Abstract: A method and apparatus is provided for setting up an IP session from a first node, across a network that uses direct virtual path routing, to a second node. The first node receives a request to set up the session. The second node is identified as the destination of the session, based on the request. A first stack of labels is obtained that defines a first path. A second stack of labels is obtained that defines a second path. A first label identifier is selected that identifies the session in a first direction. A second label identifier is selected that identifies the session in a second direction. Finally the data is transmitted from the first node to the second node and from the second node to the first node using the first stack and second stack of labels and label identifiers, respectively.

Abstract: The problems associated with prior solutions for implementing ATM or IP/MPLS in a large-scale network are overcome by providing an efficient means by which capacity in the network is more fully shared without adversely affecting call or session setup latency, and at the same time simplifying network operations. This is achieved by performing the functions of route setup, route allocation, and capacity management in a network at the edges (e.g., the interfaces existing between each ATM switch or IP/MPLS router and customers) of the network. In accordance with the principles disclosed herein, the edges contain nodes that form the interface between the ATM switches and IP/MPLS routers and the link(s) that interconnect them and the outside world. These nodes comprise controllers and other apparatus that in some cases may be incorporated in, or connected as adjuncts to, the ATM switches or IP/MPLS routers.

Abstract: A method is provided for setting up a communication from a first edge node, across a network that uses direct virtual path routing, to a second edge node. The first edge node receives a request to set up the communication. The second edge node is identified as the destination of the communication, based on the request. A first virtual path identifier (VPI) is obtained that defines a first path from the first edge node across the network to the second edge node. A second VPI is obtained that defines a second path from the second edge node across the network to the first edge node. The first and second VPIs are selected by a routing status database. A first virtual channel identifier (VCI) within the first VPI is selected. A second virtual channel identifier (VCI) within the second VPI is selected. Data is transmitted from the first edge node to the second edge node using the first VPI and first VCI, and from the second edge node to the first edge node using the second VPI and second VCI.

Abstract: An efficient approach and corresponding apparatus for selecting virtual paths in a network for routing datagrams achieves effective sharing of capacity without adversely affecting call setup latency. Edge nodes of the network assign calls to virtual paths based on the destination of the call and the current load status of each of a relatively small number of paths. Each call is assigned a VPI (Virtual Path Identifier) corresponding to the path chosen and a VCI (Virtual Circuit Identifier) corresponding to the identity of the call at that edge node. The ATM backbone nodes route calls based solely on the VPI. Capacity management and load balancing is achieved by each edge node keeping track of the load on each access link from its edge node to the rest of the network and of the congestion status of the various links in the network.

Abstract: An efficient approach and corresponding apparatus for selecting virtual paths in a network for routing datagrams achieves effective sharing of capacity without adversely affecting call setup latency. Edge nodes of the network assign calls to virtual paths based on the destination of the call and the current load status of each of a relatively small number of paths. Each call is assigned a VPI (Virtual Path Identifier) corresponding to the path chosen and a VCI (Virtual Circuit Identifier) corresponding to the identity of the call at that edge node. The ATM backbone nodes route calls based solely on the VPI. Capacity management and load balancing is achieved by each edge node keeping track of the load on each access link from its edge node to the rest of the network and of the congestion status of the various links in the network.

Abstract: A message distribution system having multiple message processors provides integrated overload control and message distribution. An additional "dummy" message processor is established by the system, to which a certain fraction of messages is assigned as though it were a real processor, thereby accomplishing rejection of that fraction of messages. The fraction to be rejected is determined in real time based on the states of the various processors, which also is the basis for the allocation among the processors of those messages that are not rejected.