CN1681254A

CN1681254A - Maintaining method of Ethernet link state

Info

Publication number: CN1681254A
Application number: CN 200410034427
Authority: CN
Inventors: 曹阳; 闫长江; 瞿祥明
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2004-04-08
Filing date: 2004-04-08
Publication date: 2005-10-12
Anticipated expiration: 2024-04-08
Also published as: CN100499485C

Abstract

The method includes following steps: when two layer 3 equipments pass through Ethernet link connection, one of the layer 3 equipments periodically send link state detection message; after opposite terminal equipment receives the message, it will make response; when the present terminal equipment doesn't receive the response from opposite terminal equipment, this link is considered have a failure, meantime disenable the router whose next-hop address is the address of opposite terminal equipment. The invention can automatically switch the failure link to spare link, and when the network failure recovers, it can switch back.

Description

A kind of ethernet link state-maintenance method

Technical field

The present invention relates to network technology, particularly the detection of network link failure and processing method.

Background technology

When connecting by the Ethernet net between router and the router or between any two three-layer network appliances, can directly connect, perhaps connect by one or more Layer 2 switch by ethernet line.Be illustrated in fig. 1 shown below; direct-connected by ethernet line; because line quality; perhaps contact badly, such phenomenon takes place through regular meeting, can operate as normal from RTA to RTB; but from RTB to RTA, but break down; RTA also can continue information is sent to RTB like this, but RTB can not return information, thereby causes network failure.When connecting by Layer 2 switch, problem is just more serious.During to the connection fault between RTB and the Layer 2 switch, RTA does not also know, causes network failure equally.After these faults took place, network can not switch on the reserve link automatically because it and do not known own fault.In order to address this is that, can only solve problem by configuration dynamic routing protocol on RTA and RTB at present.Because the agreement maintenance packet between the dynamic routing protocol all needs the other side to confirm, so, no matter be to break down in any point between RTA and the RTB, can detect by dynamic routing protocol and to break down, thereby network is switched on the backup network, thereby guarantee that whole network continues operate as normal.Do like this and bring some new problems:

1, the requirement to the network equipment increases.A lot of network equipments are not support dynamic routing protocol as three-tier switch etc.

2, increase the load of the network equipment.Start dynamic routing protocol and use static routing to compare, largely increase consumption equipment CPU and internal memory.

3, increase the complexity of network operation.At the edge of network, as the connection between three-tier switch and the router, generally all adopt static routing protocol, but, must all start dynamic routing protocol in order to guarantee the reliability of network, increased the complexity of network greatly.Especially in the MPLS network, what network will become is unusual complicated.

Summary of the invention

Based on above analysis, when two three-layer equipments connect by Ethernet link, need a kind of simple technology, detect Link State.

The present invention proposes a simple technique scheme and solve this problem, when two three-layer equipments connect by Ethernet link, one of them three-layer network appliance regularly sends the Link State detection messages, and opposite equip. is replied by opposite equip. after receiving this message.When local terminal equipment is not received replying of opposite equip. in the regular hour, just think this link fault, it is invalid with next hop address to be that the route of opposite equip. address is changed to simultaneously.

The present invention makes the network equipment can find the ethernet link fault by Ethernet link state-detection mechanism is provided, by with relevant static routing mating reaction, network is switched on the reserve link from faulty link automatically.Behind network disaster recovery, can switch back again.By using this technology, the network equipment is as long as support static routing just can realize the ethernet link backup, and is low for equipment requirements, and needs the extra resource of the equipment that takies hardly, greatly facilitates the construction and the maintenance of edge network.

Description of drawings

Fig. 1 is a network equipment networking schematic diagram;

Fig. 2 totally implements block diagram representation for the present invention;

Fig. 3 is a testing process schematic diagram of the present invention;

Embodiment

For making the purpose, technical solutions and advantages of the present invention clearer, the present invention is described in further detail below in conjunction with the drawings and specific embodiments.

As Fig. 1, be enforcement block diagram of the present invention, implement the present invention and mainly contain following steps;

Step 10: determine the next hop address that local terminal equipment will detect by configuration;

Step 11: check whether disposed the Static ARP of this next hop address, if there is not jump procedure 17;

Step 12: timed sending ARP request (can pass through broadcasting or clean culture);

Step 13: do not receive the back message using of arp request whether continuous three times, if, jump procedure 19;

Step 14: received response message, whether former interface link state is down, and whether relevant route is invalid, if do not receive jump procedure 12;

Step 15: Link State is set to effectively, and route is set to effectively

Step 16: after the route of main usefulness is effective, replace original route stand-by to have an effect automatically, jump procedure 12;

Step 17:, set up the dynamic ARP entry of this next jumping by the Dynamic ARP request

Step 18: whether set up this ARP list item, if do not set up jump procedure 12

Step 19: judge that then the opposite end is obstructed, the local terminal link is set to the down state, and it is invalid that the route entry of will being correlated with simultaneously is changed to

Step 20: after top route entry was invalid, the route of backup was had an effect automatically

As Fig. 2, get access to the MAC Address of opposite end three-layer network appliance when equipment after (no matter be or pass through static configuration) by the ARP dynamic learning, regular transmission probe messages is to opposite equip., for avoiding causing network burden, probe messages sends with the form of clean culture, the target MAC (Media Access Control) address of this message is the MAC Address of opposite end three-layer network appliance.The regular time is configurable, in order to find network failure as early as possible, reduces the unnecessary load of network simultaneously as far as possible, and recommended value is 1 second to 10 seconds.The content of probe messages can be various, can be special protocol massages, can use the prior protocols message, as clean culture ARP request message (target MAC (Media Access Control) address of ARP request message is known MAC Address), perhaps ICMP request message.Probe messages receives that with after sending together request message replying of opposite end is a detection process of finishing.

Behind the ARP list item of the IP address correspondence that the opposite end that disposes has been arranged in the ARP table, just the ARP that can be correlated with has surveyed.The form of probe messages is as follows:

Can use the ARP message request of broadcasting, message format is as follows:

Sender′s?MAC?Address	Sender′s?IP?Address	Target?MAC?Address	Target?IP?Address
Sender′s?MAC?Address	Sender′s?IP?Address	Target?MAC?Address	Target?IP?Address	00-00-0c-94-36-aa (local terminal eth MAC Address)	(172.16.10.100 local terminal IP address)	00-00-00-00-00-00 (broadcast address)	172.16.20 200 (peer IP address)

Can use the ARP message request of clean culture, message format is as follows:

Sender′s?MAC?Address	Sender′s?IP?Address	Target?MAC?Address	Target?IP?Address
Sender′s?MAC?Address	Sender′s?IP?Address	Target?MAC?Address	Target?IP?Address	00-00-0c-94-36-aa (local terminal eth MAC Address)	(172.16.10.100 local terminal IP address)	00-00-0e-95-45-55 (the mac address of arriving by static configuration or dynamic learning)	(172.16.20.200 peer IP address)

If reciever does not respond, if response times does not surpass three times, then think and the other side down continue to send then, if response is arranged, then show up.

After continuous several times (as 3 times) is not received the response message of opposite end, just think that network breaks down, in this equipment, next jumping is the route of three-layer network appliance address, opposite end, all is changed to failure state.Like this, if when in the network backup link being arranged, route stand-by will come into force, thereby network is switched on the backup link, keeps whole network operate as normal.

As Fig. 1, RTA, RTB, RTC connect by ethernet link, the middle LanSwith that passes through two layers.Connect internal network under the RTA, the internal network address network segment is 50.0.0.0/24.RTA is first outlet of in-house network to RTB, and RTA is second outlet of enterprise to RTC.Under the normal condition, two default routes are arranged on RTA, divide to be clipped to RTB and RTC, will be higher than RTC's to the priority of the route of RTB.Like this, all uplink messages all are sent to RTB, deliver to external network then.Though RTB and RTC have issued the network segment of 50.0.0.0 to external network, the priority of RTB issue will be higher than RTC, so downlink message all is sent to RTB, delivers to internal network by RTA.

When RTB after the link occurs fault of LanSwitch, according to present processing method, RTA does not also know this link occurs fault, continues message is sent to 100.0.0.2, final LanSwitch is all packet loss, thereby causes network failure.

Adopt technical scheme of the present invention, specifying 100.0.0.2 and 100.0.0.3 on RTA is the three-layer network appliance address, and all specifying 100.0.0.1 on RTB and RTC is the three-layer network appliance address.Like this, all regularly whether (every 3 seconds) transmission probe messages is normal with the detecting link state for RTA, RTB and RTC.With RTA is example, can send destination address is 100.0.0.2, target MAC (Media Access Control) address is that the ARP request message of B.B.B (is determined because of target MAC (Media Access Control) address, so be unicast message), LanSwitch delivers to this message on the RTB, RTB receives after this ARP request message, can answer an arp reply message, returns to RTA.After the link occurs fault between RTB and the Lanswtich, the probe messages of RTA also just can't send to RTB, and natural RTA just can not receive the response message of RTB yet, continuous do not have response message three times after, then think this link failure.Search that next hop address is the route of 100.0.0.2 in the routing table, and this route was lost efficacy.Like this, article one default route that RTA comes into force in Central Plains, 0.0.0.0 0.0.0.0 100.0.0.2/24 10 had just lost efficacy, the second default route, 0.0.0.0 0.0.0.0 100.0.0.3/24 20 has just come into force.Like this, uplink message is just delivered to RTC by RTA.For RTB, the continuous too response message of not receiving RTA for three times, the route that can be 100.0.0.1/24 with next hop address forbids falling, being about to route 50.0.0.0255.255.255.0 100.0.0.1/24 10 forbids falling, like this, external network just can only be received from the route of the 50.0.0.0 network segment of RTC issue.So all downlink messages just are sent to RTC, deliver to internal network by RTA.Like this, realized that RTA from fault is to the backup link of link switchover to RTA to RTC of RTB.

After RTA was normal to the link-recovery of RTB, originally forbidden route can come into force again, and network returns to again and uses the link of RTA to RTB.

Claims

1, a kind of ethernet link fault detect way, it is characterized by, local terminal Equipment Inspection configuration, local terminal equipment sends detection messages to detected link opposite equip., opposite equip. is received detection messages, reply response message, whether local terminal equipment basis receives response message is judged the state of link, and makes corresponding troubleshooting.

2, ethernet link fault detect way according to claim 1, it is characterized in that, the inspection of described local terminal equipment disposition comprises, if be checked through ARP is not quiesce protocol, then set up the dynamic address resolution agreement ARP list item of next jumping, if do not set up dynamic address resolution agreement ARP list item, then be switched to route stand-by.

3, ethernet link detection method according to claim 1 is characterized in that, described transmission message can be common dynamic address resolution agreement ARP request message clean culture ARP request message, or Internet Control Message Protocol ICMP request message.

4, ethernet link detection method according to claim 1 is characterized in that, described transmission message is the cycle transmission with 1 second to 10 seconds, does not receive corresponding message if send 3 messages continuously, then judges this link failure.

5, ethernet link detection method according to claim 1 is characterized in that, described troubleshooting comprises that the local terminal link is set to the down state, and it is invalid that the route of being correlated with is set to, and be transformed on the route stand-by.