Corrigent Systems Ltd.

Zvika Menahemi

Gideon Agmon

Alex Levit

Ronen Solomon

David Zelig

Muly Ilan

A method for communication between first and second nodes via a communication network includes assigning multiple communication ports in two or more network nodes between the first and second nodes to serve as member ports of a multi-homing group (MHG). Each of the member ports is associated with a different, respective communication path between the first and second nodes via the communication network. A frame transmitted from the first node to the second node is accepted at the two or more network nodes. The frame has a header including header fields. A single port is selected from among the member ports by applying a hashing function to one or more of the header fields at each of the two or more network nodes. The frame is forwarded via the selected port over the respective communication path that is associated with the selected port.

Hash-based multi-homing

Routing tables of OSI layer-3 network elements are modified in order to enable entry to a RPR subnet at different entry points. This enables virtual tunnels or routing paths to utilize all existing entry links to the RPR subnet and to minimize cost factors, such as the number of RPR spans required to traverse the RPR subnet from the entry point to a destination RPR node. The routing tables of RPR subnet elements are modified, such that traffic leaving different elements but destined for the same network location outside the RPR subnet may have individualized RPR exit nodes. The respective RPR exit points for the network elements are chosen to minimize cost factors, such as the number of RPR spans required to reach the exit node from each RPR node.

Layer-3 network routing with RPR layer-2 visibility

According to a disclosed embodiment of the invention, software download and installation in a ring network are synchronized in a two-stage operation in order to minimize service disruption time. In a first phase, the MPM's and UIM's in all the nodes are upgraded in parallel. Each node has two main processing modules, allowing one to be upgraded while the other continues to operate the node. In a second phase, the RIM's at the edges of a single ring segment are upgraded, one ring segment after the other, during which time each terminal node continues to operate with respect to the span connecting to its opposite side. While a span is non-operative, traffic is wrapped or diverted as necessary to maintain service of the ring. While one RIM of a node is being upgraded, the other RIM remains operational.

Growing community of inventors

Modiin, Israel

Zvika Menahemi