Abstract: In one implementation, a device makes a classification of sensor data captured by an embodied agent. The device selects, based on the classification of the sensor data, between storing the sensor data in a local memory in a local network of the embodied agent or in a cloud memory in a cloud location. The device makes a classification of a subtask of a task to be performed by the embodied agent. The device selects, based on the classification of the subtask, whether the subtask is to be performed in the local network using the local memory or in the cloud location using the cloud memory.

Abstract: Discovery enhancements for improved roaming may be provided. A client device may discover basic information for each of a plurality of Access Point (APs). Then the client device may select at least one of the plurality of APs as a seamless roaming candidate based on the basic information. Next, the client device may discover full information for the at least one of the plurality of APs.

Abstract: Network initiated Wi-Fi Stream Classification Service (SCS) may be provided. SCS flow parameters may first be negotiated. A flow may be provisioned using the SCS flow parameters. A Station (STA) may associate with an Access Point (AP), and the STA may be allowed to communicate with the AP using the flow according to the SCS flow parameters.

Abstract: In one implementation, a device may monitor Distributed Network Protocol traffic sent over a network of a supervisory control and data acquisition system. The device may identify a categorization of a network port associated with the Distributed Network Protocol traffic. The device may determine, based on the categorization of the network port, whether the Distributed Network Protocol traffic is originating from a rogue master station. The device may mitigate activity by the rogue master station.

Abstract: Client preferences signaling for buffered downlink (DL) data delivery during seamless roaming may be provided. A serving access point (AP) receives a roaming request from a client device, the roaming request indicating an intent to roam from the serving AP to a target AP and including one or more preferences for handling buffered DL data at the serving AP. The serving AP sends a roaming response to the client device, the roaming response indicating parameters for delivering buffered DL data based on the one or more preferences.

Abstract: The techniques relate to configuring a large language model to address impairment scenarios within a network. For example, the techniques include obtaining ground truth information regarding a simulation of a scenario related to an impairment with a network. A large language model is used to complete a task related to the scenario. Based on the ground truth information, an action taken by the large language model is evaluated to complete the task. Based on the evaluation, one or more actions that the large language model is allowed to perform to complete the task are updated.

Abstract: In one implementation, a device in a network maintains a link between the device and each of its neighbor nodes along a primary ring in the network. The device also maintains a link between the device and a particular node in the network that is not a neighbor of the device along the primary ring in the network. The device makes a determination that a second failure exists along the primary ring in the network. The device unblocks, based on the determination the link between the device and the particular node to allow traffic in the network to bypass the second failure.

Abstract: In one embodiment, a device makes a first determination as to whether it has wired network connectivity. The device also makes a second determination as to whether it is moving over time. The device sends an advertisement of capabilities of the device based on the first determination and on the second determination. The device joins a wireless network with a role assigned to the device based on the capabilities of the device indicated in the advertisement, wherein the role is either a fixed node role or a mobile node role.

Abstract: Generative artificial intelligence models receive prompts as input. The prompts can be augmented using contextual data retrieved by a retrieval-augmented generation process. Including sensitive data in a prompt or in contextual data can compromise the security of the sensitive data by exposing proprietary information to the respective generative artificial intelligence model. The sensitive data can be secured while remaining useful as contextual data using a secure augmentation logic. The secure augmentation logic receives sensitive data related to a plurality of entities and generates a knowledge graph that encodes relationships among the entities. The secure augmentation logic consistently anonymizes one or more entities in the knowledge graph according to a sensitive data anonymization policy. The secure augmentation logic generates a set of data representations based on the consistently anonymized knowledge graph.

Abstract: A method that may include receiving a first data signal at a first sampler input including charge pulses of a first amplitude; receiving a second data signal at a second sampler input including charge pulses of a second amplitude; transmitting the first data signal to a first transistor network including a first push-pull pair of transistors of the first transistor network; controlling the push-pull pair using a pair of clock signals, each clock signal of the pair has a complementary phase and thereby operating the first push-pull pair at a frequency lower than an overall sampling rate of one or more ADCs; and outputting a plurality of push-pull charge pulses from the sample, wherein when the push-pull charge pulses are combined at a demultiplexer, the combined pulses have a third amplitude greater than the first amplitude or the second amplitude.

Abstract: Devices, networks, systems, methods, and processes for probing a network path in a Disaggregated Scheduled Fabric (DSF) are described herein. A source device, such as a first leaf switch can receive a probe command and identify a network path associated with the probe command. The first leaf switch may generate, based on the probe command, a probe packet comprising a fabric header. The first leaf switch can identify an adjacent device such as a spine switch and transmit the probe packet to the spine switch. The spine switch can forward the probe packet to a destination device, such as a second leaf switch. The second leaf switch may generate and transmit a response probe packet in response to the probe packet. The probe packet can be utilized to measure one or more Quality of Service (QOS) parameters such as jitter, latency, packet loss, or congestion etc. for example.

Abstract: Method, system, and/or computer readable medium for roaming on a Wi-Fi network having a plurality of access points including: determining that a new AP actor provides a station (STA) actor enhanced characteristics as compared to an existing AP actor, wherein the enhanced characteristics include one or more of a stronger connection, lower congestion, and/or increased bandwidth; transferring required state from the existing AP actor to the new AP actor; and transferring at least some of data wherein the at least some data includes one or more of medium access control service data units (MSDUs), aggregated-MSDUs (A-MSDUs), medium access control protocol data units (MPDUs), and/or aggregated-MPDUs from the existing AP actors to the new AP actor.

Abstract: One implementation of the disclosure is directed to an Information Technology Service Intelligence (ITSI) that provides methods for automatically generating one or more severity level thresholds for a metric represented by a time-series data set through adaptive thresholding. One implementation may include operations to retrieve a historical time-series data set that represents a KPI or metric being monitored, perform an adaptive threshold generation process for the time-series data set that includes automatically selecting a seasonality pattern that corresponds to the time-series data set and determining a set of thresholds based thereon. Selecting a seasonality pattern may involve, for each candidate seasonality pattern, computing a silhouette score that indicates how closely a candidate seasonality pattern corresponds to the time-series data set.

Abstract: An incident arising within a user system is analyzed to determine services directly or indirectly impacted by the incident. One or more responders are also determined for the incident based on user skill sets and details of the incident. The impacted services and the determined responders are then used to determine an impact of the incident within the user system, and a response to the incident may be prepared based on the impact. Response scheduling data may also be automatically prepared for the determined responders to minimize a mean time to resolution (MTTR) for the incident.

Abstract: Presented herein are techniques for providing a Messaging Layer Security (MLS) gateway that is a verified participant in an MLS group. The MLS gateway joins an MLS group as a first verified member of the MLS group for secure messaging among participants of the MLS group. The MLS gateway transmits first secure content obtained from a first participant to one or more second participants of the MLS group. The first participant is a non-MLS capable participant, and the one or more second participants of the MLS group are verified MLS capable participants. The MLS gateway transmits second secure content obtained from the one or more second participants of the MLS group to the first participant.

Abstract: In one embodiment, a method includes receiving, by a network node, traffic within a hierarchical software-defined wide area network (SD-WAN) network. The method also includes determining, by the network node, a destination region of the traffic. The destination region is within the hierarchical SD-WAN network. The method further includes classifying, by the network node, the traffic based on a destination match condition. The destination match condition is associated with two or more destination regions.

Abstract: Presented herein is a universal Large Language Model (LLM) firewall/gateway that operates at the LLM level to protect clients and LLMs from a new threat landscape. The LLM firewall/gateway may generate alerts warning users about insecure codes that LLMs propose, as well as detecting and preventing LLMs from jailbreaking through sessions/conversations. A method is provided comprising: intercepting communications associated with a conversation between a client and a Large Language Model (LLM) service, the communications including a request message from the client to the LLM service and a response message from the LLM service to the client; deriving a context for the conversation based on the communications between the client and the LLM service; and applying one or more policies to the communications between the client and the LLM service based on the context.

Abstract: Disclosed is technology that provides an easier and more convenient method for onboarding network devices by allowing the use of user input (e.g., voice commands) to input network device and network controller configurations. The technology can receive user input and determine the intent of the user from the user input. The technology can then execute the necessary configurations and choose the appropriate controller that will provision the device onboarding configurations. Thereafter, the technology can register the device with the controller and configure the network accordingly.

Abstract: A method performed by a leader wireless access point in a multiple access point coordination system. The method includes estimating a compute time for the leader wireless access point to compute a resource assignment schedule for one or more follower wireless access points; during a transmit opportunity time interval, transmitting over a wireless channel to the one or more follower wireless access points a queue report request to solicit from each of the one or more follower wireless access points a queue report response; based on the compute time, setting a deadline by which a computing component of the leader wireless access point is to complete computing the resource assignment schedule; and while the computing component is computing the resource assignment schedule, applying one or more schemes over the wireless channel to keep the wireless channel reserved from other uses until the deadline is reached.