Celona, Inc.

Preethi Natarajan

Mehmet Yavuz

Srinivasan Balasubramanian

Rajeev Shah

Phaneendra Mettu

Method and apparatus for scalable machine learning (ML)-based frameworks for resource planning and recommendations for Enterprise Networks are disclosed. In some embodiments, a method for determining network health assessment and providing network planning and recommendations in wireless enterprise networks is provided. The method quantifies resource health of a plurality of network elements operating within the enterprise network by calculating a plurality of distance metrics between acceptable health parameters and observed network resource utilization values associated and corresponding to each selected network element. The method ranks each of the network elements based upon the distance metrics calculated for each network element. Factors that negatively impact the resource health of the network elements are ranked in accordance to how severely they impact the performance of the network elements. The method and apparatus provide suggestions and recommendations to improve network performance of the network elements. An apparatus for scalable machine-learning (ML)-based frameworks for resource planning in enterprise networks is also disclosed.

Method and apparatus for scalable ML-based frameworks for resource planning in enterprise networks

Methods and apparatus for autoconfiguring microslices in enterprise wireless communication networks using a microslice knowledgebase in a remotely located orchestration platform. Microslices are independent end-to-end logical networks operating on a shared physical infrastructure, which ensure certain Quality of Service (QoS) and Service Level Objective (SLO) requirements are met for different service types or applications. Any of the enterprise networks may be a Citizen's Broadband Radio Service (CBRS) system. A microslice knowledgebase unit in the orchestration platform configures microslices for the enterprise networks based upon communication requirements of applications, devices, and groups of devices such as service type, bandwidth requirements, using machine learning and artificial intelligence to learn communication requirements of each enterprise deployment. The microslice configuration unit in the knowledgebase may develop microslice templates based upon enterprise type, which can be used to better make recommendations to enterprise networks during initial deployment and afterwards.

Method for autoconfiguring microslices in a wireless enterprise network

A transfer management system for efficiently operating an enterprise wireless communication network (EN) at a campus location enables the EN to control congestion, selectively and smoothly admit UEs, and manage UE exits from the EN. An MNO Network (MN) footprint learning system learns the different MN footprints on the campus, relative to the BS/APs deployed by the individual enterprise, which defines a footprint of wireless coverage for each MNO that has wireless coverage. The MN footprint provides useful information that enables successfully transitioning a UE between an EN and an MNO, which provides the EN with the ability to utilize other networks to offload one or more UEs to manage congestion, delay or deny admission of UEs, and proactively exit UEs to manage congestion or for other reasons.

Method and apparatus for determining wireless MNO coverage and efficiently operating an enterprise wireless communication network

A method and apparatus for managing wireless Enterprise Networks (EN) for greater efficiency by developing inactivity timers for packet streams on bearers, monitoring the packet streams of UEs on the bearers, and if inactivity is found determining whether or not to release or modify a bearer, or maintain the current status of the UE considering factors such as congestion, expected lull time, reconstruction cost, and UE type and location. Accordingly, UEs can be more efficiently allocated the EN's available resources. To learn the timers, packet streams are monitored over a period of time, the packet streams are sorted by service type and bearer (and optionally time), and an inter-burst packet arrival time is determined. Timers are developed that provide inactivity time limits and an expected lull time that predicts the behavior of the packet streams based on service type and bearer, and optionally time.

Method and apparatus for selectively releasing user equipment devices to efficiently operate an enterprise wireless communication network

A method and apparatus for managing network loading and adjusting admission, and/or congestion and service types based on real-time analytics in an enterprise wireless communication network that includes Base Stations/Access Points (BS/APs) in communication with a number of User Equipment devices (UEs) within the coverage area at a campus location. The system dynamically learns the parameterizations for the control system, and adapts the parameters as the network usage evolves, which is useful for an efficient self-managed network. Such a system can provide a greater QoS across users and other efficiencies. The system simplifies deployment at the enterprise's campus locations because load heatmaps and service patterns are learned based on the actual campus locations.

Method and apparatus for load control of an enterprise network on a campus based upon observations of busy times and service type

A method and apparatus for monitoring network performance in near real-time by making measurements on packets received at an intermediate node in wireless communication network. The solution is useful for monitoring wireless network performance of any packetized wireless communication network that connects a client and application server, and particularly for any application running over TCP/IP protocol. A method is disclosed for measuring packet loss rate of a packet-based communication session between a Network Source (NS) and a User Equipment (UE) device at an intermediate node, in the downlink direction and the uplink direction. The measured loss is indicative of the loss in the portion of wireless network between the intermediate node and the UE. The measured packet loss rate is compared with service guarantees for the wireless network, and if the service guarantees are not being met, then resolution mechanisms can be implemented.

Method and apparatus for measuring packet loss rate via deep packet inspection at an intermediate node in a communication network

A method and apparatus for monitoring network performance in near real-time by making measurements on packets received at an intermediate node in a wireless communication network. The solution is useful for any packetized communication network that connects a client and application server, and particularly for any application running over TCP/IP protocol. A method is disclosed for measuring end-to-end qualities of a packet-based communication session between a data sender (DS) and a data receiver (DR) at an intermediate node. The measured end-to-end communication qualities may include latency and packet delay variation.

Method and apparatus for measuring end-to-end packet latency and packet delay variation via deep packet inspection at an intermediate node of a communication network

Growing community of inventors

Saratoga, CA, United States of America

Preethi Natarajan