Switching system with distributed switching fabric

Abstract

A switch encapsulates incoming information using a header, and removes the header upon egress. The header is used by both distributed ingress nodes and within a distributed core to facilitate switching. The ingress and egress elements preferably support Ethernet or other protocol providing connectionless media with a stateful connection. Preferred switches include management protocols for discovering which elements are connected, for constructing appropriate connection tables, for designating a master element, and for resolving failures and off-line conditions among the switches. Secure data protocol (SDP), port to port (PTP) protocol, and active/active protection service (AAPS) are all preferably implemented. Systems and methods contemplated herein can advantageously use Strict Ring Topology (SRT), and conf configure the topology automatically. Components of a distributed switching fabric can be geographically separated by at least one kilometer, and in some cases by over 150 kilometers.

1 . A network switch for routing packets of information, comprising: 
 a plurality of ingress switching elements, each with a plurality of input and output ports;    hardware that adds a routing header to the packets entering the input ports of the plurality of ingress elements a plurality of egress switching elements, each with a plurality of input and output ports;    a backbone that provides active switching among the plurality of ingress and egress elements; and    wherein each of the switching elements is adapted to use the routing header to pass the packets through the backbone from one of the ingress switching elements to at least one of the egress switching elements to which the one ingress element is not otherwise directly connected.    
     
     
         2 . The switch of  claim 1  wherein each of the pluralities of ingress and egress elements support a protocol providing connectionless media with a stateful connection.  
     
     
         3 . The switch of  claim 2  wherein the stateful connection comprises Ethernet.  
     
     
         4 . The switch of  claim 1  wherein each of the pluralities of ingress and egress elements has at least 8 input ports and 8 output ports.  
     
     
         5 . The switch of  claim 1  wherein at least one of the pluralities of ingress elements has a capacity of at least one gigabit/sec.  
     
     
         6 . The switch of  claim 1  wherein at least one of the pluralities of ingress elements has a capacity of at least ten gigabit/sec.  
     
     
         7 . The switch of  claim 1 , having hardware that executes a management discovery protocol that determines from time to time which of the plurality of ingress and egress elements becomes a master element.  
     
     
         8 . The switch of  claim 1 , having hardware that executes a management discovery protocol that determines which element becomes a new master on the failure of the existing master using a heart beat protocol.  
     
     
         9 . The switch of  claim 7 , wherein each of the plurality of ingress and egress elements executes the management discovery protocol.  
     
     
         10 . The switch of  claim 1 , further comprising first hardware that encapsulates packets of information with a routing header that is used to route the packets within the switch, and second hardware that removes the routing header.  
     
     
         11 . The switch of  claim 10 , wherein the routing header includes a source element address, destination element address, and a destination port address.  
     
     
         12 . The switch of  claim 10 , wherein at least some of the pluralities of egress elements are logically coupled in a cluster and the routing header includes a destination cluster address.  
     
     
         13 . The switch of  claim 10 , wherein the plurality of ingress and egress elements are coupled using a Strict Ring Topology (SRT).  
     
     
         14 . The switch of  claim 1 , further comprising an element manager unit that sends element messages to at least some of the plurality of ingress and egress elements.  
     
     
         15 . The switch of  claim 14 , wherein at least one of the element messages implements a secure data protocol (SDP) that performs ACK/NAK function on the messages.  
     
     
         16 . The switch of  claim 14 , wherein at least one of the element messages implements a port to port (PTP) protocol.  
     
     
         17 . The switch of  claim 14 , wherein at least one of the element messages implements an active/active protection service (AAPS).  
     
     
         18 . The switch of  claim 14 , further comprising hardware that maintains a list of recent element messages.  
     
     
         19 . A switching system comprising first and second clusters of switches according to  claim 1 .  
     
     
         20 . The switching system according to  claim 19 , wherein inter-cluster communication is implemented via dynamic VLAN tunnels.  
     
     
         21 . The switching system according to  claim 19 , wherein inter-cluster communication is implemented via a PTP protocol based matrix of link addresses.  
     
     
         22 . The switching system according to  claim 19 , wherein the first and second clusters are separated by at least 1 kilometer.  
     
     
         23 . The switching system according to  claim 19 , further comprising a third cluster, and where the first, second and third clusters execute an automatically configuring topology.  
     
     
         24 . A method of routing Ethernet packets, comprising: 
 providing a plurality of switch ingress modules that encapsulate the Ethernet packets into frames having intra-system routing headers;    providing a plurality of egress switch modules that remove the headers from the packets;    providing a core that actively routes the frames between ingress and egress modules, and among core elements;    wherein at least one of the ingress and one of the egress modules are distanced from one another by at least 1 km, and at least two of the core elements are distanced from one another by the distance of a least 1 km.    
     
     
         24 . The system of  claim 24 , further comprising the core connecting at least 16 ports among the plurality of ingress and egress modules.  
     
     
         25 . The system of  claim 24 , further comprising providing an optical carrier as a component of the core.  
     
     
         26 . The system of  claim 24 , further comprising operating a link aggregation across the entire system that conforms to an IEEE standard.  
     
     
         27 . The system of  claim 24 , further comprising operating a virtual LAN (VLAN) across the entire system without a need for a VLAN trunk or VLAN tagging.  
     
     
         28 . The system of  claim 24 , further comprising providing a protocol that at least temporarily prevents data from being sent to a selected one of the ports.  
     
     
         29 . The system of  claim 24 , further comprising implementing layer 3 (IP) routing among the plurality of modules.  
     
     
         30 . The system of  claim 24 , further comprising providing the plurality of modules with port based network access control capability that conforms with an IETF standard.  
     
     
         31 . The system of  claim 24 , further comprising, providing the plurality of modules with detection of layer 2 (Ethernet) to layer 7 (Application) data, and decisions on dynamic routing or handling of said data.  
     
     
         32 . The system of  claim 24 , further comprising sending management messages to at least some of the plurality of modules, wherein at least one of the messages implementing a secure data protocol (SDP) that using positive acknowledgement function on the messages, and at least another one of the element messages implementing a port to port (PTP) protocol, and at least another one of the element messages implementing an Ethernet frame based active/active protection service (AAPS).  
     
     
         33 . The system of  claim 24 , further comprising logically coupling at least some of the plurality of modules into first and second clusters, and separating the clusters by at least 1 kilometer.