Modular high-capacity switch

Abstract

A modular optical switch includes a set of optical switch modules connected in a mesh, a master controller for the whole optical node and a switch-module controller for each of the optical switch modules. The optical switch modules receive optical signals from, and transmit optical signals to, edge nodes based on connection requests received from the edge nodes. The master controller acts to select a path, using a simple or compound time-slot matching process, through the mesh of switch modules for each optical signal related to a connection request. Advantageously, the optical switch modules are fast switching, enabling the use of time-sharing schemes such as TDM, and the modular optical core node is made practical by efficient path selection at the master controller. A hybrid modular switch may include both optical and electronic switch modules, a master controller, and a switch-module controller for each of the switch modules.

What is claimed is: 
     
         1 . A switching node, comprising:
 a plurality of inlet ports;   a plurality of outlet ports;   a plurality of inner links; and   a scheduler configured to schedule transfer of data, in a specified number of time slots, from a particular inlet port to a particular outlet port by:
 designating a route set for a combination comprising the particular inlet port and the particular outlet port, the route set comprising at least one route traversing two of the inner links between the particular inlet port and the particular outlet port; 
 selecting a candidate time slot; and 
 for the candidate time slot, selecting a candidate route in the designated route set by:
 determining, from a calendar associated with the particular inlet port, an occupancy state of the particular inlet port for the candidate time slot; 
 determining, from a calendar associated with the particular outlet port, an occupancy state of the particular outlet port for the candidate time slot; 
 determining, from a calendar associated with a first inner link in said candidate route, an occupancy state of the first inner link in the candidate route for the candidate time slot; and 
 when the occupancy state of the particular inlet port, the particular outlet port and the first inner link in the candidate route are each determined as vacant:
 considering the candidate route to be a first available route; and 
 considering the candidate time slot to be an allocable time slot. 
 
 
   
     
     
         2 . The switching node of  claim 1 , wherein the scheduler is configured to designate a respective route set for each of a plurality of combinations of a respective inlet port and a respective outlet port, each route set comprising at least one route traversing two of the inner links. 
     
     
         3 . The switching node of  claim 1 , wherein the scheduler is configured to select candidate routes by considering routes in the designated route set in a cyclic order. 
     
     
         4 . The switching node of  claim 3 , wherein the scheduler is further configured to associate a respective vacancy value with each of the candidate routes, the respective vacancy value for a route being a total number of vacant time slots per time frame for the route. 
     
     
         5 . The switching node of  claim 4 , wherein the scheduler is further configured:
 to repeat the selection of a candidate route and the determination of occupancy states of the particular input port, the particular outlet port and a first inner link on the candidate route to find a second available route from said candidate routes;   to associate a second vacancy value with the second available route; and   to select said second available route when the second vacancy value associated with the second available route is greater than the vacancy value associated with the first available route.   
     
     
         6 . The switching node of  claim 5 , wherein the scheduler is further configured to repeat the selection of candidate time slots in a predetermined order, the determination of occupancy states and the selection of candidate routes as long as the specified number of time slots exceeds the number of allocatable time slots. 
     
     
         7 . The switching node of  claim 6 , wherein the predetermined order is a sequential order of time slots in a time frame having a predetermined number of time slots. 
     
     
         8 . A data structure for simple and compound time-slot matching over a number of time slots to be considered, the data structure for use in a switch module in a modular switch comprising a plurality of switch modules having inlet ports and outlet ports, the data structure comprising:
 a first matrix having:
 a number of rows equal to a first product of a maximum number of said inlet ports and a maximum number of switch modules; and 
 a number of columns equal to said number of time slots; 
   a second matrix having:
 a number of rows equal to a second product of a maximum number of said outlet ports and a maximum number of switch modules; and 
 a number of columns equal to said number of time slots; 
   a third matrix having:
 a number of rows equal to a third product of said maximum number of said inlet ports and a maximum number of time slots to be considered; and 
 a number of columns equal to a maximum number of said outlet ports; 
   a fourth matrix having:
 a number of rows equal to a fourth product of said maximum number of said outlet ports and a maximum number of time slots to be considered; and 
 a number of columns equal to said maximum number of inlet ports. 
   
     
     
         9 . The data structure of  claim 8  wherein each element in each of the matrices is a 1-bit word indicating an occupancy state. 
     
     
         10 . A method of using the data structure of  claim 8  for first-order time-slot matching, the method comprising:
 finding an entry in a row in said first matrix indicating a vacant occupancy state for an inlet port of a given one of said plurality of switch modules; and 
 finding an entry in a row in said second matrix indicating a vacant occupancy state for an outlet port of said given one of said plurality of switch modules. 
 
     
     
         11 . A method of using the data structure of  claim 8  for second order time-slot matching, the method comprising:
 finding an entry in a row in said first matrix indicating a vacant occupancy state for an inlet port of a first one of said plurality of switch modules; 
 finding an entry in a row in said second matrix indicating a vacant occupancy state for an outlet port of a second one of said plurality of switch modules; and 
 finding an entry in a row in said third matrix indicating a vacant occupancy state for an inner link to said second one of said plurality of switch modules. 
 
     
     
         12 . A method of using the data structure of  claim 8  for third-order time-slot matching, the method comprising:
 finding an entry in a row in said first matrix indicating a vacant occupancy state for an inlet port of a first one of said plurality of switch modules; 
 finding an entry in a row in said second matrix indicating a vacant occupancy state for an outlet port of a second one of said plurality of switch modules; 
 finding an entry in a row in said third matrix indicating a vacant occupancy state for an inner link to a third one of said plurality of switch modules; and 
 finding an entry in a row in said fourth matrix indicating a vacant occupancy state for an inner link to said second one of said plurality of switch modules.