Scalable optical-core network

Abstract

A network comprising a large number of electronic edge nodes interconnected through bufferless optical switch planes so that a signal from any edge node to any other edge node traverses only one switch plane scales to a capacity of hundreds of petabits per second while providing global geographic coverage. Each edge node is time-locked to each optical switch plane to which it connects to enable loss-free time-sharing of the network core despite the absence of buffers in the core. In an alternate implementation, a relatively small number of electronic switch units may be employed in a predominantly-optical core. In addition to scalability and high performance, the simple structure of the network significantly simplifies addressing and routing.

What is claimed is: 
     
         1 . A method of switching through a switch plane in a network, the method comprising:
 switching signals through a plurality of first optical switch units;   switching said signals through a plurality of second optical switch units, each of said first optical switch units having a first reconfiguration rate greater than a second reconfiguration rate of each of said second optical switch units; and   switching said signals through a plurality of third optical switch units, each of said third optical switch units having a third reconfiguration rate greater than said second reconfiguration rate.   
     
     
         2 . The method of  claim 1  wherein said first reconfiguration rate is at least an order of magnitude greater than said second reconfiguration rate and said third reconfiguration rate is at least an order of magnitude greater than said second reconfiguration rate. 
     
     
         3 . The method of  claim 2  further comprising a process of periodic reconfiguration of at least one of said second optical switch units based on configuration-change instructions received from a global reconfiguration server. 
     
     
         4 . The method of  claim 2  further comprising performing second-order time-slot matching processes to schedule connections for said signals through said first optical switch units and said third optical switch units. 
     
     
         5 . The method of  claim 2  comprising, prior to said switching through said plurality of first optical switch units, performing a time-alignment of said signals at said first optical switch units. 
     
     
         6 . The method of  claim 2  further comprising:
 integrating each said first optical switch unit with a respective third optical switch unit to form an integrated switch unit; and 
 exchanging time indicators between a controller of said integrated switch unit and controllers of a respective set of edge nodes to align signals received from said set of edge nodes at input ports of said integrated switch unit