Patent Application
20090219808
This application is based upon and claims the benefit of the priority of Japanese patent application No. 2008-047737, filed on Feb. 28, 2008, the disclosure of which is incorporated herein in its entirety by reference thereto.
The present invention relates to a network system and management method therefor, and particularly to a layer 2 ring network system and management method therefor.
Multiple Spanning Tree Protocol (abbreviated as MSTP) is a protocol for controlling a layer 2 multi-ring network. MSTP (standardized as IEEE 801.1s) defines an STP (Spanning Tree Protocol) tree as an instance and configures one instance for a plurality of VLANs (Virtual Local Area Networks). One problem with the MSTP is a slow convergence time.
Multi-ring network control employing a ring protocol is utilized to cope with this problem.
As a technique for managing a layer 2 network having a multi-ring configuration, for instance, a technique in which the network is managed by combining rings and arcs, as shown in
However, in the management technique shown in
In a state 3a shown in
Each of the ring domains 32 and 33 which are managed as arcs, monitors the state of the arc by performing health check between the layer 2 switches on both ends of the associated arc, and each ring domain manages itself independently (see the state 3a). For instance, health check is performed between the layer 2 switches 304 and 308 on both ends of the arc of the ring domain 32, and health check is performed between the layer 2 switches 307 and 310 on both ends of the arc of the ring domain 33.
When a failure occurs in the arc, the ring domain open a blocking port (block port) managed by the ring domain. When a failure 31 a occurs in the link between the layer 2 switches 309 and 310 in the ring domain 32 which is managed in the arc (a state 3b in FIG. 6B), the layer 2 switch 308 opens a blocking port 30b, which has been configured as a port to the layer 2 switch 309, secures a path, and recovers the failure.
Further, when another failure 31b occurs in the link between the layer 2 switches 304 and 305 in the ring domain 32 (a state 3c in
As a result, even though the paths of the layer 2 network still exist, the layer 2 switches 305, 306, and 310 cannot recover tie communication paths to other layer 2 switches. Therefore, the path between the layer 2 switches 306 and 307 is blocked by the blocking port 30c, the link between the layer 2 switches 309 and 310 is cut off by the failure 31a, and the link between the layer 2 switches 304 and 305 is cut off by the failure 31b.
In another technique for managing a layer 2 network having a multi-ring structure, the network is managed by monitoring shared link and newly configuring a blocking port when a failure occurs in the shared link as shown in
For instance, in a state 3y shown in
Further, in a state 3z shown in
Patent Document 1 discloses a data relay apparatus that relays data in a network that shares a part of a plurality of rings and avoids occurrence of a loop path by providing a block in each of the rings. The data relay apparatus transmits a failure notification packet only to a predetermined redundant ring, when a failure is detected in a shared portion of a ring, and sets a block that cuts off a main signal by passing through a control packet at a port where the failure is detected. In other words, in a multi-ring network of which a plurality of rings share a part, when a failure occurs in the shared portion, the data relay apparatus in the shared portion can recover the failure without forming a super loop, by selecting a ring as an initial primary ring, blocking a port on the side where the failure has occurred, and transmitting a trap packet that notifies the failure only to the primary ring. The technique disclosed in Patent Document 1 newly sets blocking ports, as described with reference to
[Patent Document 1]
Japanese Patent Kokai Publication No. JP2006-279279A
The entire disclosure of Patent Document 1 is incorporated herein by reference thereto. The following is an analysis of the related art by the present inventor.
The ring protocol managements method which have been described with reference to
The first problem is that, when the failures 31a and 31b occur simultaneously as shown in
In the technique shown in
The second problem is that, in the technique shown in
In Patent Document 1, a blocking port is added after a failure has occurred. Adding a blocking port introduces a change in the network configuration and in the state of topology, and hence complicates the management process. Further, in Patent Document 1, if failures and recoveries are repeated, it will be impossible to predict the eventual state of the blocking ports.
Accordingly, it is an object of the present invention to provide a switch node, network system, and management method therefor capable of recovering a path when a link failure occurs in a multi-ring network.
The present invention, which seeks to solve one or more of the above problems, is configured as follows.
According to a first aspect of the present invention, there is provided a method for managing a ring network which includes a plurality of ring domains. The method comprises:
a switch node managing a shared link shared by a plurality of ring domains (for example, first and second ring domains); and
when the switch node detects a failure in the shared link, the switch node instructing the plurality of ring domains first and second ring domains) sharing the shared link to make at least one ring domain (for example, a first ring domain) expand to another ring domain (for example, a second ring domain). The one ring domain and the another ring domain form an expanded ring.
In the present invention, there are provided the switch nodes that manage the shared link on both ends of the shared link shared by the plurality of ring domain. The switch nodes each monitor the shared link to detect whether or not a failure has occurred in the shared link.
In the present invention, in the plurality of ring domains, the locations of blocking ports are preset and managed by the master nodes of the ring domains, respectively. No new blocking port is created. The locations of the blocking ports are kept unchanged when the ring domain is expanded.
In the present invention, a ring domain expands to another ring domain, based on priority given to the plurality of ring domains sharing the shared link, when a failure is detected in the shared link. A ring domain of relatively higher priority out of the plurality of ring domains sharing the shared link manages the shared link and the ring domain of relatively lower priority expands to a ring domain of relatively higher priority when a failure is detected in the shared link.
In the present invention, when the switch node that manage the shared link detects a failure in the shared link, the switch node transmitting a trap that notifies a failure in the shared link to a ring domain of relatively higher priority.
In the present invention, when a master node of the ring domain receives a trap that notifies a failure in the shared link from the switch node that manages the shared link, the master node creates an expanded ring domain, in accordance with ring domain information included in the trap. The master node of the ring domain transmits a flush packet, which is transmitted on occurrence of a ring domain failure, to the ring domain in which the master node is included, to demand path change within the ring network.
In the present invention, a transit node, on receipt of a trap that notifies a failure in the shared link from the switch node that manages the shared link, creates an expanded domain in accordance with information included in the trap. The transit node performs path change thereafter, when the transit node receives the flush packet.
According to another aspect of the present invention, there is provided a ring network system which comprises:
a plurality of ring domains; and
a switch node that manage a shared link shared by a plurality of ring domains (for example, first and second ring domains). When the switch node detects a failure in the shared link, the switch node instructs the plurality of ring domains that share the shared link (for example, first and second ring domains) to make at least one ring domain (for example, a first ring domain) expand to another ring domain (for example, a second ring domain). The one ring domain and another ring domain form an expanded ring.
In the system according to the present invention, there are provided the switch nodes that manage the shared link on both ends of the shared link shared by the plurality of ring domains. The switch nodes each monitor the shared link to detect whether or not a failure has occurred in the shared link.
In the system according to the present invention, each of the ring domains includes a master node. In the plurality of ring domains, locations of blocking ports are preset and managed by the master nodes of the ring domains, respectively. No new blocking port is created and the locations of blocking ports are kept unchanged when the ring domain is expanded. In the present invention, a ring domain is made expanded to another ring domain based on priority given to the plurality of ring domains that constitute the shared link. In the present invention, a ring domain of relatively higher priority out of the plurality of ring domains that share the shared link manages the shared link and a ring domain of relatively lower priority (for example, a first ring domain) may be made expanded to a ring domain of relatively higher priority (for example, a second ring domain).
In the present invention, when the switch node that manages the shared link detects a failure in the shared link, the switch node transmits a trap that notifies a failure in the shared link to a ring domain of relatively higher priority.
In the present invention, when a master node of the ring domain receives a trap that notifies a failure in the shared link from the switch node that manages the shared link, the master node creates an expanded domain in accordance with ring domain information included in the trap, and transmits a flush packet, which is transmitted on occurrence of a ring domain failure, to the ring domain in which the master node is included to demand path change within the ring network.
In the present invention, a transit node, on receipt of the trap that notifies a failure in the shared link from the switch node that manages the shared link creates an expanded domain according to information included in the trap, and performs path change thereafter if the transit node receives a flush packet which is transmitted on occurrence of a ring domain failure.
According to the present invention, there is provided a switch node, located on an end of a link shared by ring domains, that monitors the shared link and controls so that at least one ring domain out of a plurality of ring domains that constitute the shared link expands to another ring domain when the switch node detects a failure in the shared link. In the present invention, the switch node transmits a trap that notifies a failure in the shared link to a ring domain of relatively higher priority.
According to the present invention, even when multi-link failures occur in a multi-ring network of a layer 2 network, layer 2 paths can be recovered by switching paths as long as the layer 2 paths still exist.
Still other features and advantages of the present invention will become readily apparent to those skilled in this art from the following detailed description in conjunction with the accompanying drawings wherein only exemplary embodiments of the invention are shown and described, simply by way of illustration of the best mode contemplated of carrying out this invention. As will be realized, the invention is capable of other and different embodiments, and its several details are capable of modifications in various obvious respects, all without departing from the invention. Accordingly, the drawing and description are to be regarded as illustrative in nature, and not as restrictive.
In the present invention, ring domains each are managed in a ring, not in an arc. For instance, in a state 1a in
When the SLH detects a failure in the shared link, the SLH transmits a message to expand the domain to any one of ring domains.
The layer 2 switch, on receipt of a flush packet which is transmitted at a time of recovery from the link failure, deletes the expanded domain.
The layer 2 switch, on receipt of the message to expand a domain from the SLH, expands the domain.
According to the present invention, in a multi-ring network including a layer 2 network, even when multiple link failures occur, communication paths can be maintained by securing layer 2 paths and switching paths. While the configuration of the links depends on the number of closed paths, the links can be configured freely. The priority of the links can be set freely as well. A master node can freely set one of opposing switches as a blocking port, however, no blocking port is newly created after a failure has occurred and there is no need for changing blocking ports.
In
a ring (ring domain) 1 constituted by layer 2 switches 101, 102, 103, 104, 107, and 108;
a ring (ring domain) 2 constituted by layer 2 switches 104, 105, 108, 109, and 110; and
a ring (ring domain) 3 constituted by layer 2 switches 106, 107, 108, 109, and 110.
Although
The layer 2 switch group constituting the ring 1 belongs to a ring domain 11.
The layer 2 switch group constituting the ring 2 belongs to a ring domain 12.
The layer 2 switch group constituting the ring 3 belongs to a ring domain 13.
In each ring domain, there is provided a master node (M) that manages the respective ring domain. The master nodes in the ring domains 11, 12, and 13 are the layer 2 switches 102, 105 and 106 respectively.
The layer 2 switches 104 and 108 belong to both the ring domains 11 and 12 and are configured as the SLHs (Shared Link Holders) that manage a link shared by the ring domains 11 and 12.
The layer 2 switches 107 and 108 belong to both the ring domains 11 and 13 and are configured as the SLHs that manage a link shared by the ring domains 11 and 13.
The layer 2 switches 108, 109, and 110 belong to both the ring domains 12 and 13.
The layer 2 switches 108 and 110 are configured as the SLHs that manage a link shared by the ring domains 12 and 13. The configuration information of the layer 2 multi-ring network described above may be transferred to each layer 2 switch from the storing means of a predetermined node (layer 2 switch).
The SLHs managing the same shared link transmit a heartbeat message (transmitted as a control packet) to each other and monitors whether or not any failure has occurred in the shared link. The layer 2 switches located on both ends of a shared link (for instance the layer 2 switches 108 and 110) are configured as the SLHs. The normality of the shared link is monitored by having the SLHs transmit the heartbeat message to each other.
When the SLHs detect a failure in the shared link due to the timeout of the heartbeat message or the link down of the shared link, the SLHs perform control to expand one of the ring domains to another ring domain (
Further, if another failure 11b occurs in the state 1d (refer to
In the present example, the control over the ring network can be maintained against any pattern of link failures since the ring configuration is redefined by expanding the ring domains when a link failure occurs in the shared link.
Further, in the present exemplary embodiment, since the position of the blocking port is managed by the initially set master node (M) of each ring domain, the position of the blocking port remains unchanged, no matter how the ring domains are expanded, and the path information is easily graspable. In other words, in the present exemplary embodiment, since no blocking port is newly created, it is possible to grasp the position of the blocking port even during the occurrence of a failure. As a result, backup path design can be simplified and the cost of managing the network topology can be reduced.
In each ring domain, one master node (M) that manages the respective ring domain is provided.
The master node (M) in each ring domain monitors the state of the ring using a health check packet, and when a ring, is formed, the master node avoids formation of a network loop by blocking one of the ports of the ring.
In
The layer 2 switch 102, the master node of the ring domain 11, sets up a block 10a at a port on the side of the layer 2 switch 101. The layer 2 switch 102 periodically transmits a health check packet in the ring domain 1.
The layer 2 switch 105, the master node of the ring domain 12, sets up a blocking port 10b on the side of the layer 2 switch 104. The layer 2 switch 105 periodically transmits a health check packet in the ring domain 12.
The layer 2 switch 106, the master node of the ring domain 13, sets up a blocking port 10c on the side of the layer 2 switch 107. The layer 2 switch 106 periodically transmits a health check packet in the ring domain 13.
The other layer 2 switches 101, 103, 104, 107, 108, 109, and 110 are transit nodes.
In each ring domain, the master node and transit nodes (nodes that are not the master node) respectively manage the following states shown in
The layer 2 switches 104 and 108 belong to both the ring domains 11 and 12 and are configured as the SLHs that manage the link shared by the ring domains 11 and 12.
The layer 2 switches 107 and 108 belong to both the ring domains 11 and 13 and are configured as the SLHs that manage the link shared by the ring domains 11 and 13.
The layer 2 switches 108, 109, and 110 belong to both the ring domains 12 and 13.
The layer 2 switches 108 and 110 are configured as the SLHs that manage the link shared by the ring domains 12 and 13.
The SLHs transmit the heartbeat message to each other over the shared link and monitors whether or not any failure has occurred in the shared link.
The layer 2 switches 104, 107, 108, 109, and 110 each on the shared links determine which ring domain manages the shared link according to the priority of the ring domain.
In the present exemplary embodiment, between die ring domains 11 and 12, it is assumed that the ring domain 12 has a higher priority. Therefore, it is determined that the shared link between the layer 2 switches 104 and 108 is managed by the ring domain 12.
When the failure 11a occurs in the shared link between the layer 2 switches 104 and 108 (refer to the “x”-marked section in
When the layer 2 switch 105 receives the trap (T) of link failure from the layer 2 switch 104, the layer 2 switch 105 cancels the block state 10b set up at a port on the side of the layer 2 switch 104 (refer to
In order to demand path change, the layer 2 switch 105 flushes its MAC (Media Access Control) table. The layer 2 switch 105 transmits flush packets (F), issued upon the occurrence of a failure, to the ring domain 12 in order to flush its MAC table (
The layer 2 switches 104, 105, 108, 109, and 110 constituting the ring domain 12 which has the link between the layer 2 switches 104 and 108 shared with ring domain 11, on receipt of the trap (T) notifying the failure 11a in the shared link between the layer 2 switches 104 and 108, create newly an expanded ring domain 11′ (refer to
After the failure 11a in the shared link between the layer 2 switches 104 and 108 has occurred, the health check packet for the expanded ring domain 11′ transmitted by the layer 2 switch 102 travels through the layer 2 switches 103, 104, 105, 110, 109, 108, 107, and 101, and returns to the layer 2 switch 102 (refer to
After the failure 11a in the shared link between the layer 2 switches 104 and 108 has occurred and the ring domain 11 has been expanded to the ring domain 11′ to cover (include) ring domain 12 (refer to
After the failure 11a in the shared link between the layer 2 switches 104 and 108 has occurred and the ring domain 11 has been expanded to the ring domain 12 (refer to
As described, the network is redefined as a network constituted by two rings: the ring domains 11′ and 13.
From this state, as shown in
When the layer 2 switch 106 in the ring domain 13 receives the trap (T) notifying a link failure from the layer 2 switch 107, the layer 2 switch 106 cancels the block state 10c set up at the port on the side of the layer 2 switch 107 (refer to
In order to demand path change, the layer 2 switch 106 flushes its MAC (Media Access Control) table, and transmits the flush packets (F), issued upon the occurrence of a failure, to the ring domain 13 in order to flush the MAC tables (
The states of the layer 2 switches 106, 107, 108, 109, and 110 are changed to states in which they belong to a further expanded ring domain 11″ (refer to
The operation of the SLH in the present exemplary embodiment, described with reference to
As shown in
As a result of the examination by the SLHs, when an abnormality, such as a link down or the timeout of a heart beat message, occurs in the shared link (705), the SLHs detects a failure in the shared link (706) and the SLHs expand the ring domain of lower priority (out of the ring domains sharing the shared link) to cover the ring domain of higher priority (707). Further, the SLHs transmit a trap notifying the failure in the shared link to the other layer 2 switches in order to expand the ring domain (708). At this time, the SLHs transmit the trap (T) notifying the failure in the shared link to the ring domain of higher priority.
The above procedure performed by each SLH may well be implemented by a computer program executed on a computer (CPU) in the SLH.
Next, the operation of the master node that has received the trap notifying the failure in the shared link in the present exemplary embodiment will be described with reference to a flowchart shown in
Before the failure has occurred, the master node is in the “Complete” state (601). After the failure has occurred, the health check packet transmitted (602) times out (603), and the state of the master node changes to the “Fail” state (604).
Then the master node opens a blocking port set up by the master node itself (605) and performs path change (the initialization of the MAC tables) (606). When the failure has occurred in the shared link, the master node receives a trap (T) notifying the failure in the shared link.
In the case where the master node receives the trap (“YES” in 607), the master node expands the ring domain according to ring domain information included in the trap (609).
Then, in order to bring the paths within the ring domains into a normal state, the master node transmits the flush packet (608), issued upon the occurrence of a failure in the ring domain, and changes the paths within (or via) the rings. When the master node does not receive the trap (“NO” in 607), the master node transmits the flush packet (608).
The above procedure performed by the master node may be implemented by a computer program executed on a computer (CPU) of the master node.
Next, the operation of the transit nodes, which have received the trap notifying the failure in the shared link, in the present exemplary embodiment will be described with reference to a flowchart shown in
Since the transit nodes are not directly involved in the ring path control, the state of the transit nodes does not matter. The transit node in the “LinkUp” state (801) changes to the “LinkDown” state (806) when it detects a link down on itself.
Whether in the “LinkUp” or “LinkDown” state, the transit node checks if the trap notifying the failure in the shared link (803) has been received. When the trap is received by the transit node, the transit node expands the domain according to the information included in the trap (807).
Thereafter, the transit node receives the flush packet (issued upon the occurrence of a failure) transmitted from the master node (804), flushes the MAC table (805), and performs path change.
The above procedure performed by the transit node may be implemented by a computer program executed on a computer (CPU) of the transit node.
According to the present exemplary embodiment, the expansion of the ring domains can be controlled by having the SLHs, the master nodes, and the transit nodes operate as described.
As described above, the present exemplary embodiment has the following benefits.
Upon the occurrence of any pattern of link failures in the layer 2 multi-ring network, the path can be recovered by expanding an existing ring domain and adjusting to a newly created ring network, as long as layer 2 paths exist.
When the path control is performed, any one or more of the existing blocking ports are put into either a blocking or transfer state. Therefore the position of the blocking ports remains unchanged, and the path state of the layer 2 ring network upon the occurrence of a failure is easily graspable.
The present invention can be applied to layer 2 switches constituting a multi-ring network or to a network apparatus such as a router.
The disclosure of the aforementioned Patent Document 1 is incorporated into the present document by reference.
It should be noted that other objects, features and aspects of the present invention will become apparent in the entire disclosure and that modifications may be done without departing the gist and scope of the present invention as disclosed herein and claimed as appended herewith.
Also it should be noted that any combination of the disclosed and/or claimed elements, matters and/or items may fall under the modifications aforementioned.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2008-047737 | Feb 2008 | JP | national |
