Tropic Networks Inc.

Cisco Technology, Inc.

Alcatel-lucent Canada Inc.

Tropic Network Inc.

Byoung-Joon Lee

Sudhakar Ganti

Walter Carpini

Udo Mircea Neustadter

Cuong Tu Dang

Anand Srinivasan

Vincent Chi Chiu Wong

James Benson Bacque

Steve Joseph Grabner

Hans Frederick Johnsen

Richard Charles Vieregge

Barry Mark

John Willem Van Gulik

Methods, systems and apparatus are provided which prevent frame mis-ordering in an explicitly routed network when switching from a first path to a second path. In a first embodiment, last frame and first frame markers are transmitted, and a receiver buffers all frames received between receipt of the last frame marker and the first frame marker. In a second embodiment, a last frame marker is sent to which an acknowledgement must be received before transmitting on the second path. In a third embodiment, only a first frame marker is sent. The receiver, upon receipt, signals a network card for the first path to stop forwarding frames, also buffers frames for some period. In a fourth embodiment, the merge point node forwards frames over different paths with unique markings so the outgoing card can distinguish them.

Methods and systems preventing frame mis-ordering in explicitly routed networks

A failure protection between interconnected adjacent Resilient Packet Rings (RPRs) in a multiple RPR network is provided. Two paths, a regular message path and a protection path, are provided between two adjacent RPRs. The regular path is used for routing inter-ring messages when no failure has occurred on the path. Messages are rerouted through the protection path when a failure occurs on the regular path. Each of these paths has two RPR interface nodes (one for each RPR) that are connected to an interconnection device (a layerbridge or a layer-router) through interconnection links. Procedures for detecting failures and generating notifications for message rerouting and fault reports are executed at the interconnection devices. The procedures use periodic keep alive messages for diagnosing network segment and interconnection device failures. The fault detection and message rerouting are accomplished in less than 50 ms.

Protection system and method for resilient packet ring (RPR) interconnection

Cascaded policing methods and systems are provided which allow lower priority traffic to benefit from otherwise unused capacity allocated to higher priority traffic of a given customer/service with multiple classes of service. The method involves policing packets of a first class in accordance with at least one policing parameter associated with the first class, and policing packets of a second class in accordance with at least one policing parameter associated with the second class in a manner which gives to the second class at least a portion of a traffic throughput afforded to the first class by at least one of said at least one policing parameter, such as a rate guarantee or burst tolerance, associated with the first class of traffic which is not being used by the packets of the first class. The method is easily adapted to an arbitrary number of different traffic classes.

Cascaded policing systems and methods

A routing system and method are provided which use a composite cost in identifying routes. This allows a simple way of identifying the best route taking multiple metrics into account for each link simultaneously. The system allows for the inclusion of pruning constraints, and allows for various objectives such as bin packing or load balancing to be achieved.

Multi-constraint routing system and method

The present invention provides a method, apparatus and article of manufacture for increasing network processing node interconnect capacity and reducing maximum hop count in a scalable multidimensional ring network by creating additional rings. Initially a node identification algorithm is selected and an initial network processing node in the scalable multidimensional ring network is selected as a first node in a new ring. The node identification algorithm is applied to the selected node to calculate a subsequent node in the new ring. The calculated node is then made the selected node. The applying and selecting steps terminate when applying the node identification algorithm to the selected node results in the calculated subsequent node being equal to the initial node, thereby creating the new ring. A new initial node is then selected and the new ring creation process continues. When a new initial node is selected that is already a member of a new ring the entire process terminates, thereby creating all the new rings in the new scalable multidimensional ring network.

Scalable multidimensional ring networks

Methods and components are provided which ensure both the improved use of network resources and adequate performance of best effort (BE) traffic by intelligently distributing the BE traffic demands at connection level with corresponding scaling weights, and without reserving bandwidth. A weighted sum of the best effort (BE) class connections (or LSPs in MPLS context) in a link is used as a path selection criterion, where each BE connection is weighted by its service volume. Path selection for a requested BE service volume is performed by creating a virtual topology in which all links in a network have weighted BE metrics updated to include the effects of the requested BE service volume, and identifying a best path through the virtual topology taking into account the weighted BE metrics.

Growing community of inventors

Nepean, Canada

Byoung-Joon Lee