects at least one of an object within a space that is included in the video and an audio signal and selectively filters the media stream to exclude the object or a portion of the audio signal based on at least one of participant list information, learned background information, or learned voices of participants of the collaboration session.

Abstract: In one embodiment, a method includes receiving energy efficiency data from a plurality of nodes within a network. The method also includes determining an energy efficiency node quotient for each of the plurality of nodes within the network to generate a plurality of energy efficiency node quotients and determining an energy efficiency path quotient for each of a plurality of paths within the network to generate a plurality of energy efficiency path quotients. The method further includes determining one or more policies associated with the plurality of paths and selecting a path from the plurality of paths based at least on the plurality of energy efficient path quotients and the one or more policies.

Abstract: Presented herein are techniques for assigning Ultra-Wideband (UWB) anchors for client ranging. A location server can estimate a coarse location of a mobile device using a localization technique other than a UWB localization technique. The localization technique can involve multiple wireless access points or other radio devices. The location server can define an area around the coarse location to identify a set of candidate anchors for UWB ranging. The set of candidate anchors can be disposed within the area and include at least a subset of the radio devices. The location server can modify the set of candidate anchors to create a modified set of candidate anchors that includes only UWB-enabled devices. The location server can select a primary anchor from the modified set of candidate anchors and send a command to cause a UWB ranging procedure to be initiated between the primary anchor and the mobile device.

Abstract: Presented herein are techniques for assigning Ultra-Wideband (UWB) anchors for client ranging. A control device can monitor UWB ranging between a mobile device and a primary anchor. In response to determining that a signal strength between the mobile device and the primary anchor is below a threshold, the control device can identify anchors for which the mobile device has had a signal strength above the threshold during a period of time, and select one of the anchors as a new primary anchor for the mobile device. For example, the control device can select the new primary anchor based on a relative collision tolerance mapping for the new primary anchor and at least one other anchor within a UWB range of the new primary anchor. The control device can send a command causing UWB ranging to be performed between the mobile device and the new primary anchor.

Abstract: In one embodiment, a method comprises: determining, by a constrained network device in a low power and lossy network (LLN), a self-estimated density value of neighboring LLN devices based on wirelessly receiving an identified number of beacon message transmissions within an identified time interval from neighboring transmitting LLN devices in the LLN; setting, by the constrained network device, a first wireless transmit power value based on the self-estimated density value; and transmitting a beacon message at the first wireless transmit power value, the beacon message specifying the self-estimated density value, a corresponding trust metric for the self-estimated density value, and the first wireless transmit power value used by the constrained network device for transmitting the beacon message.

Abstract: In one embodiment, a method comprises: determining, by a constrained network device in a low power and lossy network (LLN), a self-estimated density value of neighboring LLN devices based on wirelessly receiving an identified number of beacon message transmissions within an identified time interval from neighboring transmitting LLN devices in the LLN; setting, by the constrained network device, a first wireless transmit power value based on the self-estimated density value; and transmitting a beacon message at the first wireless transmit power value, the beacon message specifying the self-estimated density value, a corresponding trust metric for the self-estimated density value, and the first wireless transmit power value used by the constrained network device for transmitting the beacon message.

Abstract: Techniques for utilizing a cloud service to compute an end-to-end SLA-aware path using dynamic software-defined cloud interconnect (SDCI) tunnels between a user device and an access point-of-presence (POP) node and inter-POP tunnels of the SDCI. The cloud service may include a performance aware path instantiation (PAPI) component including a POP database for storing performance metrics associated with the POPs of the SDCI, an enterprise policy database for storing user specific policies, and/or a path computation component. The path computation component may compute the path, based on the user specific policies, performance metrics associated with the POP nodes, and/or real-time contextual data associated with the user device and/or destination device. The path may include a first tunnel between the user device and the most optimal access POP node of the SDCI and a second tunnel between the access POP node, through the internal POP nodes, and to the destination device.

Abstract: First data indicative of information that a packet is part of a DDoS attack is received at a management network device. A DDoS remediation network device to be used for remediation of packets associated with the DDoS attack is determined from the first data. Second data, indicative of the DDoS attack and indicative of the DDoS remediation network device, is transmitted from the management network device to an edge network device. The second data is configured to cause the edge network device to route packets associated with the DDoS attack to the DDoS remediation network device.

Abstract: A method includes obtaining performance characterization values from endpoints managed by a first fog node at a first hierarchical level in a hierarchy of fog nodes. The method includes changing a first operating characteristic of the wireless network based on the performance characterization values. The first operating characteristic affects the operation of one or more of the endpoints. The method includes transmitting a portion of the performance characterization values to a second fog node at a second hierarchical level in the hierarchy of fog nodes. The method includes changing a second operating characteristic of the wireless network based on an instruction from the second fog node. The second operating characteristic affects the operation of the first fog node and/or other fog nodes at the first hierarchical level. Changing one or more of the first operating characteristic and the second operating characteristic satisfies an operating threshold for the wireless network.

Abstract: Disclosed are methods, systems, and non-transitory computer-readable media for using a sponsor as a proxy for multi-factor authentication of a first user account for a first user when a primary multi-factor authentication mechanism is unavailable to the first user account, comprising registering the sponsor in a multi-factor authentication chain of trust associated with the first user account; requesting verification of an identity of the first user from the sponsor; receiving, from the sponsor, a verification of the identity of the first user; and granting access to a service to the first user account.