System and method for hierarchical multicast routing in ATM networks

• US 5,831,975 A
• Filed: 04/04/1996
• Issued: 11/03/1998
• Est. Priority Date: 04/04/1996
• Status: Expired due to Fees

• Current Assignee: Lucent Technologies, Inc. (Nokia Corporation)
• Original Assignee: Lucent Technologies, Inc. (Nokia Corporation)
• Inventors: Venkateswaran, Ramanathan, Kumar, Vijay Pochampalli, Chen, Xiaoqiang, Raghavendra, Cauligi Srinivasa
• Primary Examiner(s): NGUYEN, CHAU T
• Application Number:
  US08/627,637
• Time in Patent Office: 943 Days
• Field of Search: 370/255, 370/256, 370/408, 370/254, 370/407, 370/400, 370/395, 370/389, 370/390, 370/471, 395/200.72
• US Class Current: 370/256
• CPC Class Codes: H04L 12/185 with management of multicast group membership
• CPC Class Codes: H04L 12/1877 Measures taken prior to transmission
• CPC Class Codes: H04L 2012/5619 Network Node Interface, e.g. tandem connections, transit switching
• CPC Class Codes: H04L 2012/563 Signalling, e.g. protocols, reference model
• CPC Class Codes: H04L 2012/5642 Multicast/broadcast/point-multipoint, e.g. VOD
• CPC Class Codes: H04L 45/10 Routing in connection-oriented networks, e.g. X.25 or ATM
• CPC Class Codes: H04L 45/16 Multipoint routing
• CPC Class Codes: H04L 45/48 Routing tree calculation
• CPC Class Codes: H04Q 11/0478 Provisions for broadband connections

Abstract

The present invention extends the PNNI protocols to support hierarchical multicast routing and signaling for ATM networks. The invention utilizes an extension to a core-based tree algorithm. Instead of a single core node, core nodes are maintained in each peer-group and at each level of the hierarchy. The advantage of this is that one single core node is not overloaded. Additionally, this increases fault-tolerance because there are no single points of failure. As would be understood, the present invention is highly scaleable because of the hierarchical nature of PNNI. In addition, the scheme supports multiple senders and dynamic membership changes to the multicast group. Quality of service requirements can be negotiated during connection setup and are guaranteed during the course of the connection. Though some additional topological information has to be flooded in the peer-groups to compute efficient multicast routes, the overheads to the connection management are minimal. The multicast tree is built incrementally and the cost of the tree is comparable to the cost of the Steiner Tree computed using some standard heuristics.

Claims

1. A method for multicasting cells in a communications network, said communications network including a plurality of nodes coupled to one another by links, said method comprising the steps of:

dividing said communications network into a hierarchical arrangement of peer groups including logical peer groups representative of a collection of nodes at a lower level of said hierarchical arrangement, wherein a peer group includes at least one of said nodes therein;

building a multicast tree for a multicast group of nodes in said network which includes all participant nodes involved in a multicast, wherein a participant node is either a sender or receiver of data for said multicast group, said step of building including the steps of;

selecting core nodes for each of said peer groups within said multicast group, wherein a node wanting to become part of said multicast group must register with a core node in its peer group;

flooding core node identity information locally within each of said peer groups including said logical peer groups, wherein said nodes of a peer group need only maintain said identity information about said core nodes of direct ancestor peer groups; and

selecting a peer group leader for each of said peer groups in said network for aggregating topology information of nodes in said peer group and flooding said topology information in higher level peer groups, said peer group leader flooding collected topology information from higher level peer groups into lower level peer groups, said collected topology information including a list of logical core nodes of ancestor peer groups for each said multicast group, wherein said peer group leader may be a different node than said core node in a peer group,

wherein said cells are able to be efficiently multicast by way of said multicast tree to said nodes in said multicast group.

2. The method of claim 1, wherein said communications network is an ATM network and said cells are ATM cells.

3. The method of claim 1, wherein a core node is active if a participant node is in the same peer group to which the core node belongs, and further including requiring all active core nodes of a specific multicast group to be part of said multicast tree for said specific multicast group.

4. The method of claim 3, further including requiring a core node to be part of said multicast tree if an intermediate non-participant node on the multicast tree belongs to that same peer group.

5. The method of claim 4, further including pruning a corresponding core node and corresponding intermediate non-participant nodes from said multicast tree if there is no participant node in the corresponding peer group, and if said pruning does not disconnect said multicast tree.

6. The method of claim 1, wherein said topology information includes node and link state information.

7. The method of claim 1, further including attaching a non-core node wanting to join said multicast group to said core node within its peer group, wherein said core node joins the core node of a parent peer-group on said multicast tree in a recursive fashion if said core node is not part of said multicast tree.

8. The method of claim 1, wherein a non-core participant node leaving said multicast tree removes itself and a link between said non-core participant node and its neighbor, provided said non-core participant node has exactly one neighbor;

wherein if said non-core participant node has more than one neighbor, it remains on said multicast tree as a non-participant node, and a core node with no participant nodes in its peer group leaving said multicast tree is required to prune itself therefrom provided said pruning does not disconnect said multicast tree.

9. The method of claim 1, further including the steps of:

sending a SETUP message from a node wanting to join said multicast group towards a core node within its peer group; and

reserving resources on a link for a connection, when a SETUP message passes on said link from one node to another, a path to said core node being expressed in terms of a designated transit list (DTL), where said DTL is an approximate path to be followed by signaling messages.

10. The method of claim 9, wherein said DTL created by a source can be modified by an ingress node to the peer-group, an egress node to the peer-group and a core node of the peer-group.

11. The method of claim 9, further including sending a RETRACE message, wherein said RETRACE message traverses only on links already traversed by a SETUP message in an opposite direction.

12. The method of claim 9, further including sending a CONNECT message, wherein said CONNECT message is sent by a node to which a joining node attaches, said CONNECT message traversing on all the links on which resources are reserved.

13. The method of claim 9, further including sending a RELEASE message, wherein said RELEASE message may be sent by any intermediate nodes, in case a connection cannot be established, said RELEASE message also being sent to terminate a connection, wherein said RELEASE message traverses on all the links on which resources are reserved.

14. The method of claim 9, wherein upon receiving said SETUP message, if the node is not on the multicast tree, said node checks the DTL and forwards a message to a next node in the DTL with a pointer, wherein if said node is an egress border node, said node saves a list of the nodes visited by said SETUP message and forwards said list across a border link, and wherein if said node is an ingress border node, said node computes a path to a core node of the peer-group, said path being converted to a new DTL and pushed onto a stack, wherein said SETUP message is then forwarded according to the new DTL towards the core node.

15. The method of claim 9, wherein upon receiving said SETUP message, if the node is a core node, said core node first checks if it is active, wherein if not active, said core node computes a path to a next peer-group in the DTL, wherein said path is converted to a new DTL and pushed onto a stack, and wherein if any node on said path has already been visited by said SETUP message, said SETUP message is changed to a RETRACE message.

16. The method of claim 15, wherein entries in the DTL are removed until a first entry on the top of said stack is a visited node and no other entry in the DTL is a visited node, said RETRACE message being forwarded towards a node from which said SETUP message was received, wherein if there are no visited nodes in said path, said SETUP message is forwarded as per said new DTL.

17. The method of claim 9, wherein upon receiving said SETUP message, if a core node is not active and the DTL stack is empty, said core node forwards said SETUP message towards a core node of a parent peer-group, wherein a path to a core node of said parent peer-group and a corresponding new DTL is computed and a check for a visited node is made, wherein if a visited node is found, said SETUP message is changed to a RETRACE message and the DTL is modified, and wherein if no visited nodes are found, the SETUP message is forwarded as per the new DTL.

18. The method of claim 11, wherein upon receiving a RETRACE message, if a first entry on top of the DTL stack does not match with a node ID or an ancestor ID, the RETRACE message is forwarded towards a node from which it received the SETUP message, wherein if the entry matches, a path to a next entry as per the DTL is computed, a check for visited nodes being made, wherein if there is a visited node, the DTL is modified and the RETRACE message is forwarded towards the node from which it received the SETUP message and if there is no visited node, the RETRACE message is changed to a SETUP message and forwarded to the next node as per the new DTL.

19. The method of claim 12, wherein for said CONNECT message, resources that are reserved on a link are committed, the node forwarding said message on all the links on which a SETUP was received, with the exception of a link on which said CONNECT is received.

20. The method of claim 13, wherein for said RELEASE message, resources that are reserved on a link are released, the node forwarding said message on all the links on which a SETUP was received/sent, with the exception of a link on which said RELEASE message is received.

21. A shared communications network for multicasting cells, said communications network including a plurality of nodes coupled to one another by links, said network comprising:

a hierarchical arrangement of peer groups making up said communications network, wherein a peer group includes at least one of said nodes therein;

one or more multicast trees for a multicast group of nodes in said network, which multicast trees include all participant nodes involved in a multicast, wherein a participant node is either a sender or receiver of data for said multicast group;

a core node selected for each of said peer groups within said multicast group, wherein a node wanting to become part of said multicast group must register with said core node in its peer group, wherein core node identity information is flooded locally within each of said peer groups, and said nodes of a peer group need only maintain said identity information about said core nodes of direct ancestor peer groups; and

a peer group leader for each of said peer groups in said network for aggregating topology information of nodes in said peer group and flooding said topology information in higher level peer groups, wherein a list of logical core nodes of ancestor peer groups is flooded in a peer group by each said peer group leader, wherein said peer group leader may be a different node than said core node,

wherein said cells are able to be efficiently multicast by way of said multicast tree to said nodes in said multicast group.

22. The network of claim 21, wherein said communications network is an ATM network and said cells are ATM cells.

23. A method for multicasting cells in a communications network, said communications network including a plurality of nodes coupled to one another by links, said method comprising the steps of:

dividing said communications network into a hierarchical arrangement of peer groups, wherein a peer group includes at least one of said nodes therein;

building a multicast tree for a multicast group of nodes in said network which includes all participant nodes involved in a multicast, wherein a participant node is either a sender or receiver of data for said multicast group, said step of building including the steps of;

selecting core nodes for each of said peer groups within said multicast group, wherein a node wanting to become part of said multicast group must register with a core node in its peer group;

flooding core node identity information locally within each of said peer groups, wherein said nodes of a peer group need only maintain identity information about said core nodes of direct ancestor peer groups; and

selecting a peer group leader for each of said peer groups in said network for aggregating topology information of nodes in said peer group and flooding said topology information in higher level peer groups, wherein a list of logical core nodes of ancestor peer groups is flooded in a peer group by each said peer group leader.