Agilent Technologies, Inc.

Jonathan P R Lacey

Douglas Michael Baney

Brian Elliot Lemoff

Mark C Nowell

David G Cunningham

Steve Methley

A process for discovering a path from a source node to a destination node through a network by using 'collisions' of randomly-propagating 'feeler' packets originating from both the source node and the destination node. A discovered path is reported to the source node by the collision-detecting node where it may be stored and updated responsively to reports of new feeler packet collisions. Paths discovered and reported may be analyzed at either the collision-detecting node or the originating node to remove loops. The random collision-detecting path-discovery procedure reduces the operational traffic overhead associated with other exponentially-proliferating discovery methods. The feeler packets are propagated randomly through the network topology, thereby imposing relatively uniform path-discovery traffic effects in the network. Path discoveries arising from feeler-packet collisions always reflect current network topology and traffic conditions. The origination rate of feeler packets may be adjusted responsively to changes in demand, cost or other parameters.

Adaptive path discovery process for routing data packets in a multinode network

An optical switch, preferably an add/drop switch, includes a minimal number of two-state switching arrangements, thereby facilitating a reduction of switch complexity. Light pulses entering the switch via a given input port may be directed to either a corresponding output port or a corresponding drop port by the operation of the switching arrangements. Light pulses entering the switch via a given add port may be directed to a corresponding output port or may be disbursed by the operation of the switching arrangements. The switching arrangements are toggled between reflective and transmissive states. In the preferred embodiment, the two-state switching arrangements are controlled by the manipulation of index matching fluid within trenches. Additionally, in the preferred embodiment, the switching arrangements are configured such that a given light pulse interacts with only one trench, thereby minimizing the overall signal loss and loss non-uniformities within the switch.

Optical add/drop switch utilizing a minimal number of switching crosspoints

An optical switching node and method for operating same is disclosed. If a signal received over an optical fiber is destined for the switching node at which the signal is received, the signal is processed conventionally. If there is a signal on the fiber that is destined for a different optical switching node, then a node controller will determine whether that signal requires equalization and/or regeneration. A network manager instructs the node controller whether the signal requires wavelength conversion. In accordance with the invention, if the optical signal on the fiber is destined for a different optical switching node and requires none of equalization, regeneration and wavelength conversion, the optical signal remains in the optical domain and is switched directly to an appropriate output fiber. If the node controller and network manager determine that the optical signal that is destined for a different optical switching node requires any combination of equalization, regeneration, and/or wavelength conversion, then the signal is further processed in the optical switching node to provide those functions.

Growing community of inventors

Menlo Park, CA, United States of America

Jonathan P R Lacey