Telefonaktiebolaget Lm Ericsson (publ)

Farnaz Moradi

Christofer Flinta

Andreas Johnsson

Catalin Meirosu

Jawwad Ahmed

Lackis Eleftheriadis

Elena Fersman

Leonid Mokrushin

Péter Hága

Selim Ickin

Ramamurthy Badrinath

Junaid Shaikh

Wolfgang John

David Lindero

Daniel Lindström

Chunyan Fu

Valentin Kulyk

Wenfeng Hu

Xiaoyu Lan

Saurabh Singh

Pontus Sköldström

Rolf Stadler

Mathew Clegg

Rafael Pasquini

Bertrand Pechenot

Tim Josefsson

Pontus SKÖLDSTRÖM

A method performed by a central server node in a distributed machine learning environment is provided. The method includes: managing distributed machine learning for a plurality of local client nodes, such that a first set of the plurality of local client nodes are assigned to assist training of a first central model and a second set of the plurality of local client nodes are assigned to assist training of a second central model; obtaining information regarding network conditions for the plurality of local client nodes; clustering the plurality of local client nodes into one or more clusters based at least in part on the information regarding network conditions; re-assigning a local client node in the first set to the second set based on the clustering; and sending to the local client node a message including model weights for the second central model.

Distributed machine learning using network measurements

Embodiments herein relate to a method performed by a network node () for handling monitoring of applications and/or services in a communication network. The network node () obtains an indication associated with a monitoring session to monitor a metric of a service and/or application; and also obtains a location of deployment of the service and/or application. The network node identifies one or more ongoing monitoring sessions for monitoring one or more metrics based on the metric associated with the indication and the location of deployment of the application or service. The network node () then selects a monitoring process based on the identification; and triggers the selected monitoring process.

Network node, monitoring node and methods performed therein

An automated improving of quality of service of a data center. Transients of a power grid fed to a power supply unit are monitored by a probe. Information on transients is provided across an interface to a server of the data center. Based on characteristics of the transients, a reliability of the data center is subjected to automated updating. A request for migration of workload requiring a higher reliability than the updated reliability can be sent to a central management. When the central management has identified another data center that can meet the required reliability, the central management migrates or relocates the workload to the another data center.

Methods and arrangements for automated improving of quality of service of a data center

Systems and methods are provided for generating a predicted number of viewers in support of selecting an advertisement to be displayed during an advertisement break. The method includes: generating, by an advertisement break class identifier module, a plurality of advertisement break classes based on historical advertisement break information and prediction errors generated by a regression model analysis of advertisement break data samples; evaluating, by the advertisement break class identifier module, the plurality of advertisement break classes; redefining and rearranging, by the advertisement break class identifier module, advertisement break classes from the plurality of advertisement break classes based on the step of evaluating which results in an updated set of advertisement break classes; forwarding, by the advertisement break class identifier module, the updated set of advertisement break classes and trained model information to a prediction module; receiving, modifying and returning prediction model information to the prediction module by a training function in the advertisement break class identifier module; generating, by the prediction module, the prediction of a number of viewers for the advertisement break, wherein a regression model uses the updated set of advertisement break classes and data associated with a specific advertisement break to generate the prediction; and sending, by the prediction module, the prediction of a number of viewers for the advertisement break to a scheduler of advertisements.

Dynamic viewer prediction system for advertisement scheduling

A method for predicting a location of a fault in a system is described. The method includes obtaining a faulty unit pertaining to a Supervised Self-Organizing Map (SSOM), wherein the faulty unit has been derived from a sample of machine-level metrics for the system and the SSOM has been trained with a first layer of machine-level metrics and a second layer of service-level metrics. The method further includes expanding a circle originating at the faulty unit until a number of normal units of the SSOM falls within the circle, computing, for each machine-level metric, a score based on differences between the faulty and normal units, wherein a set of scores comprises the respective score for each machine-level metric, and selecting a sub-set of the set, wherein the sub-set comprises the greatest scores of the set, whereby the fault is predicted as located according to metrics represented by the sub-set.

Method and system for fault localization in a cloud environment

A monitoring controller and a method for monitoring network performance are provided. The comprises determining, by a monitoring controller of a first host, a change in a communication between a first and a second application instance by observing application traffic exchanged between the first and second application instance, the first application instance being executed on the first host and the second application instance being executed on a second host. The method further comprises initiating, by the monitoring controller, a monitoring instance in the first host for exchanging test packets between the first host and the second host for monitoring network performance.

Monitoring controller and a method performed thereby for monitoring network performance

A method and network device for resource-aware dynamic monitoring of application units is described. Instantiation of a monitoring element is caused. A first usage status of resources is obtained. Configuration parameters of the monitoring element are set based upon the first usage status of the resources. The monitoring element is scheduled based upon the first usage status of the resources. A second usage status of the resources is obtained. A determination is performed of whether to update the monitoring element based upon the second usage status of the resources. Responsive to determining that the monitoring element is to be updated, at least one of the following is performed: (i) updating the one or more configuration parameters of the monitoring element based upon the second usage status of the resources, and (ii) rescheduling the monitoring element based upon the second usage status of the resources.

Dynamic management of monitoring tasks in a cloud environment

A method and a node for distributed performance monitoring in a communication network comprising a set of nodes. Each node in the set is configured to operate alternatively as a monitor node or as a sensor node and wherein each monitor node in the network is configured to perform measurements together with a corresponding sensor node. The monitor node determines (from a priority list) the sensor node which has the highest priority and sends a first query to that sensor node. The sensor node on its side has a listening period for receiving queries. When the listening period expires, the sensor node selects the monitor node having the highest query priority and performs measurements together with said selected monitor node.

Method and node for distributed network performance monitoring

A processor implemented method for monitoring the performance of a message queue for communication between processes is described. The method includes, by an initiating process, publishing a trigger message to a first topic in the message queue, wherein the trigger message identifies one or more performance metrics and at least two processes that communicate via the message queue. The method further includes, by one or more of the at least two processes, retrieving the trigger message from the first topic in the message queue; performing a measurement on one or more messages exchanged between the processes over the message queue, wherein the measurement determines the one or more performance metrics; and publishing the one or more determined performance metrics to a second topic in the message queue. The method also includes, at a receiving process, retrieving the published performance metrics from the second topic in the message queue.

Message queue performance monitoring

It is disclosed a server node, a virtual container and methods performed therein for establishing a monitoring function monitoring performance metrics of data traffic of a first task container in a virtual environment. The methods comprise instantiating a monitoring function, receiving data traffic, communicated between the first task container and a second task container, which data traffic is mirrored to a monitoring container. The methods also comprise executing the monitoring function monitoring the performance metrics of the mirrored data traffic. Monitoring and task processes are separated, which allows multiple VNFs to share the same monitor reducing overhead. The disclosure allows for an automatic instantiation and configuration of monitor functions deployed in containers, eliminating the need for manual selection of measurement points, and thus reducing operational expenditure.

Growing community of inventors

Stockholm, Sweden

Farnaz Moradi