JP2014007568A - Multicast forwarding system and multicast path switching method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To set up a bypass route for a multicast path without advanced function processing at relay nodes, such as signaling, when bypassing by protection the link where a failure occurred.SOLUTION: The multicast forwarding system comprises means in which: an Egress node transmits a fault notification to an Ingress node upon detecting that a CC packet multicast forwarded from the Ingress node did not arrive; the Ingress node deletes the Egress node that transmitted the fault notification from primary destination Egress nodes in a destination registration table, registers the Egress node as a secondary destination Egress node, and relabels both primary and secondary labels; and a relay node multicast forwards a multicast packet to relevant destination IFs according to a primary or secondary label of the multicast packet transmitted from the Ingress node, referring to a multicast forwarding table that associates labels with destination IFs.

Description

  The present invention relates to a multicast transfer system and a multicast path switching method related to protection against a relay node or relay link failure in a multicast transfer system that multicasts video or other data via a network.

FIG. 5 shows a configuration example of a conventional multicast transfer system.
In FIG. 5, the multicast forwarding system shown here has a tree configuration in which four Egress nodes EN1 to EN4 are connected from one Ingress node IN via two relay nodes RN1 and RN2. Egress nodes EN1 to EN2 are connected to relay node RN1, and Egress nodes EN3 to EN4 are connected to relay node RN2. Furthermore, a standby Ingress node IN ′, relay nodes RN1 ′ and RN2 ′, and egress nodes EN1 ′ to EN4 ′ are provided for detouring the failure section by protection when a failure occurs.

   Each of the Ingress node IN and the relay nodes RN1 and RN2 includes a multicast forwarding table that forwards only to the delivery destination that has requested viewing. For example, when EN1, EN2, EN3, and EN4 transmit a viewing request for channel # 1, EN1 and EN2 are set as distribution destination IFs in channel # 1 of the multicast forwarding table of RN1, and the multicast forwarding table of RN2 In the channel # 1, EN3 and EN4 are set as distribution destination IFs. Then, RN1 and RN2 are set as distribution destination IFs in channel # 1 of the IN multicast forwarding table.

  In recent years, MPLS-TE (Multiprotocol Label Switching-Traffic Engineering) technology capable of realizing high-reliability and high-quality communication by protection has attracted attention, and application of this technology has also been realized in multicast distribution. As a protection method in a network using the MPLS-TE technique, there are an FRR (Fast Reroute) method and a path protection method. In the FRR method, a detour LSP is set for all links or nodes to make a detour as close as possible to the fault location (Non-Patent Document 1). On the other hand, in the path protection method, a bypass path is set by end-to-end, and the Ingress node IN that receives the failure notification switches to the bypass path.

  As one of the path protection methods, as shown in FIG. 6, there is a method in which only a failed Sub-LSP is switched to a detour path by performing signaling for each Sub-LSP and detecting a signaling error at the time of the failure. It has been proposed (Patent Document 1).

JP 2008-206050 A

RFC4090 Fast Reroute Extensions to RSVP-TE for LSP Tunnels

  In the FRR method, in order to protect all the routes of the multicast path, detour paths are set for all relay links or relay nodes, and monitoring by signaling and OAM and segment protection according to the failure are necessary. The processing load on the node increases.

  On the other hand, in the path protection method, an end-to-end detour path is set, so switching processing can be performed only by switching the transmission end at the Ingress node, but there is no influence of failure by switching the multicast path. There is a problem that a momentary interruption due to switching to the route occurs.

  In response to this problem, the method proposed in Patent Document 1 detects PathErr and the like in RSVP-TE signaling at the relay node, notifies the Ingress node, and bypasses only the Sub-LSP affected by the failure at the Ingress node. Switching to LSP eliminates instantaneous interruption at nodes that are not affected by the failure. However, the technique proposed in Patent Document 1 is based on a network composed of MPLS devices that operate RSVP-TE, and detects a message such as PathErr in signaling at a relay node. For this reason, high function processing such as signaling is required in the relay node.

  The present invention relates to a multicast transfer system and a multicast path that can set a detour path for a multicast path without performing high-function processing such as signaling at a relay node when detouring a fault section due to protection when a fault occurs. An object is to provide a switching method.

  According to a first aspect of the present invention, there is provided a multicast forwarding system in which a multicast packet transmitted from an ingress node is multicast-forwarded to a plurality of egress nodes via a relay node. The Ingress node receives a failure notification from the Egress node when an error is detected, and the Ingress node receives the failure notification from the Egress node. Delete from the previous Egress node and register it as a standby distribution destination Egress node, and simultaneously replace the active and standby labels corresponding to the combination of the active distribution destination Egress node and the combination of the standby distribution destination Egress node. The relay node associates the label with the delivery destination IF. Referring to the multicast forwarding table comprises means for multicast transfer destination IF corresponding to the working system labels and spare system labels multicast packet transmitted from the Ingress node.

  In the multicast forwarding system of the first invention, the Ingress node exclusively manages the active delivery destination Egress node and the standby delivery destination Egress node registered in the delivery destination registration table.

  In the multicast forwarding system of the first invention, the Ingress node transmits a connectivity confirmation packet to each multicast path including all Egress nodes as a delivery destination, and notifies the failure from the Egress node that has detected the non-delivery of the connectivity confirmation packet. And the Egress node that is the source of the failure notification is identified.

  In the multicast forwarding system according to the first aspect of the invention, the Ingress node transmits a connectivity confirmation packet to each of all multicast paths as a delivery destination, and the failure including the multicast path from the Egress node that has detected non-delivery of the connectivity confirmation packet. The notification is received, and the Egress node and multicast path that are the source of the failure notification are identified.

  According to a second aspect of the present invention, in the multicast path switching method of the multicast forwarding system for multicast forwarding multicast packets transmitted from the Ingress node to a plurality of Egress nodes via relay nodes, the Egress node is connected by multicast forwarding from the Ingress node. When a failure is detected, the failure notification is sent to the Ingress node. The Ingress node receives the failure notification from the Egress node and uses the Egress node that sent the failure notification as the current destination registration table. Delete from the primary distribution destination Egress node and register it as a standby distribution destination Egress node, and replace the active and standby labels corresponding to the combination of the active distribution destination Egress node and the combination of the standby distribution destination Egress node respectively. At the same time, the relay node sets the label and destination IF Referring to the multicast forwarding table attached, multicast transfer destination IF corresponding to the working system labels and spare system labels multicast packet transmitted from the Ingress node.

  In the multicast route switching method according to the second invention, the Ingress node exclusively manages the active delivery destination Egress node and the standby delivery destination Egress node registered in the delivery destination registration table.

  In the multicast route switching method of the second invention, the Ingress node transmits a connectivity confirmation packet to each multicast path including all Egress nodes as a delivery destination, and the Egress node that has detected the failure of the connectivity confirmation packet has failed. The notification is received, and the Egress node that has transmitted the failure notification is identified.

  In the multicast route switching method of the second invention, the Ingress node transmits a connectivity confirmation packet to each of all multicast paths as a delivery destination, and includes the multicast path from the Egress node that has detected the non-delivery of the connectivity confirmation packet. The failure notification is received, and the Egress node and multicast path that are the transmission source of the failure notification are identified.

  The present invention transmits a failure notification to the Ingress node when the Egress node detects non-delivery of the connectivity confirmation packet multicast-forwarded from the Ingress node, and identifies the Egress node that has transmitted the failure notification to the Ingress node. By deleting from the active distribution destination and rewriting the working label and the standby label of the multicast packet in addition to the standby distribution destination, the relay node is not affected by the high-level processing such as signaling and is affected by the failure. A detour route of the multicast path can be set for the route.

  Also, because the connectivity confirmation packet is copied and transferred along the multicast path and a failure is detected at the egress node, unlike the segment protection, the number of connectivity confirmation packets that are sent and received even if the number of egress nodes connected to one relay node increases. Can be prevented.

It is a figure which shows the Example of the multicast transmission system of this invention. It is a figure which shows the operation example of the failure detection and notification by CC common system. It is a flowchart which shows the example of a failure switching process procedure of the Ingress node in CC common system. It is a figure which shows the operation example of the failure detection and notification by CC separate system. It is a figure which shows the structural example of the conventional multicast transmission system. It is a figure explaining the example of a signaling process of the conventional multicast transmission system.

FIG. 1 shows an embodiment of a multicast forwarding system of the present invention.
In FIG. 1, the multicast forwarding system of the present embodiment has a tree configuration in which four Egress nodes EN1 to EN4 are connected from one Ingress node IN via two relay nodes RN1 and RN2. Egress nodes EN1 to EN2 are connected to relay node RN1, and Egress nodes EN3 to EN4 are connected to relay node RN2. Furthermore, a standby Ingress node IN ′, relay nodes RN1 ′ and RN2 ′, and egress nodes EN1 ′ to EN4 ′ are provided for detouring the failure section by protection when a failure occurs.

   Ingress nodes IN and IN ′ include a common distribution destination registration table 11, a distribution destination label determination table 12, and a multicast forwarding table 13. The relay nodes RN1 and RN 'have a common multicast forwarding table 14, and the relay nodes RN2 and RN2' have a common multicast forwarding table 15.

  In the distribution destination registration table 11, for each channel, an active label corresponding to the combination of the active distribution destination Egress nodes and a standby label corresponding to the combination of the standby distribution destination Egress nodes are registered. Rewriting is performed. Here, when a failure occurs between the relay node RN1 and the Egress node EN1 in the channel # 1 and the detour path is switched, the active distribution destination Egress nodes are changed from (EN1, EN2, EN3, EN4) to (EN2). , EN3, EN4), and the standby delivery destination Egress node is rewritten to EN1. Details of the rewriting procedure will be described later.

  In the delivery destination label determination table 12, labels corresponding to all combinations of delivery destination Egress nodes are preset for both the active system and the standby system. For example, the label L15 is set for the distribution destination Egress nodes EN1, EN2, EN3, and EN4, and the label L6 is set for the distribution destination Egress nodes EN2, EN3, and EN4.

  In the multicast forwarding tables 13, 14, and 15, a delivery destination IF corresponding to the label is preset for both the active system and the standby system. For example, since the label L15 corresponds to the egress nodes EN1 to EN4 connected to the relay nodes RN1 and RN2, it is set as L15 / RN1 and RN2 in the multicast forwarding table 13 of IN, and L15 in the multicast forwarding table 14 of RN1. / EN1 and EN2 are set, and L15 / EN3 and EN4 are set in the multicast forwarding table 15 of RN2.

  Hereinafter, based on the configuration of FIG. 1, an outline of operations of the Ingress node IN and the relay node RN1 when a failure occurs between the relay node RN1 and the Egress node EN1 to switch to a detour path will be described.

   The Ingress node IN transmits a connectivity confirmation packet (CC packet), and the relay nodes RN1 and RN2 perform copy transfer according to the multicast transfer tables 14 and 15 similarly to the multicast packet without monitoring the CC packet. Egress nodes EN1, EN2, EN3, and EN4 receive CC packets and check connectivity, respectively.

  When a failure occurs, the failure is detected by the CC packet not reaching the Egress node. The Egress node EN1 that has detected the failure notifies the Ingress node IN of the failure via the unicast path. The Ingress node IN identifies the Egress node EN1 that is the source of the failure notification, and replaces the working label and the backup label to be added to the multicast packet, so that the Egress node affected by the failure in the working multicast path Switch only EN1 to the standby system. That is, as shown in the distribution destination registration table 11 of the Ingress node IN in FIG. 1, the combinations of the active distribution destination Egress node and the active label are changed from (EN1, EN2, EN3, EN4 / L15) to (EN2, EN3, EN4). / L6), and the combination of the standby delivery destination Egress node and the standby label is rewritten from (none / none) to (EN1 / L1).

  As described above, in the multicast forwarding system of the present invention, only the route affected by the failure is detected according to the failure detection due to the non-delivery of the CC packet in the egress node without performing high-function processing such as signaling in the relay node. It is possible to detour by switching the transmission end (path protection) in the Ingress node.

  In addition, when the active distribution destination Egress node and the standby distribution destination Egress node are exclusively managed and the active and standby labels are set, there are multiple channels for multicast distribution, each with different distribution destinations. In addition, it is possible to share the delivery destination label determination table for all Egress nodes. Further, the number of labels and the table size can be reduced.

  Here, CC packets are transmitted to each multicast path including all Egress nodes as distribution destinations, and CC packets are transmitted to each multicast path, and CC packets are transmitted to all configured multicast paths. The description will be divided into CC individual methods for individualizing CC packets for all multicast paths.

(CC common method)
FIG. 2 shows an operation example of failure detection / notification by the CC common method. FIG. 3 shows an example of failure switching processing procedure of the Ingress node in the CC common method.

  In FIG. 2, the Ingress node IN transmits a CC packet to a multicast path including all Egress nodes as distribution destinations. Here, the CC packet is transmitted to each multicast path including all Egress nodes EN1 to EN4 as distribution destinations. When the egress node EN1 detects a failure due to the non-delivery of the CC packet, it notifies the failure via the unicast path. At this time, since the CC packet is common to each multicast path including all the egress nodes, it is impossible to determine which multicast path has failed. However, since CC packets are not transmitted to a multicast path having some egress nodes as distribution destinations, the number of CC packets can be reduced in the entire system.

  In FIG. 3, when an Egress node that detects a failure due to a non-delivery of a CC packet transmits a failure notification, the Ingress node IN receives the failure notification (S0), and identifies the Egress node that transmitted the failure notification (S1). ), A channel in which the Egress node of the failure notification transmission source is registered as the active distribution destination Egress node of the distribution destination registration table (11 in FIG. 1) is searched (S2). In the example of FIG. 1, channel # 1 including Egress node EN1 is searched for as the active distribution destination Egress node. Next, the Egress node (EN1) of the failure notification transmission source of the channel is deleted from the active distribution destination Egress node and added as a standby distribution destination Egress node at the same time (S3).

  Next, with respect to the channel in which the active distribution destination Egress node and the standby distribution destination Egress node are changed, the distribution destination label determination table (12 in FIG. 1) is referred to, and labels (corresponding to combinations of the distribution destination Egress nodes respectively) L6, L1) is selected, and the working label and the spare label in the distribution destination registration table (11 in FIG. 1) are rewritten (S4). In the example of FIG. 1, the working system label is rewritten from L15 to L6, and the spare system label is not rewritten from L1. Then, the rewritten working label (L6) is assigned to the multicast packet transferred from the Ingress node to the working system, and the rewritten protection label (L1) is assigned to the multicast packet transferred to the protection system (S5). As a result, the multicast packet is copied / transferred to the distribution destinations corresponding to the working label and the standby label according to the multicast forwarding table (13, 14, 15 in FIG. 1).

  As described above, it is possible to switch only the Egress node affected by the failure in the active system to the standby system by switching the label without performing high-function processing such as signaling in the relay node.

(CC individual method)
FIG. 4 shows an operation example of failure detection / notification by the CC individual method.
In the CC individual method, CC packets are individually transmitted to all multicast paths. Therefore, the Egress node that detects a failure due to the non-delivery of the CC packet must send the failure path information to the Ingress node along with the failure notification. Can do.

  In FIG. 4 (1), the Ingress node IN identifies the working label (L15) of the failed multicast path and the Egress node (EN1) that transmitted the failure notification. Then, the Egress node (EN1) of failure notification is deleted from the active delivery destination Egress node of the channel for which the identified active label (L15) is set, and is added to the standby delivery destination Egress node. That is, the process of step S2 in the procedure of the CC common method shown in FIG. 3 can be omitted, and the processes after step S3 are the same.

  In the example of FIG. 4 (2), the Ingress node IN identifies the working label (L5) of the failed multicast path and the Egress node (EN2) that transmitted the failure notification. The same applies hereinafter.

  Thus, in the CC individual method, since CC packets for the multicast path are individually set, it is possible to determine which multicast path is faulty and switch only the channel using the corresponding multicast path in the Ingress node. As a result, when multiple multicast paths go through the same section, when only a specific multicast path fails due to a setting error in the multicast forwarding table, instantaneous switching to a normal multicast path occurs. Can be prevented.

IN Ingress node RN Relay node EN Egress node 11 Distribution destination registration table 12 Distribution destination label determination table 13, 14, 15 Multicast forwarding table

Claims (8)

In a multicast forwarding system that multicasts multicast packets sent from Ingress nodes to multiple Egress nodes via relay nodes,
The Egress node comprises means for transmitting a failure notification to the Ingress node when detecting a non-delivery of the connectivity confirmation packet multicast-forwarded from the Ingress node,
The Ingress node receives a failure notification from the Egress node, deletes the Egress node of the failure notification from the active distribution destination Egress node of the distribution destination registration table, and registers it as a standby distribution destination Egress node. , Comprising means for simultaneously replacing the active label and the standby label corresponding to the combination of the active distribution destination Egress node and the combination of the standby distribution destination Egress node,
The relay node refers to a multicast forwarding table in which a label and a delivery destination IF are associated with each other, and performs multicast forwarding to a delivery destination IF corresponding to an active label and a standby label of the multicast packet transmitted from the Ingress node A multicast transmission system comprising: In the multicast forwarding system according to claim 1,
The multicast forwarding system, wherein the Ingress node exclusively manages an active delivery destination Egress node and a standby delivery destination Egress node registered in the delivery destination registration table. In the multicast forwarding system according to claim 1,
The Ingress node transmits a connectivity confirmation packet to each multicast path including all Egress nodes as distribution destinations, receives a failure notification from the Egress node that has detected the failure of the connectivity confirmation packet, and transmits the failure notification. A multicast forwarding system characterized by identifying the original Egress node. In the multicast forwarding system according to claim 1,
The Ingress node transmits a connectivity confirmation packet to each of all multicast paths as a delivery destination, receives a failure notification including the multicast path from the Egress node that has detected the failure of the connectivity confirmation packet, and receives the failure notification. A multicast forwarding system characterized by identifying a source egress node and a multicast path. In a multicast route switching method of a multicast forwarding system for multicast forwarding a multicast packet transmitted from an Ingress node to a plurality of Egress nodes via a relay node,
The Egress node sends a failure notification to the Ingress node when detecting a non-delivery of the connectivity confirmation packet multicast-forwarded from the Ingress node,
The Ingress node receives a failure notification from the Egress node, deletes the Egress node of the failure notification from the active distribution destination Egress node of the distribution destination registration table, and registers it as a standby distribution destination Egress node. In addition, the active system label and the standby system label corresponding to the combination of the active system distribution destination Egress node and the combination of the standby system distribution destination Egress node are changed at the same time,
The relay node refers to a multicast forwarding table in which a label and a delivery destination IF are associated with each other, and performs multicast forwarding to a delivery destination IF corresponding to an active label and a standby label of the multicast packet transmitted from the Ingress node. A multicast route switching method characterized by the above. The multicast route switching method according to claim 5, wherein
The multicast path switching method, wherein the Ingress node exclusively manages an active distribution destination Egress node and a standby distribution destination Egress node registered in the distribution destination registration table. The multicast route switching method according to claim 5, wherein
The Ingress node transmits a connectivity confirmation packet to each multicast path including all Egress nodes as distribution destinations, receives a failure notification from the Egress node that has detected the failure of the connectivity confirmation packet, and transmits the failure notification. A multicast path switching method characterized by identifying an original egress node. The multicast route switching method according to claim 5, wherein
The Ingress node transmits a connectivity confirmation packet to each of all multicast paths as a delivery destination, receives a failure notification including the multicast path from the Egress node that has detected the failure of the connectivity confirmation packet, and receives the failure notification. A multicast route switching method characterized by identifying a source egress node and a multicast path.

Images