Patent Application
20070286069
This is a continuation of International Application PCT/CN2006/000170, with an international filing date of Jan. 26, 2006. The International Application claims priority of Chinese application 200510004845.9, with a filing date of Feb. 2, 2005.
1. Field of the Invention
The present invention relates to the technical field of DCN (Data Communication Network), in particular, to a method for realizing working/standby transmission path.
2. Background of the Invention
Among many networks transmitting data based on IP, a DCN is a packet communication network specialized for carrying management information and distributed signaling message, and belongs to the scope of telecommunication supporting network and does not transmit the final traffic flow of users directly.
Specifically, a physical transmission link in a DCN refers to an in-fiber control channel and an out-of-fiber control channel. The in-fiber control channel can be an ECC (Embedded Control Channel), such as a DCC (Data Communications Channel) in an SDH (Synchronous Digital Hierarchy), or a GCC (General Communication Channel), an dedicated optical monitoring channel in an OTN (which only transmits management information and distributed signaling message, and does not transmit service message) can be an Ethernet transmission path.
A network layer of the DCN conforms to TCP/IP protocol. In other words, each node in a DCN can query a preset routing table according to destination IP address information encapsulated in a transmitted packet to obtain next hop node information and then forward the packet. The routing table configured in each node can be based on static route configuration protocol, or dynamic routing protocol. At present, dynamic routing protocol is usually adopted to generate a routing table because dynamic routing protocol, because it has the advantage that network failure can be prevented from occurring on a packet by dynamically modifying the routing table configuration. Thus, a DCN can dynamically refresh the routing table configured in each node according to the network failure status. After the whole-network routing table is refreshed synchronously, a packet to be transmitted can be delivered to the destination node, keeping away from the paths on which failure occurs. Therefore, in a DCN, by realizing working/standby transmission path by means of routing tables in conjunction with dynamic routing protocol, strong fault-tolerance capacity can be obtained.
However, in the method for realizing working/standby transmission path in a DCN by means of routing tables in conjunction with dynamic routing protocol, after a failure occurs on the network, the nodes on a transmission path on which the failure occurs need to diffuse the failure information to other nodes hop by hop from near to far according to dynamic routing protocol, and then each node re-computes the routing table. During the update of the routing table (i.e., before the whole-network routing table is refreshed synchronously), it is possible that the routing table in each node is not consistent with each other. As a result, part of the packets can not be forwarded to the destination node correctly and be discarded due to a limitation on forwarding times or related queue overflow. Moreover, the time needed for the update of the routing table relates to the scale of the network. For a small size network, the time can be controlled at second level, but for a large-size network, the time can be several minutes, which sometimes is intolerable for the transmission of the management information and distributed signaling message and can cause the loss of some very important management information. For example, failure notice message and so on may be discarded during the transmission process.
In addition, according to a principle of MPLS (Multi-Protocol Label Switching) technology, a plurality of logical LSPs (Label Switched Paths) can be established on the network layer of a DCN. Thus, when a packet to be transmitted passes through an LSP, it can be forwarded directly according to the label value (Label) encapsulated in the packet, so it can be avoided to look up the routing table. For a packet, LSP is a “Tunnel”, i.e., a virtual packet forwarding link. Thus, by establishing two or more non-overlapped LSPs between any two nodes, a mutual protection relation can be formed among these LSPs. In other words, when one LSP is broken down due to network failure, the packet can be switched to other standby LSPs for forwarding. Thus, MPLS 1+1 protection or 1:1 (N) protection solution (for MPLS 1+1 protection or 1:1 (N) protection solution, reference can be made to ITU-T G.7712) can be formed. The advantage of the method for realizing working/standby transmission path in a DCN via LSP in the MPLS lies in that: a packet can be rapidly switched from an LSP on which failure occurs to other standby LSPs for forwarding. The duration of the switching process can be shorter than 50 ms, so that loss of packets during the transmission process can be reduced.
However, in the method for realizing working/standby transmission path in a DCN via LSP in the MPLS, two or more LSPs need to be pre-established between any two nodes on which a communication message is to be delivered. Thus, for a DCN with N nodes, if it is required to establish an LSP connection between any two nodes, the total number of LSPs to be established will be in direct proportion to the square of node number N. Therefore, establishment and maintenance workload of establishing and maintaining N square LSPs is very heavy, and the extensibility of DCN will be limited. In addition, when two or more failures occur on the network resources, failures may be occur on the working LSP and standby LSP simultaneously. Therefore, communication can be broken down. In such case, the reliability of packet forwarding will be lower than that of the method for realizing working/standby transmission path by means of routing tables in conjunction with dynamic routing protocol.
An object of the present invention is to provide a method for realizing working/standby transmission path that can improve the reliability of packet forwarding and reduce maintenance workload and maintenance complexity of the working/standby transmission path.
To solve the above problem, the invention employs the following solutions.
A method for realizing working/standby transmission path, including:
configuring a standby LSP for a physical transmission link in a network; and
forwarding a packet to be transmitted on the physical transmission link to the standby LSP for transmission when a failure occurs on the physical transmission link.
The network is a network for transmitting data via IP.
In the method for realizing working/standby transmission path, N pairs of LSPs are configured for each physical transmission link, and transmission directions of each pair of LSPs are opposite to each other, N being a natural number.
In the method for realizing working/standby transmission path, configuring a standby LSP for the physical transmission link is accomplished by one of:
using RSVP-TE (Resource Reservation Protocol-Traffic Extension);
using CR-LDP (Constraint-Based Label Distribution Protocol); and
configuring a label forwarding table in each node in the network.
The physical transmission link is physically independent of a physical transmission link through which the standby LSP passes.
The method for realizing working/standby transmission path further includes: when the physical transmission link on which failure occurs recovers to normal, forwarding the packet transmitted on the standby LSP to the physical transmission link that has recovered to normal for transmission.
A cost value of the standby LSP is configured as greater than a cost value of the physical transmission link protected by the standby LSP and less than a sum of cost values of physical transmission links through which the LSP passes.
In the method for realizing working/standby transmission path, if at least two pairs of LSPs are configured for the physical transmission link, when the failure occurs on the physical transmission link, the packet to be transmitted on the physical transmission link is forwarded to a standby LSP with a minimum cost value for transmission.
In the method for realizing working/standby transmission path, forwarding the packet to be transmitted on the physical transmission link on which the failure occurs to a standby LSP for transmission is accomplished by one of:
modifying a list item directed to the physical transmission link on which the failure occurs in a routing table saved in a source node on the physical transmission link on which the failure occurs to a list item directed to the standby label switched path; and
redirecting the packet transmitted on the physical transmission link on which the failure occurs to a standby LSP for transmission in an interface driver of the physical transmission link on which the failure occurs.
The physical transmission link is an in-fiber control channel or an out-of-fiber control channel.
The method for realizing working/standby transmission path further comprises detecting whether a failure occurs on the physical transmission link and the standby LSP by periodicity sending a failure detection message to the network.
The method for realizing working/standby transmission path further comprises detecting whether a failure occurs on a physical transmission link composed of the in-fiber control channel via optical network alarm detecting mechanism.
In the method for realizing working/standby transmission path, when the failure occurs on the physical transmission link while the standby LSP works normally, two end nodes of the physical transmission link on which the failure occurs delay delivering the failure information to other nodes via dynamic routing protocol; when the failure occurs on the physical transmission link and the standby LSP simultaneously, the two end nodes of the physical transmission link on which the failure occurs deliver the failure information to other nodes in time via dynamic routing protocol.
In the method for realizing working/standby transmission path according to the invention, by configuring a standby LSP for each physical transmission link in a DCN, a corresponding physical transmission link can be protected using the standby LSP, so that when a failure occurs on the physical transmission link, a packet transmitted on the physical transmission link can be forwarded to the standby LSP for transmission. Thus, when a failure occurs on a physical transmission link on which a packet is to be transmitted, the packet can be forwarded to a standby LSP for transmission in time. As a result, the reliability of packet forward can be improved, and the loss of packets during the transmission process can be reduced greatly and thus transmitting management information and distributed signaling message in a DCN is facilitated.
Meanwhile, because the number of the LSPs configured for the physical transmission links is in direct proportion to the number of the physical transmission links to be protected. In other words, N (N is a natural number, which is usually taken as 1) pairs of LSPs are configured for a physical transmission link, wherein the transmission directions of each pair of LSPs are opposite to each other. Thus, a disadvantage can be avoided that a total number of LSPs to be established should be in direct proportion to the square of node number N when the prior art method for realizing DCN working/standby transmission path via LSP in the MPLS is employed. As a result, the number of LSPs to be configured can be reduced greatly. Thus, the maintenance workload and maintenance complexity of the working/standby transmission path in a DCN can be reduced.
The embodiments of the invention will now be illustrated in detail in conjunction with the drawings.
The invention can be applied to a network for transmitting data via IP, such as DCN or Internet, etc. It will now be illustrated by taking the case in which the invention is applied to a DCN as an example. In the method for realizing working/standby transmission path based on a DCN, a corresponding standby LSP is configured for each physical transmission link in the DCN, so that when failure occurs on a physical transmission link, a packet transmitted on the physical transmission link can be forwarded to a standby LSP for transmission in time, thus loss of packets can be reduced; at the same time, the maintenance workload and maintenance complexity of the working/standby transmission path can be reduced.
Step S10: A standby LSP is configured for each physical transmission link in the DCN.
In the DCN, the physical transmission link can be composed of an in-fiber control channel, such as ECC; or, the physical transmission link can be composed of an out-of-fiber control channel (such as physical transmission media using out-of-fiber Ethernet as a DCN). The method of the invention can be applied to the physical transmission links composed of either of the two transmission media.
Because LSPs of MPLS technology usually support the unidirectional transmission of packets, and the LSP configured for a physical transmission link here should have the property of bidirectional transmission so as to protect the physical transmission link in bidirectional transmission modes respectively, two LSPs in opposite directions need to be configured for each physical transmission link. In other words, a pair of LSPs with opposite transmission directions needs to be configured.
Certainly, N (N is a natural number, which is usually taken as 1) pairs of LSPs can also be configured for each physical transmission link in the DCN wherein each pair of LSPs has opposite transmission directions. In other words, a plurality of different LSPs can be configured for a physical transmission link. Thus, when failure occurs on the physical transmission link and one or more LSPs for the physical transmission link simultaneously, the packet can also be forwarded to a normal LSP for the physical transmission link for transmission. Therefore, it can be further ensured that packets will not be lost during the transmission process, so that it is favourable to improve the reliability of the transmission of management information and distributed signaling message in a DCN.
The protected physical transmission link and the physical transmission link through which the standby LSP passes are physically independent; i.e., failure-uncorrelated. In other words, the physical transmission link through which the configured LSP passes and the protected physical transmission link are mutually independent in a physical layer, and occupy different physical transmission links respectively. Thus, the failure occurring on the standby LSP can be uncorrelated to the failure occurring on the protected physical transmission link. As a result, a failure occurs on a physical transmission link will not cause a failure to occur on the standby LSP, so that the object of protecting the physical transmission link via the standby LSP can be attained.
Because N pairs of LSPs with opposite transmission directions are configured for a physical transmission link to protect the physical transmission link, the number of the LSPs configured is in direct proportion to the number of the physical transmission links. Thus, the disadvantage in the method of the prior art for realizing DCN working/standby transmission path employing MPLS technology can be avoided that a number of LSPs to be configured is a square number of the node number N in the DCN, and a problem of increasing the maintenance workload and maintenance complexity of the working/standby transmission path and so on can be avoided.
In a topology structural diagram of a DCN, a configured LSP and a physical transmission link it protects are in a relation of parallel links. In other words, the physical transmission link and the standby LSP originate from the same source node and end at the same destination node.
Meanwhile, it should be ensured that a packet is forwarded to the physical transmission link, rather than the configured standby LSP, for transmission when the physical transmission link works normally in the DCN. Because in normal condition, the standby LSP is a long path compared with the physical transmission link to be protected and its routing is not optimized.
The process of configuring a standby LSP for each physical transmission link in the DCN can be accomplished according to RSVP-TE (Resource Reservation Protocol-Traffic Extension); or, it can be accomplished according to CR-LDP (Constraint-Based Label Distribution Protocol). Alternatively, a static LSP can also be realized by configuring a label forwarding table in each node of the DCN respectively. For the above signaling protocol for realizing LSP, reference can be made to the related standard text specified in IETF (Internet Engineering Task Force). Specifically, for RSVP-TE, reference is made to RFC3209; and for CR-LDP, reference is made to RFC3212.
Meanwhile, to ensure that the working/standby transmission path in the DCN realized in the method of the invention can transmit packets reasonably, the cost value of the standby LSP can be configured as greater than the cost value of the physical transmission link to be protected by the standby LSP and less than the sum of the cost value of each physical transmission link through which the standby LSP passes, so that when a failure occurs on the physical transmission link, the packet can be forwarded to the standby LSP for transmission; and when the physical transmission link works normally or recovers to normal, the packet can be automatically forwarded to the physical transmission link for transmission (because the cost value of the physical transmission link is less than that of the standby LSP, and the packet flow transmitted via the LSP originally can be automatically forwarded to the physical transmission link with a smaller cost value for transmission after the routing table is updated).
For the case in which at least two pairs of LSPs are configured for each physical transmission link, when a failure occurs on a physical transmission link, packets to be transmitted on the physical transmission link is first forwarded to a standby LSP with the minimum cost value for transmission.
In the DCN as shown in
Each node in the DCN delivers the configuration of its physical transmission link and the corresponding standby LSP to each of the other nodes in the network via a Link State Advertisement message according to dynamic routing protocols, such as OSPF (Open Shortest Path First). Thus, each node in the DCN can obtain a path topology state diagram. For the specific formation of the path topology state diagram, reference is made to
As shown in
In OSPF protocol, a Hello message is regularly sent via the management plane layer in the DCN, by this way, working state of the working/standby transmission path in a DCN can be detected. In other words, the specific working state of each physical transmission link and the corresponding standby LSP can be detected. Meanwhile, to reduce a transmission bandwidth occupied during the sending process of the Hello message, the sending period of the Hello message as set cannot be too short. The sending period is about 10 seconds usually. However, to shorten a route convergence process of each node in the DCN, the sending period of the Hello message should be shortened appropriately; for example, it can be set as 1 second or shorter. In other words, the sending period of the Hello message as a failure discovery mechanism needs to be compromised between the rapidity of failure detection and the dedicated overhead of resources.
When the physical transmission link in the DCN is composed of an ECC, an optical network alarm detecting mechanism based on port level can also taken as a supplementary of the Hello message detecting mechanism. For example, in an SDH network, working state of a physical transmission link in the DCN can be determined by detecting an alarm message of a regeneration section or multiplex section. In other words, if alarm of LOS (Lost Of Signal) or LOF (Lost Of Frame) appears, it indicates that break-down failure occurs on an associated physical transmission link. Thus, a failure positioning efficiency of the physical transmission link can be improved greatly via the optical network alarm detecting mechanism. When a failure occurs, a packet transmitted on a related physical transmission link on which the failure occurs can be switched to the standby LSP for transmission as quickly as possible. Because the physical transmission link in the DCN is usually bidirectional, if only unidirectional transmission failure occurs on the physical transmission link, the downstream nodes of the physical transmission link on which the failure occurs will detect the failure state immediately. An SDH, a Sonet (Synchronous Optical Network), an OTN (Optical Transport Network) and so on can notify the opposite node of the failure information in REI/RDI (Remote Error Indication/Remote Defect Indication) mode. Thus, no matter it is unidirectional transmission failure or bidirectional transmission failure that occurs on the physical transmission link, the nodes on both sides of the physical transmission link on which the failure occurs can detect the failure state quickly.
When a failure occurs on a physical transmission link in the DCN and the standby LSP works normally, the two end nodes of the physical transmission link on which the failure occurs need to delay delivering the failure information to other nodes in the network using dynamic routing protocol; but when the failure occurs on the physical transmission link and the standby LSP simultaneously, the two end nodes of the physical transmission link on which failure occurs need to deliver the failure information to other nodes in the network in time via dynamic routing protocol. In other words, the delayed delivery of failure information only applies to the situation in which the failure occurs on the physical transmission link while the standby LSP works normally; in other situations, the failure state of the transmission path needs to be delivered immediately. By using the optical network alarm detecting mechanism based on port, the time for failure recovery of a physical transmission link in the DCN can be decreased to within 50 ms. Therefore, the reliability of packet forwarding in the DCN can be further improved.
The reasons that the node in a DCN delays delivering the failure state information to other nodes in the network via dynamic routing protocol (such as OSPF protocol and so on) when the failure occurs on a physical transmission link lie in that:
(1) if the two end nodes of the physical transmission link on which the failure occurs deliver the failure state information to the whole network immediately when the failure is detected, the IP routing table of each node in the network will be changed, and during this process, it is possible that the IP routing table in each node is inconsistent with each other, thus the communication process between each node in the DCN will possibly be influenced;
(2) after the failure occurs on a physical transmission link, because the packet sent by the two end nodes of the physical transmission link on which the failure occurs may not reach the destination node normally, the sending node usually resends the packet for several times. Thus, the traffic between any two nodes in the DCN and between each node and the network management system will be increased, and at this point, “message storm” may take place in the network. If the state update information of dynamic routing protocol, based on OSPF is delivered and the routing table is re-computed by each node according to the route state update information at the same time, the communication traffic and computation load will be further increased. Therefore, the two end nodes of the physical transmission link on which the failure occurs delay delivering the failure state information (the actual delay time is set according to specific network situation), so the related node can deliver the route state update information in time on an appropriate occasion, and the IP routing table of each node can be updated and optimized in time, because the “message storm” has ended at this moment. Therefore, the influences of the update process of IP routing tables on the network load can be reduced greatly.
As shown in
At the same time, in the DCN, the two end nodes of the physical transmission link on which the failure occurs can diffuse the failure information to other nodes in the network via OSPF protocol or IS-IS (Intermediate System to Intermediate System) protocol. For the specific standard of IS-IS protocol, reference can be made to ISO Standard No. 8473 (Intermediate System to Intermediate System Intra-Domain Routeing Exchange Protocol for use in Conjunction with the Protocol for Providing the Connectionless-mode Network Service).
As shown in
In addition, when the packet is transmitted on the standby LSP, if the failure state of the corresponding physical transmission link is eliminated during the process, i.e., the corresponding physical transmission link recovers to normal working state, the packet transmitted on the standby LSP needs to be forwarded to the physical transmission link that recovers to normal for transmission. Because after the physical transmission link is restored, the state of the related physical transmission link changes from Down (which means that the working state of the physical transmission link is abnormal) to Up (which means that the working state of the physical transmission link is normal). At this point, the situation in which the physical transmission link recovers to normal can also be detected based on the Hello message of OSPF protocol, thus the dynamic route configuration can be updated again via OSPF protocol, so as to update the whole-network IP routing table. Moreover, because the cost value of the physical transmission link that recovers to normal is less than that of the standby LSP, the packet flow transmitted on the standby LSP previously will be automatically forwarded to the physical transmission link that has recovered to normal working state for transmission.
As shown in
During the above process, forwarding the packet to be transmitted on the physical transmission link on which the failure occurs to the standby LSP for transmission can be accomplished by modifying a list item directed to the physical transmission link on which the failure occurs in the IP routing table saved in a source node on the physical transmission link on which the failure occurs as directed to a list item of the corresponding standby LSP. In other words, when a node in the DCN detects that a failure occurs on a physical transmission link related to the node and a corresponding standby LSP works normally, the node can modify the routing table items in which the forwarding interface is directed to the failed physical transmission link in the IP routing table saved therein as directed to the forwarding interface of the corresponding LSP. Thus, the packet can be forwarded from the physical transmission link on which the failure occurs to a normal LSP for transmission.
In addition, forwarding the packet to be transmitted on the physical transmission link on which the failure occurs to the standby LSP for transmission can also be accomplished by redirecting the packet transmitted on the physical transmission link on which the failure occurs to the standby LSP for transmission in an interface driver of the physical transmission link on which the failure occurs. In other words, the packet actually transmitted is redirected in the interface driver of the physical transmission link, which means that the packet previously forwarded on the physical transmission link (herein, the packet refers to packet other than OSPF protocol packet, because OSPF protocol packets need to reflect the actual working state of the network) is forwarded to a preconfigured standby LSP for transmission. In the redirecting process, the IP routing table in the DCN does not need to be modified.
The above two modes will now be illustrated by an example:
Now, an embodiment of the method for realizing working/standby transmission path based on DCN according to the invention will be illustrated in conjunction with the structural diagram shown in
First of all, the packet is forwarded to node C via physical transmission link C-F; node C forwards the packet via nodes E and D to node B through the standby LSP of physical transmission link B-C, and this forwarding process is realized via label forwarding in the standard MPLS technology, and for the packet forwarding on IP network layer, in this forwarding process, a packet is directly forwarded via node C to node B (because MPLS forwarding is a forwarding mode between data link layer and network layer, which is invisible to the network layer); then, node B forwards the packet to node A. It can be seen that: if the technical solution of the invention is not employed, in other words, if the standby LSP of physical transmission link B-C does not exist, the forwarding process may fail, because the update of the routing table is not completed when the failure state information sent by node F reaches node C.
Additional advantages and modifications will readily occur to those skilled in the art. Therefore, the invention in its broader aspects is not limited to the specific details and representative embodiments shown and described herein. Accordingly, various modifications and variations can be made without departing from the spirit or scope of the invention as defined by the appended claims and their equivalents.
| Number | Date | Country | Kind |
|---|---|---|---|
| 200510004845.9 | Feb 2006 | CN | national |
| PCT/CN06/00170 | Jan 2006 | CN | national |
