Nortel Networks Corporation

Martin Biddiscombe

Paul Anthony Kirkby

Subramaniam Sabesan

Subramanian Sabesan

Nigel Lawrence Bragg

Clive Colin Hayball

Julian Frank Cable

Philip Davies

Glenn Denney

Gordon Bradley

A protocol data unit is routed along a path between a first node and a second node within a communication network. The protocol data unit carries routing information which specifies the path in the form of a list of identifiers of the links to be followed towards the second node. At each intermediate node, the link identifier for the link just traversed with a link identifier for the corresponding link in the reverse direction towards the first node. At any point along the path the protocol data unit carries information which allows it to be carried towards the second node, or to follow a reverse path towards the first node. This reverse path can be used to report a fault to the first node.

Routing of protocol data units within a communication network

DNS in its native form cannot identify a good or best server. Traditional Service Providers deploy a centralised approach to global traffic management based on enhancements to DNS, but do have the capability to augment this with edge-based server selection because they do not own/operate an edge network. Their resolution of DNS requests is typically restricted to identifying candidate servers within an edge domain rather than selecting the 'best' server. The invention proposes a method of handling Internet resource requests whereby a local lookup server receiving a request searches for a best server able to satisfy the resource request. The local lookup server finds the best server by searching a database for a resource record associated with the best server, and then uses the resource record to retrieve an identifier of a series of executable instructions that are executed to locate the best server. The resource requests and responses may hold additional information relating to operational characteristics which may used to determine the best server. The invention proposes a DNS record type comprising a user-defined field for conveying this additional information in the resource request and in the response. The invention also proposes architecture for handling the resource request comprising a network server that provides the requested resource using the best server, as well as a communications network comprising this architecture.

Apparatus and method for managing internet resource requests

A communication network comprises a plurality of nodes which exchange topology information. Some of the nodes are of a conventional type, which propagate topology information in an unrestricted manner while others of the nodes are of a radius-restricted type, which propagate topology information for a restricted distance within the network. Upon receiving a topology advertisement from another node, a node determines what type of node sent the advertisement, and adjusts its operation accordingly. A radius-restricted node forwards a topology advertisement with a metric which indicates the aggregate distance travelled by that advertisement. When a radius-restricted node receives a topology advertisement from a non radius-restricted node, it forwards the topology advertisement with a metric set at a value which will prevent other radius-restricted nodes from installing the topology advertisement. This prevents flooding of unnecessary information.

Restricted dissemination of topology information in a communication network

The Invention provides a distributed resource allocation Method. The distributed resource allocation method is based on the concept of proportional fairness. In this method, users express their willingnesses to pay to the local distributed controller which is associated with those users. The bandwidth allocated to each user is determined by the local controller with which they are associated and the allocation takes into account the relative demands on all the resources in the network and the time taken for those demands to reach the distributed local controllers. The distributed controllers each use only locally available information at their respective nodes, but take into account the delays between nodes. The locally available resource (i.e. bandwidth) is divided in proportion to the willingnesses to pay expressed by the various users entering their particular node. On a ring (and in other networks where the only place that traffic streams are multiplexed is at the periphery) this guarantees that no data will be lost (due to cell/packet collision) once it has been launched.

Method of allocating resources in a telecommunications network

The invention provides a resource allocation method which is based on the concept of proportional fairness. In this method, users express their willingnesses to pay for network resources. A controller determines how the users' willingnesses to pay are to be divided between the resources in order to determine the relative demands for the resources. Each resource is then divided between those users using it in proportion to how much they are willing to pay for the use of their share of it. The allocation takes into account the relative demands on all the resources in the network. A user may increase his share of the network resources allocated to him by increasing the value of his willingness to pay. The method may be applied either by a central controller (network manager), or it may be delegated to controllers (network element managers/customer premises equipment managers) distributed across the network.

Growing community of inventors

Harlow, United Kingdom

Martin Biddiscombe