British Telecommunications Public Limited Company

British Telecommunications Corporation

Khalifa University

Emirates Telecommunications Corporation

Carla Di Cairano-Gilfedder

Robert J Briscoe

Arnaud Jacquet

Gilbert Kwame Owusu

Anthony Conway

Beum Seuk Lee

Kin Poon

Kjeld Jensen

David Lesaint

Gilbert Owusu

Siddhartha Shakya

Anne Liret

Mathias Kern

Timothy Glover

Raphael Dorne

Anne Linet

Rupal Mandania

Jiyin Liu

Samuel Gaffney

Pierre Desport

Frederic Lardeux

Chenlu Ji

Beum Seuk-Lee

Robert John Briscoe

Alessandro Salvatori

A computer implemented method of scheduling maintenance tasks for a target device, the device having a unique device type, comprising: retrieving, from a sequential transactional database, a series of transactions for each of a plurality of devices having the device type, each transaction corresponding to a change of state of one of the other devices; training a recurrent neural network based on each retrieved series of transactions as a predictor of a subsequent change of state of a device having the device type; retrieving, from a sequential transactional database, a series of transactions for the target device, each transaction corresponding to a change of state of the device; predicting a next maintenance task for the target device based on the series of transactions for the target device applied to the recurrent neural network; scheduling the predicted next maintenance task for the target device.

Automated device maintenance

A computer implemented method to specify a software defined network (SDN) for deployment by an SDN controller, the SDN controller being adapted to configure a plurality of physical network components to deliver communication services, the method including generating a model of the physical network components by accessing each of a plurality of data sources; defining the model including an indication of a location, type and interconnections of physical components based on the data sources; associating each indication in the model with a degree of confidence of accuracy of the indication based on the degree of confidence information from the data sources; accessing a set of rules defining relationships between types of physical component; refining the model based on the rules including adjusting a degree of confidence of accuracy of indications in the model based on satisfaction of the rules; and selecting a subset of network components in the refined model for inclusion in the SDN specification.

Software defined network specification

A data management process determines, from user-implemented provisional reservations () for data processing resources, a projected total capacity requirement for each said data processing resource, by maintaining a record () recording previous such reservations made by each user and comparing each reservations with records () of the actual resources used, to provide an estimate of resources required to meet the projected capacity requirement, and to provide data for a demand management processor (), which control associated configurable data processing equipment () to provide the resources required to meet the estimated capacity required. The process takes account of over- and under-ordering of capacity by comparing each reservation () with the use actually made (), and includes a record () of ad-hoc (unreserved) usage.

Network resource management

A computer implemented method of adapting a software defined network (SDN), the SDN operating with a set of network appliances in communication via network connections and the SDN comprising a control component operable in communication with at least a subset of the network appliances to control communication via the SDN, the method including receiving a representation of the logical arrangement of the SDN including a definition of appliances configured to provide communication for the SDN and network connections therebetween; receiving a measure of a performance metric for the SDN; receiving a forecast demand for the SDN; using a heuristic search to search a library of possible adaptations to the logical arrangement to identify a sequence of adaptations of the SDN that, when implemented, satisfy the forecast demand for the SDN and provide an improved measure of the performance metric of the SDN; applying the identified sequence of adaptations to the SDN by modifying the logical arrangement for use by the control component in subsequent communication via the SDN.

Dynamic network adaptation

A method for deploying virtualized network functions (VNFs) as virtualized implementations of logical network devices in a virtualized computing environment for providing network services, the method including determining a VNF deployment plan including a specification, for each point in time over a planning period, of zero or more VNF instances to be released at the point in time and zero or more VNF instances to be deployed at the point in time, wherein the deployment plan is determined by an optimization function based on: a constraint on a number of VNFs in all states of undeployed, being deployed, deployed and being release; constraints based on a demand for VNFs defined by characteristics of the VNFs and the virtualized computing environment; and constraints based on a capacity of the virtualized computing environment to accommodate a resource consumption of the VNFs, and wherein the optimization function is configured to minimize one or more of: a number of VNFs in a deployed state; and a resource cost of deploying and releasing VNFs; and executing the VNF deployment plan to deploy VNFs for the virtualized computing environment.

Virtualized network function deployment

A method of dimensioning a link in a telecommunications network, and a device for implementing the method, the method including determining, for a first plurality of sets of traffic observations, wherein each set of the first plurality of sets includes a series of traffic observations each indicating the amount of traffic on a link within an observation interval of a particular duration for that set, a peak value of a traffic observation of the series of traffic observations within each set of the first plurality of sets and an average value of the series of traffic observations within each set of the first plurality of sets; calculating a first peak to average ratio for each observation interval duration based on the determined peak and average values within each set of the first plurality of sets; determining a goodness of fit value for a power function of the first peak to average ratio against its respective observation interval duration; selecting a dimensioning interval duration based on the determined goodness of fit value; and estimating a capacity of the link based on the selected dimensioning interval duration.

Telecommunications network

A method for designating a class of data content for determining how the content will be handled by a content delivery network (CDN) into one of: a transport designation by which content is transported from an entrance node to one or more edge nodes; a caching designation by which the content is stored at one or more edge nodes; and a hybrid designation by which content is stored at a node in the CDN for subsequent transport to one or more edge nodes in the CDN, the method comprising: determining a set of one or more transport designation options; determining a set of one or more caching designation options; determining a set of one or more hybrid designation options; determining a normalised measure resource requirement for each of the transport, caching and hybrid designation options; and selecting one designation option for the class of content from all of the transport, caching and hybrid designation options by an optimisation function for selecting the designation option based on the determined normalised resource requirement for each option.

Content designation for a content delivery network

Methods and apparatus are disclosed for signalling congestion being caused by data items such as packets, received at a network element such as a router, in a communications network such as the Internet, or being caused by items otherwise requiring service or capacity from a shared resource. Preferred embodiments of the method involve identifying whether or not received data items received at a network element are capable of carrying congestion indications such as ECN marks, and for those that are capable, assigning congestion indications to the data items in dependence on a queue length characterization based on a substantially current, instantaneous measurement of the length of the queue, whereas for those that are not capable, a sanction such as dropping may be applied in dependence on a different queue length characterization based on a weighted moving average of current, recent, and less recent measurements of the length of the queue.

Signalling congestion

Growing community of inventors

London, United Kingdom

Carla Di Cairano-Gilfedder