KR101046009B1 - How to detect network failure - Google Patents

Abstract

The present invention, when the GRE tunnel is generated between the other equipment and the other equipment connected via the data transmission path, the step of transmitting the GRE PIC state information of the other equipment from the other equipment to the one equipment; Grasping GRE PIC state information of the other device by the one device; And generating a GRE tunnel in the data transmission path when the GRE PIC state is normal, and blocking the data transmission path and activating a spare path when the GRE PIC state is abnormal.

Description

Method for Detecting a Network Failure}

BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a diagram of the case where autonegotiation function is used between homogeneous routers.

2 is a diagram illustrating a problem when an auto-negotiation function is used between heterogeneous routers.

3 is a diagram illustrating a problem when no auto-negotiation function is used between different routers.

4 is a diagram illustrating a problem when a failure occurs in a transmitting end of a router.

5 is a diagram for explaining a problem in GRE PIC failure.

6 is an operation flowchart according to an embodiment of the present invention.

7A and 7B are diagrams illustrating setting and operation of the S-Ping.

8 is a diagram illustrating a failure detection method by an S-Ping operation when a failure occurs at a transmitting end of a router.

9 is a view for explaining a method of switching to a spare path when a failure is detected using S-Ping.

10 is a structural diagram of a GRE header.

FIG. 11 illustrates a method of switching paths using a GRE header value when a GRE PIC fails. FIG.

The present invention relates to a method for detecting a failure that occurs when a GRE tunnel is created. In particular, when the auto negotiation function is not used, a failure detection for preventing traffic loss due to continuous transmission of packets to a failed device. It is about a method.

In the Ethernet section, autonegotiation is used on the interfaces of devices for duplex, speed, and status checks. When the auto negotiation function is used, when a failure occurs in the transmitting terminal, the receiving terminal detects this and reports the failure of the transmitting terminal. Such auto-negotiation function is properly performed between the same equipment, but there is a problem in that the use parameters, thresholds, applied technologies, etc., are different from each other, and are not performed properly.

1 shows the results of autonegotiation measurement between homogeneous routers. Routers 1 and 2 have a transmitting end and a receiving end, respectively. The transmitting end A of the router 1 is connected to the receiving end B of the router 2, and the transmitting end C of the router 2 is connected to the receiving end D of the router 1. As shown in the table, when at least one of the two routers fails without distinguishing the transmission / reception line, the traffic is transferred to the SBY path by bringing down the state of the corresponding router.

Figure 2 shows the results of automatic negotiation between heterogeneous routers. If only one of the two routers is broken, the problem is that the state of the router is displayed as up state rather than down state, which is the actual state (in parentheses). Because of this problem, it is common not to adopt auto negotiation between heterogeneous routers.

However, when a transmission error occurs when the auto-negotiation function is not used, the reception terminal does not report a failure, and thus a transmission terminal does not know that its interface is down. If you perform static routing on the physical service path (hereinafter referred to as the ACT path) and perform dynamic routing on the spare path (hereinafter referred to as the SBY path), the counterpart equipment designated as the next link, even though that static path is down Since the interface is still delivered in the up state, the traffic is not transferred to the SBY path, but the traffic is continuously sent to the down ACT path. Therefore, all of the transmitted traffic is lost.

3 illustrates a problem that may occur when the auto negotiation function is not used between heterogeneous routers. As shown in the table, if the receiving end (B or D) fails, the router including the receiving end is down.However, if the transmitting end (A and B) fails, the router including the transmitting end maintains the up state. Instead, a problem arises where it does not transfer to the new path.

FIG. 4 illustrates a problem that may occur when there is a failure in the transmitting end of router 2 when the auto negotiation function is not used between heterogeneous routers as shown in FIG. 3. The path from router 1 to router 2 is the ACT path, and the path from router 1 to routers 3 and 4 is the SBY path. As shown in FIG. 3, even if a failure occurs in the transmitting terminal C of the router 2, the router 2 is displayed as an up state, and thus the router 2 is designated as the next router in the routing information of the router 1. Therefore, when transmitting data from router 1 to network 10.10.10.0, the ACT path is used as it is and is not transferred to the SBY path. At this time, since the transmitting end C of Router 2 fails, the packet does not reach the network 10.10.10.0.

Also, GRE PIC down problem, which is the biggest problem in generating Generic Routing Encapsulation (GRE) tunnel, is mentioned. GRE PIC is the name of the router unit used to provide the GRE tunnel. GRE PIC is a logical element that operates independently from the physical transmit / receive interface and is processed after the interface path. Therefore, even if the current GRE PIC is down, there is a problem that one router cannot notify the other router that the other router is down. Because of this problem, even if the GRE PIC of the other router is down, traffic destined for the downlink GRE tunnel continues to be lost, and all of the traffic is lost, and these traffic are directed to the routing engine and eventually down to the routing engine. .

5 is a diagram illustrating a problem when a failure occurs in the GRE PIC. In the drawing, routers 1 and 2 have transmitting terminals A and C and receiving terminals B and D, respectively, which are normally operating. For example, when the GRE PIC of Router 1 is up and the GRE PIC of Router 2 is down, Router 1 does not know the GRE PIC state of Router 2, which causes a problem of recognizing an up state.

The present invention introduces a new Status-Ping (S-Ping) similar to the ping used in the prior art to solve the problems of the prior art, and automatically shuts down and no shutdown commands of the router. In order to solve the problem that the state of the router remains up when the transmitter is down, the routing path is automatically switched to the SBY path.

In addition, an object of the present invention is to prevent a router from going down despite the down of the GRE PIC by providing a mechanism for differently setting and transmitting a value of a field that is not used in the GRE header when it is down and normal.

The present invention, when the GRE tunnel is generated between the other equipment and the other equipment connected via the data transmission path, the step of transmitting the GRE PIC state information of the other equipment from the other equipment to the one equipment; Grasping GRE PIC state information of the other device by the one device; And generating a GRE tunnel in the data transmission path when the GRE PIC state is normal, and blocking the data transmission path and activating a spare path when the GRE PIC state is abnormal.

Hereinafter, with reference to the accompanying drawings will be described in detail an embodiment of the present invention.                     

6 is a flowchart illustrating a failure detection method according to the present invention. Initially, the ACT path is set along the optimal path and data is transmitted through the set ACT path (S100). In the next step, the S-Ping is set (S200), and the S-Ping request is periodically sent (S300) to check whether the S-Ping response is normally received (S400).

If the S-Ping response is normally received, since the other party's router is in a normal state, data is continuously transmitted through the same path (S500). If it is determined that the GRE tunnel is generated (S510), if it is not the case of generating the GRE tunnel, data is continuously transmitted through the ACT path (S100). When the GRE tunnel is generated, the GRE PIC is down (S520), and if not, the data is continuously transmitted through the ACT path (S100). If the GRE PIC is down, the PIC status bit of the GRE header is set to '1' to be transmitted (S530) to inform the user that the down is down, and traffic is transferred to the SBY path (S540).

If the S-Ping response is not normal, a blocking command is performed on the corresponding interface (S600), and traffic is transferred to the SBY path (S610). Subsequently, it is checked whether the ACT path is restored (S620), and if not, the traffic is transferred to the SBY path (S610). If it is recovered, execute the block prohibition command to the ACT interface (S630), and transfers traffic to the ACT path (S640).

FIG. 7A illustrates an S-Ping setting item in router 1, and FIG. 7B is a diagram illustrating an operation according to the item set in FIG. 7A (see FIG. 6). "Interface: 20.20.20.0" represents the address of the receiving end B of Router 2. "Repeat time: 2 (sec)" represents a time interval for sending an S-Ping message from Router 1. "Duration: 2-30 (1: month, 2: day, 3: hour, 4: minute)" represents the period of time for sending an S-Ping message from Router 1. "Threshold of death: 2 (default: 1)" represents the number of response failures that are recognized as down when the router 2 does not respond to the S-Ping from the router 1. "Rollback Threshold: 5 (default: 3)" indicates the number of responses that Router 2 recognizes as normal when there is a response to S-Ping and returns to the original path.

In FIG. 7B, '!' Indicates that Router 1 received a response from Router 2, and '.' Indicates that Router 1 did not receive a response from Router 2. '.', Indicated by the indication number 10, indicates that there is no third response. Therefore, since "Threshold of death" is currently set to 2, Router 1 switches from the ACT path to the SBY path from that point on. '!', Indicated by the indication number 20, indicates that there is a sixth response after the transfer to the SBY path. Since "Rollback Threshold" is set to 5, Router 1 returns from the SBY path to the original ACT path from that point on.

Unlike the conventional extended ping, the S-Ping has been reduced to 5 setting items. The difference is that the setting threshold and measurement of the S-Ping are linked to the router's interface blocking / blocking to allow switching to the SBY path. There is.

8 shows a state in which a router is correctly displayed by using S-Ping even when a failure occurs in the transmitting terminal C of the router. Router 2 does not respond to the S-Ping message sent from Router 1 when the transmitting end fails. After the set number of times, Router 1 recognizes Router 2 as down and switches over to the SBY path. If the router 1 fails, router 2 cannot respond to the S-Ping message sent by router 2, and router 2 recognizes that router 1 is down and switches to the SBY path.

9 is a diagram illustrating a process of automatically switching to the SBY path through S-Ping interworking. It is assumed that routers 1 and 2 have transmitting terminals A and C and receiving terminals B and D, respectively, A is connected to B and C with D, and that a transmitting terminal C of router 2 has failed. The ACT path connects to Network 10.10.10.0 through Router 1, Router 2, and the SBY path connects to Network 10.10.10.0 through Router 1, Router 3, and Router 4.

Router 1 forwards S-Ping messages using the ACT path. Router 2 does not respond to the S-Ping message due to a failure in the sender (C). Router 1 cuts the interface connected to router 2 when the response is not transmitted to the predetermined number of messages from router 2, and then switches to the SBY path.

10 shows the structure of a GRE header. The reserved field (Reserved0) of the GRE header consists of a total of 12 bits. Currently, 1 to 5 bits are bits already defined in RFC1701, and 6 to 12 bits are not used. In this embodiment, one bit of 6 to 12 bits which are not used is used to convey the state of the GRE PIC. For example, the bit may be set to '0' to indicate an up state, and the bit may be set to '1' to indicate a down state. If the corresponding bit is '1', switch to SBY path.

FIG. 11 is a diagram illustrating a process of switching to an SBY path by transmitting a GRE PIC state using a GRE header. FIG. It is assumed that routers 1 and 2 have transmitting terminals A and C and receiving terminals B and D, respectively, A is connected to B and C with D, and GRE PIC of router 2 has failed. The ACT path connects to Network 10.10.10.0 through Router 1, Router 2, and the SBY path connects to Network 10.10.10.0 through Router 1, Router 3, and Router 4.

Router 1 forwards traffic along the ACT path. If the GRE PIC of Router 2 fails, Router 2 changes the GRE PIC status bit in the GRE header to '1' and sends it to Router 1. Router 1 recognizes that GRE PIC of Router 2 is down and switches over to the SBY path.

By applying the present invention, when the automatic negotiation function is not used, the packet is discarded because the state of the corresponding device is displayed as normal despite the failure occurring at the transmitting end of the counterpart device. In addition, even when an error occurs in the GRE PIC of the other device when creating the GRE tunnel, the other device may not recognize the problem.

Claims (4)

delete In a case where a GRE tunnel is created between one device and another device connected through a data transmission path, Transmitting GRE PIC state information of the other device from the other device to the one device; Grasping GRE PIC state information of the other device by the one device; And Creating a GRE tunnel in the data transfer path if the GRE PIC state is normal; and blocking the data transfer path and activating a spare path if the GRE PIC state is abnormal. Network failure detection method comprising a. The method of claim 2, wherein the GRE PIC state information is transmitted by using an unused field among fields of a GRE header. Periodically transmitting a confirmation message from one device to the other device connected through the data transmission path with the one device; Blocking the data transmission path and transmitting data through the spare path if the one device does not continuously receive the response message for the confirmation message from the other device for a predetermined period of time; Blocking the preliminary path and activating the blocked data transmission path to transmit data when the one device continuously receives a response message from the other equipment through the blocked data transmission path for a predetermined period of time; Transmitting GRE PIC state information of the other device from the other device to the one device when the GRE tunnel is created between the one device and the other device through the data transmission path; And Creating a GRE tunnel in the data transfer path if the GRE PIC state is normal; and blocking the data transfer path and activating a spare path if the GRE PIC state is abnormal. Network failure detection method comprising a.

Images