CN101651616B

CN101651616B - Main link recovery control method, exchanger system and exchangers

Info

Publication number: CN101651616B
Application number: CN2009100923256A
Authority: CN
Inventors: 苏楠枝
Original assignee: Hangzhou H3C Technologies Co Ltd
Current assignee: New H3C Technologies Co Ltd
Priority date: 2009-09-04
Filing date: 2009-09-04
Publication date: 2011-09-28
Anticipated expiration: 2029-09-04
Also published as: CN101651616A

Abstract

The invention discloses a main link recovery control method, an exchanger system and exchangers. When a main link fails and does not cause the exchanger system to split, a new main link can be found by the interaction of all exchangers as long as a downlink exchanger of the failure main link has other direct connection or cascade links leading to a main exchanger. In this way, even if the exchangers on both ends of the failure main link do not have other direct connection links, reconfiguration for reselecting the main exchanger is not initiated, thereby avoiding that the exchanger system cannot transmit data or even lose the data within longer time, and further enhancing the reliability of the exchanger system.

Description

Main link recovery control method and switch system and switch

Technical field

The present invention relates to the link-recovery control technology, be particularly related to a kind of main link recovery control method, based on optical-fibre channel (Fibre Channel, FC) network and can realize a kind of switch system that primary link recovers and all kinds of switches in this switch system.

Background technology

In the existing switch system based on the FC network; usually can be by setting up optical fiber (Build Fabric between each switch; BF) configuration and select out a host exchange, other switches are non-host exchange; primary link as herein described just is that non-host exchange is used for the link of communicating by letter with host exchange, and primary link both can be link direct-connected between non-host exchange and the host exchange, also can be link by other non-host exchange cascades.

For every non-host exchange, the primary link that is used for sending to host exchange data can be referred to as the upstream primary link, be used for and will then can be referred to as the downstream primary link from the data forwarding of host exchange to the primary link in downstream.That is to say, divide according to data flow, primary link can be divided into upstream primary link and downstream primary link, either one or the other, host exchange does not have the upstream primary link but has at least one downstream primary link, but not host exchange then has only a upstream primary link and has zero or at least one downstream primary link.

In the specific implementation, no matter be that host exchange also is non-host exchange, all have the interconnected pattern fibre channel port of some switches (E_Port), between host exchange and the non-host exchange, and can set up available non-primary link by the link negotiation between the E_Port earlier between each non-host exchange, thereafter by the selected primary link from non-primary link alternately between host exchange and each the non-host exchange, write down the E_Port that it connects the downstream primary link by host exchange again, write down it by each non-host exchange and connect the upstream, the E_Port of downstream primary link so just can realize non-host exchange communicating by letter by primary link and host exchange.

Referring to Fig. 1, suppose to have switch 1～7, switch 1 to be host exchange:

Link ISL1 between switch 1 and the switch 2 is the downstream primary link of switch 1, the upstream primary link of switch 2;

Link ISL3 between switch 2 and the switch 3 is the downstream primary link of switch 2, the upstream primary link of switch 3, and promptly the primary link between switch 3 and the switch 1 is that cascade, the switch 3 of ISL1 and ISL3 do not have the downstream primary link;

Link ISL4 between switch 2 and the switch 5 is non-primary link;

Link ISL2 between switch 1 and the switch 4 is the downstream primary link of switch 1, the upstream primary link of switch 4;

Link ISL5 between switch 4 and the switch 5 is the downstream primary link of switch 4, the upstream primary link of switch 5, and promptly the primary link between switch 5 and the switch 1 is the cascade of ISL2 and ISL5;

Link ISL6 between switch 5 and the switch 6 is the downstream primary link of switch 5, the upstream primary link of switch 6, and promptly the primary link between switch 6 and the switch 1 is the cascade of ISL2, ISL5, ISL6;

Link ISL7 between switch 6 and the switch 7 is the downstream primary link of switch 6, the upstream primary link of switch 7, and promptly the primary link between switch 7 and the switch 1 is that cascade, the switch 7 of ISL2, ISL5, ISL6, ISL7 do not have the downstream primary link.

In the practical application, primary link between host exchange and the direct-connected non-host exchange or the primary link between any two non-host exchanges all might lose efficacy, and may occur following two kinds of situations when primary link lost efficacy:

Wherein a kind of situation, switch system is owing to the inefficacy of primary link is split into two or more, referring to Fig. 2 a, suppose that ISL2, ISL4 lost efficacy, switch system then is divided into two switch systems being divided as dotted line broken line among Fig. 2 a as shown in Figure 1, just need initiate BF this moment respectively and reshuffle again selected host exchange in two switch systems.

Another kind of situation, switch system is owing to the inefficacy of primary link is divided, referring to Fig. 2 b, suppose that ISL5 lost efficacy, then the switch 5 in the switch system still links to each other with switch 2 by non-primary link ISL4 as shown in Figure 1, at this moment, if also there are other direct-connected links between the switch at inefficacy primary link two ends, for example except failed ISL5, also have other direct-connected links between switch 5 among Fig. 2 b and the switch 2, then can from other links, select one in the following manner and be used as new downstream primary link:

Switch (can be referred to as upstream switches) for inefficacy primary link upstream extremity, after detecting its downstream primary link inefficacy, switch (can be referred to as downstream switch) by aforementioned other chain road direction inefficacy primary link downstream sends an exchange optical fiber parameter (Exchange Fabric Parameters, EFP) message, if at the stabilized fiber (Fabric_Stabible_Timeout that fixes time, F_S_TOV, be generally about 5 seconds) in receive that the switch of downstream switch replys (SW_ACC) message, then use these other links as new downstream primary link, correspondingly, downstream switch is then responded the SW_ACC message as long as receive the EFP message from aforementioned other links, and use receives these other links of EFP message as new upstream primary link; If upstream switches is not received the SW_ACC message that downstream switch is responded in F_S_TOV, or the switch that receives the downstream switch response is refused (SW_RJT) message, then triggering BF reshuffles, if and downstream switch is not received the EFP message in 2F_S_TOV, also can trigger BF and reshuffle.

As above as seen, existing primary link reset mode is only applicable between two direct-connected switches of multilink, thereby lost efficacy at primary link and not cause under the situation of switch system division, even if downstream switch can also be connected to host exchange by other non-primary links, also very likely initiating BF reshuffles, and the time that BF reshuffles needs 3F_S_TOV at least, promptly about 15 seconds, thereby cause switch system can't transmit data even obliterated data in a long time, and then make that the reliability of switch system is not high.

Summary of the invention

In view of this, the invention provides a kind of main link recovery control method and can realize all kinds of switches in the switch system that primary link recovers, can under primary link lost efficacy the situation that does not cause the switch system division, improve the reliability of switch system.

A kind of main link recovery control method provided by the invention, this control method comprises the steps:

During downstream switch that switch lost efficacy as the upstream primary link, the interconnected pattern fibre channel port of all normal switches E_Port sends primary link and recovers to survey the PIRD message to it; If have E_Port after the upstream primary link lost efficacy, default first to receive that switch replys the SW_ACC message at interval, then this E_Port place link be recorded as new upstream primary link;

Switch is a non-host exchange and during as the intermediary switch in other links, if receive the E_Port place link of PIRD message is that non-primary link and this intermediary switch exist the upstream primary link, then responding the SW_ACC message, write down this non-primary link in E_Port place to the E_Port that receives the PIRD message is the downstream primary link, and transmits the PIRD message, receives the SW_ACC message of responding then to the upstream of this intermediary switch primary link E_Port; If the E_Port place link that receives the PIRD message is for the upstream primary link of this intermediary switch and also have other normal E_Port except that this E_Port, then send the PIRD message to other normal all E_Port, and the SW_ACC message that other normal E_Port receive is forwarded to the E_Port that receives the PIRD message, upstream, the E_Port place primary link that the E_Port place link that also will receive SW_ACC is recorded as the upstream primary link, will receive the PIRD message is recorded as the downstream primary link; If the E_Port place link that receives the PIRD message is the downstream primary link of this intermediary switch, then transmits PIRD messages, receive the SW_ACC message of responding then to other all primary link E_Port;

Switch is a host exchange and during as the intermediary switch in other links, responds the SW_ACC message to the E_Port that receives the PIRD message, and transmits PIRD messages, receives the SW_ACC message of responding then to other all primary link E_Port; If receiving the E_Port of PIRD message is non-primary link, then this non-primary link in E_Port place is recorded as new downstream primary link;

During upstream switches that switch lost efficacy as the downstream primary link, if have in default second interval of E_Port after the downstream primary link lost efficacy and receive the PIRD message, then respond the SW_ACC message to this E_Port, if and this E_Port place link is non-primary link, then this E_Port place link is recorded as new downstream primary link.

Switch is a non-host exchange and during as the intermediary switch in other links, after other normal all E_Port send the PIRD message, have more than other normal E_Port and receive the SW_ACC message, the SW_ACC message that then will receive at first is forwarded to the E_Port that receives the PIRD message, and the E_Port place link that will receive SW_ACC at first is recorded as the upstream primary link, the upstream primary link that will receive the E_Port place of PIRD message is recorded as the downstream primary link;

During downstream switch that switch lost efficacy as the upstream primary link, have more than receiving the SW_ACC message in default first interval of an E_Port after the upstream primary link lost efficacy, the E_Port place link that then will receive the SW_ACC message at first is recorded as new upstream primary link.

Include the switch name of the command code that identifies this message, described inefficacy primary link middle and upper reaches switch and the switch name of described inefficacy primary link middle and lower reaches switch in the PIRD message.

Described switch name is global unique identification WWN.

A kind of switch provided by the invention, the upstream primary link with some switches interconnected pattern fibre channel port E_Port and this switch lost efficacy, and this switch comprises:

Packet sending unit sends primary link to all normal E_Port and recovers to survey the PIRD message;

Recover timing unit, after the upstream primary link loses efficacy, picked up counting, and timing finishes when first interval is preset in arrival;

Recover decision unit,, then this E_Port place link is recorded as new upstream primary link if there is E_Port to receive that switch replys the SW_ACC message before finishing recovering the timing unit timing.

Have more than an E_Port and recovering the timing unit timing when receiving the SW_ACC message before finishing, recover the E_Port place link that decision unit further will receive the SW_ACC message at first and be recorded as new upstream primary link.

Include the command code that identifies this message, the switch name of inefficacy upstream primary link other end upstream switches and the switch name of described switch in the PIRD message.

Described switch name is global unique identification WWN.。

Another kind of switch provided by the invention has some switches interconnected pattern fibre channel port E_Port and this switch and is not chosen as host exchange and is arranged in any other links outside the inefficacy primary link, and this switch comprises:

The link identification unit is judged to receive any E_Port place Linktype that primary link recovers to survey the PIRD message;

Transmit decision unit, if receive the E_Port place link of PIRD message is that non-primary link and described switch exist the upstream primary link, then responding the SW_ACC message, write down this non-primary link in E_Port place to the E_Port that receives the PIRD message is the downstream primary link, and transmits the PIRD message, receives the SW_ACC message of responding then to the upstream of described switch primary link E_Port; If to receive the E_Port place link of PIRD message be the upstream primary link of described switch and also have other normal E_Port except that this E_Port, then send the PIRD message to other normal all E_Port, and the SW_ACC message that other normal E_Port receive is forwarded to the E_Port that receives the PIRD message, the upstream primary link that the E_Port place link that also will receive SW_ACC is recorded as the upstream primary link, will receive the E_Port place of PIRD message is recorded as the downstream primary link; If receiving the E_Port place link of PIRD message is the downstream primary link of described switch, then transmits PIRD messages, receive the SW_ACC message of responding then to other all primary link E_Port.

Have more than other normal E_Port and receive the SW_ACC message, then transmit the SW_ACC message that decision unit further will receive at first and be forwarded to the E_Port that receives the PIRD message, and the E_Port place link that further will receive SW_ACC at first is recorded as the upstream primary link, the upstream primary link that will receive the E_Port place of PIRD message is recorded as the downstream primary link.

Described switch name is global unique identification WWN.。

Another switch provided by the invention, the downstream primary link with some switches interconnected pattern fibre channel port E_Port and this switch lost efficacy, and this switch comprises:

Recover timing unit, when the downstream primary link loses efficacy, picked up counting, and when second interval is preset in arrival, finish timing;

Recover decision unit, recover to survey the PIRD message if there is E_Port before recovering timing unit timing end, to receive primary link, then respond switch and reply the SW_ACC message to this E_Port, if and this E_Port place link is non-primary link, then this E_Port place link is recorded as new downstream primary link.

Include the switch name of the command code that identifies this message, described switch and the switch name of inefficacy downstream primary link other end downstream switch in the PIRD message.

Described switch name is global unique identification WWN.

As seen from the above technical solution, when the present invention does not cause the switch system division in the primary link inefficacy, because upstream switches and intermediary switch might be host exchange, also might be direct-connected or be cascaded to host exchange by primary link, therefore, as long as the downstream switch of inefficacy primary link possesses other links direct-connected or cascade towards upstream switches or intermediary switch, promptly possess other links direct-connected or cascade towards host exchange, then can utilize and find new primary link alternately between each switch, like this, even if not existing other direct-connected links also can not initiate BF between the switch at inefficacy primary link two ends reshuffles, thereby can avoid switch system can't transmit data in a long time, even obliterated data, and then can improve the reliability of switch system.

Description of drawings

Fig. 1 is the structural representation of existing a kind of switch system example;

Fig. 2 a was for causing the schematic diagram of switch system division owing to primary link loses efficacy in the switch system example as shown in Figure 1;

Fig. 2 b does not cause the schematic diagram of switch system apportion for primary link in the switch system example as shown in Figure 1 lost efficacy;

Fig. 3 is the exemplary flow schematic diagram among the main link recovery control method embodiment one of the present invention;

Fig. 4 a～4b is the example schematic among the main link recovery control method embodiment one of the present invention;

Fig. 5 is the exemplary flow schematic diagram among the main link recovery control method embodiment two of the present invention;

Fig. 6 a～6b is the example schematic among the main link recovery control method embodiment two of the present invention;

Fig. 7 is the exemplary flow schematic diagram among the main link recovery control method embodiment three of the present invention;

Fig. 8 is the example schematic among the main link recovery control method embodiment three of the present invention.

Embodiment

For making purpose of the present invention, technical scheme and advantage clearer, below with reference to the accompanying drawing embodiment that develops simultaneously, the present invention is described in more detail.

In the embodiment of the invention, primary link between host exchange and direct-connected non-host exchange or the primary link between any two non-host exchanges lost efficacy and (need to prove, inefficacy primary link as herein described, for mode direct-connected between non-host exchange and the host exchange, promptly be meant the complete direct-connected primary link of inefficacy; And for the mode of cascade between non-host exchange and the host exchange, then be meant a part that lost efficacy in the complete primary link, be not the complete primary link that partial failure is arranged) when not causing the switch system division, as long as the downstream switch of inefficacy primary link possesses other links direct-connected or cascade towards host exchange, then can utilize and find new primary link alternately between each switch, like this, even if not existing other direct-connected links also can not initiate BF between the switch at inefficacy primary link two ends reshuffles, thereby can avoid switch system can't transmit data in a long time, even obliterated data, and then can improve the reliability of switch system.

In the switch system that primary link lost efficacy, switch (comprising host exchange and non-host exchange) is divided, also can be regarded as according to the role and divide according to its position at the inefficacy primary link, can be divided into following three kinds:

The downstream switch that the upstream primary link lost efficacy, the switch of the primary link downstream that promptly lost efficacy (need to prove, for the situation of non-host exchange and the complete primary link of the direct-connected formation of host exchange, the downstream switch that the upstream primary link lost efficacy promptly is the terminal non-host exchange of communicating by letter with host exchange of complete primary link; Constitute the situation of complete primary link for non-host exchange and host exchange cascade, when the inefficacy primary link is positioned at complete primary link latter end, the downstream switch that the upstream primary link lost efficacy promptly is the terminal non-host exchange of communicating by letter with host exchange of complete primary link, when the inefficacy primary link was positioned at the stage casing of complete primary link or connect first section of host exchange, the downstream switch that the upstream primary link lost efficacy promptly was the non-host exchange that is used for the complete primary link of cascade);

Other primary links outside the inefficacy primary link or the intermediary switch in the non-primary link (need to prove that for the situation of non-host exchange and the complete primary link of the direct-connected formation of host exchange, intermediary switch can be the non-host exchange in other link; Constitute the situation of complete primary link for non-host exchange and host exchange cascade, when the inefficacy primary link is positioned at first section of complete primary link, intermediary switch can be the non-host exchange in other link, when the inefficacy primary link is positioned at the stage casing of complete primary link or latter end, because detouring to other link of other end upstream switches, the downstream switch of inefficacy primary link one end might pass through host exchange, therefore, intermediary switch both might be in other link non-host exchange, also might be host exchange);

And, the upstream switches that the downstream primary link lost efficacy, the switch of primary link upstream extremity of promptly losing efficacy (need to prove that for the situation of non-host exchange and the complete primary link of the direct-connected formation of host exchange, the upstream switches that the downstream primary link lost efficacy promptly is a host exchange; Constitute the situation of complete primary link for non-host exchange and host exchange cascade, when the inefficacy primary link is positioned at first section of complete primary link, the upstream switches that the downstream primary link lost efficacy promptly is a host exchange, when the inefficacy primary link was positioned at the stage casing of complete primary link or latter end, the upstream switches that the downstream primary link lost efficacy promptly was the non-host exchange that is used for the complete primary link of cascade).

Main link recovery control method in the embodiment of the invention need be carried out corresponding step respectively by above-mentioned three kinds of switches:

1) switch lost efficacy as the upstream primary link downstream switch the time:

Downstream switch to it all normal (Up), promptly the E_Port by link negotiation sends primary link and recovers detection (Principal ISL Recovery Detect, PIRD) message, because the E_Port of inefficacy upstream primary link is Up no longer, thereby, send the PIRD message from the E_Port of other all Up, can detect other links direct-connected or cascade that except the inefficacy primary link, whether exist towards host exchange;

If have and receive the SW_ACC message that upstream switches or intermediary switch are replied at the PIRD message in the F_S_TOV of E_Port after the upstream primary link lost efficacy, then the E_Port place link of expression reception SW_ACC message can lead to host exchange, thereby this E_Port place link is recorded as new upstream primary link; Preferably, if have more than receiving the SW_ACC message in the F_S_TOV of an E_Port after the upstream primary link lost efficacy, then can only the E_Port place link that receives the SW_ACC message at first be recorded as new upstream primary link, and the E_Port of the follow-up SW_ACC of receiving message is ignored;

Do not receive the SW_RJT message that upstream switches or intermediary switch are replied if all receive SW_ACC message or all E_Port that upstream switches or intermediary switch are replied at the PIRD message in the F_S_TOV of all E_Port after the upstream primary link lost efficacy, then represent except the inefficacy primary link, to no longer include other links and can lead to host exchange, thereby trigger BF and reshuffle.

When 2) switch was as the intermediary switch in other links outside the inefficacy primary link, under a kind of situation, intermediary switch was not chosen as host exchange, promptly as non-host exchange, under the another kind of situation, intermediary switch is a host exchange:

2.1) be the previous case of non-host exchange for intermediary switch:

2.1.1) the E_Port place link that receives the PIRD message is non-primary link:

If there is the upstream primary link in this intermediary switch, represent that then this intermediary switch can be used as the new upstream switches of inefficacy primary link middle and lower reaches switch, thereby to the E_Port response SW_ACC message that receives the PIRD message, and the non-primary link in E_Port place of recorder PIRD message is new downstream primary link, and, also transmit the PIRD message to the upstream of this intermediary switch primary link, and the SW_ACC message responded of the switch that receives this intermediary switch upstream primary link other end, thereby receive the SW_ACC message in downstream switch, perhaps downstream switch and other intermediary switch that is not chosen as host exchange receive the E_Port that record behind the SW_ACC message sends the PIRD message, can realize that downstream switch passes through to set up between this intermediary switch and the host exchange primary link of new cascade;

If there is not the upstream primary link in this intermediary switch, represents that then this intermediary switch not can be used as the new upstream switches of inefficacy primary link middle and lower reaches switch, thereby respond the SW_RJT message to the E_Port that receives the PIRD message.

2.1.2) the E_Port place link that receives the PIRD message is the upstream primary link of this intermediary switch:

If except that this E_Port that receives the PIRD message, also there is the E_Port of other Up, then transmit the PIRD message to other normal all E_Port, to seek other link of alternative inefficacy primary link, and the SW_ACC message of the response that the E_Port of other Up is received is forwarded to the E_Port that receives the PIRD message, the E_Port place link that also will receive SW_ACC is recorded as the upstream primary link, the upstream primary link change that receives the E_Port place of PIRD message is recorded as the downstream primary link, thereby receive the SW_ACC message in downstream switch, after perhaps downstream switch and other intermediary switch that is not chosen as host exchange receive the SW_ACC message, record sends the E_Port of PIRD message, can realize that downstream switch passes through to set up between this intermediary switch and the host exchange primary link of new cascade; Preferably, after other normal all E_Port transmit the PIRD message, if there is E_Port to receive the SW_ACC message more than other Up, then the SW_ACC message that receives at first can be forwarded to the E_Port that receives the PIRD message, and the E_Port place link that will receive SW_ACC at first is recorded as the upstream primary link, the upstream primary link that will receive the E_Port place of PIRD message is recorded as the downstream primary link;

In addition, after other normal all E_Port transmit the PIRD message, if the E_Port of all other Up all receives the SW_RJT message of response, represent that then this intermediary switch not can be used as the new upstream switches of inefficacy primary link middle and lower reaches switch, transmit the SW_RJT message to the E_Port that receives the PIRD message;

If except that this E_Port that receives the PIRD message, there is not the E_Port of other Up, represent that then this intermediary switch can't utilize the downstream primary link that inefficacy primary link middle and lower reaches switch is connected to other and can lead to the host exchange intermediary switch, thereby respond the SW_RJT message to the E_Port that receives the PIRD message.

2.1.3) the E_Port place link that receives the PIRD message is the downstream primary link of this intermediary switch:

The PIRD message that intermediary switch is received from the downstream primary link normally receives 2.1.1), 2.1.2) the PIRD message transmitted to the host exchange direction of the intermediary switch of both of these case.

As previously mentioned, inefficacy primary link downstream switch has found new primary link to communicate by letter with host exchange, but this moment, this inefficacy primary link upstream switches is not also known, can not trigger BF and reshuffle if upstream switches is received the PIRD message in 2F_S_TOV.

Therefore, when intermediary switch receives the PIRD message at the downstream primary link, the PIRD message received is transmitted so that inefficacy primary link upstream switches can be received the PIRD message to all primary link E_Port (comprising its upstream primary link E_Port and other downstreams primary link E_Port), thereby the primary link upstream switches of avoiding losing efficacy has found in downstream switch and triggers BF under the situation of new primary link and reshuffle.

2.2) be the latter event of host exchange for intermediary switch:

The downstream switch of link one end of need not to lose efficacy is surveyed towards the link of this inefficacy link other end upstream switches again, but directly and host exchange set up primary link and get final product, therefore, the intermediary switch that is chosen as host exchange can directly be responded the SW_ACC message (if the E_Port of reception PIRD message is non-primary link to the E_Port that receives the PIRD message, then also need this non-primary link in E_Port place is recorded as new downstream primary link), thereby receive the SW_ACC message in downstream switch, after perhaps downstream switch and other intermediary switch that is not chosen as host exchange received the SW_ACC message, the E_Port of record transmission PIRD message can realize setting up between downstream switch and the host exchange primary link of new direct-connected or cascade; The intermediary switch that is chosen as host exchange is except responding the SW_ACC message, also transmit the PIRD message so that inefficacy primary link upstream switches can be received the PIRD message, receive the SW_ACC message of responding from other all primary link E_Port then to other all primary link E_Port.

3) switch lost efficacy as the downstream primary link upstream switches the time:

Receive the PIRD message that intermediary switch sends if having in the 2F_S_TOV of E_Port after the downstream primary link lost efficacy, then respond the SW_ACC message so that this SW_ACC message can be responded to the downstream switch of the primary link other end that lost efficacy or respond to this downstream switch and other intermediary switch to this E_Port; Further, be non-primary link if receive this E_Port place link of PIRD message in the 2F_S_TOV, then this E_Port place link is recorded as new downstream primary link;

If all do not receive the PIRD message in the 2F_S_TOV of all E_Port after the downstream primary link lost efficacy, then trigger BF and reshuffle.

In the practical application, the PIRD message can comprise the mutual Connection Service of expression switch optical fiber (Switch Fabric Internal Link Service, SW_ILS) heading of message, and the command word of expression message content, the command code that comprises sign PIRD message in the command word again, the switch name of inefficacy primary link middle and upper reaches switch, and the switch name of inefficacy primary link middle and lower reaches switch, switch name can be selected global unique identification (World Wide Name for use, can it not made an amendment when WWN), intermediary switch is transmitted the PIRD message.

Below, respectively at the different situations of intermediary switch and mutual in conjunction with between three kinds of switches, enumerate three embodiment main link recovery control method is elaborated again.

Main link recovery control method embodiment one

Fig. 3 is the exemplary flow schematic diagram among the main link recovery control method embodiment one of the present invention.As shown in Figure 3, be that the non-primary link E_Porr of non-host exchange, intermediary switch receives that the PIRD message is an example with intermediary switch, the main link recovery control method in the present embodiment comprises the steps:

All normal E_Port send the PIRD message to it for step 301, downstream switch;

The E_Porr that step 302, intermediary switch are positioned at non-primary link receives the PIRD message that downstream switch sends;

Step 303, there is the upstream primary link in intermediary switch, thereby the non-primary link in E_Port place from recorder PIRD message to its E_Port that receives the PIRD message that respond SW_ACC message and is new downstream primary link, also transmit the PIRD message so that the PIRD message can directly or indirectly arrive the upstream switches of inefficacy primary link by this upstream primary link, and receive the SW_ACC message of responding by the upstream primary link E_Port of this intermediary switch to the upstream of this intermediary switch primary link E_Port;

In the practical application, if the upstream switches of this intermediary switch promptly is the upstream switches of inefficacy primary link, the upstream switches of primary link of then losing efficacy receives the PIRD message that intermediary switch sends at E_Port in 2F_S_TOV, then respond the SW_ACC message, and can not trigger BF and reshuffle to this E_Port;

If the upstream switches of this intermediary switch is not the upstream switches of inefficacy primary link, then no matter the upstream switches of this intermediary switch is that host exchange also is non-host exchange, the upstream switches of inefficacy primary link all can continue to transmit the PIRD message to other primary link, so that can receive the PIRD message and know that the inefficacy primary link is substituted by other links according to the PIRD message of transmitting through this intermediary switch upstream primary link E_Port; Correspondingly, the upstream switches of inefficacy primary link receives the PIRD message that intermediary switch sends at E_Port in 2F_S_TOV, then responds the SW_ACC message to this E_Port, and can not trigger BF and reshuffle;

In this step,, then respond SW_RJT message and process ends to its E_Port that receives the PIRD message if there is not the upstream primary link in intermediary switch;

Step 304, downstream switch connect the E_Port of intermediary switch, receive in the F_S_TOV after the upstream primary link lost efficacy that switch replys the SW_ACC message, and this E_Port place link is recorded as new upstream primary link.

So far, this flow process finishes.

In the above-mentioned flow process, intermediary switch has received the PIRD message from non-primary link, make downstream switch in the inefficacy primary link lead to original upstream switches of inefficacy primary link and no longer direct-connected with this original upstream switches by intermediary switch, perhaps make in the inefficacy primary link downstream switch by intermediary switch lead to host exchange and no longer by the original upstream switches in the inefficacy primary link towards host exchange, thereby recovered and host exchange between complete primary link.

Fig. 4 a～Fig. 4 b is the example schematic among the main link recovery control method embodiment one of the present invention.

Referring to Fig. 4 a, there are switch 1～6, switch 1 to be host exchange, wherein,

Link ISL7 between switch 4 and the switch 6 is the downstream primary link of switch 4, the upstream primary link of switch 6, and promptly the primary link between switch 6 and the switch 1 is the cascade of ISL2, ISL7;

Link ISL6 between switch 5 and the switch 6 is non-primary link.

Suppose that ISL5 lost efficacy, then switch 4 is intermediary switch for the upstream switches of inefficacy primary link, switch 5 for the downstream switch of inefficacy primary link, all the other switches.

So, switch 5 can send the PIRD message to the E_Port of its all Up, and switch 6 can be received this PIRD message.

For switch 6, when its after receiving the PIRD message on the non-primary link ISL5 because it has upstream primary link ISL7, thereby upstream the E_Port of primary link ISL7 transmits the PIRD message and arrives switch 4;

Switch 4 is as the upstream switches of inefficacy primary link ISL5, in 2F_S_TOV, receive the PIRD message after, to write down non-primary link ISL7 be new downstream primary link and respond the SW_ACC message; Because switch 4 be upstream switches, thereby switch 4 can not transmit the PIRD message to switch 1 again, thereby switch 1 does not also just need the message to its primary link E_Port forwarding PIRD;

Switch 6 is new downstream primary link to switch 5 response SW_ACC messages and with its non-primary link ISL6 change record again;

Switch 5 is received the SW_ACC message, then non-primary link ISL6 change can be recorded as new upstream primary link.

Like this, originally by the direct-connected switch 5 of ISL5 and switch 4, change into by switch 6 and lead to switch 4, promptly, between switch 5 and the switch 1 former just like shown in the dashed double, by the complete primary link that ISL2 and ISL5 cascade form, just be replaced and revert to complete primary link shown in the solid line double-head arrow, that form by ISL2 and ISL7 and ISL6 cascade.

Referring to Fig. 4 b, there are switch 1～7, switch 1 to be host exchange, wherein,

Link ISL4 between switch 2 and the switch 5 is non-primary link;

So, switch 5 can send the PIRD message to the E_Port of its all Up, and switch 2 and switch 6 can be received this PIRD message.

For switch 2, when its after receiving the PIRD message on the non-primary link ISL4, because it has upstream primary link ISL1, thereby can to respond the SW_ACC messages and write down ISL4 to switch 5 be new downstream primary link, and switch 5 receives that in F_S_TOV it is new upstream primary link that the SW_ACC message then can write down ISL4; And switch 2 also is forwarded to host exchange again with the PIRD message, is switch 1;

Switch 1 can be responded the SW_ACC message to switch 2 after receiving the PIRD message; And, switch 1 also belongs to intermediary switch and receives the PIRD message from its downstream primary link ISL1, thereby switch 1 is also transmitted the PIRD message to other all primary link E_Port, so that in 2F_S_TOV, receive the PIRD message, do not reshuffle thereby can not trigger BF as the switch 4 of inefficacy primary link upstream switches.

Like this, between switch 5 and the switch 1 former just like shown in the dashed double, by the complete primary link that ISL2 and ISL5 cascade form, just be replaced and revert to complete primary link shown in the solid line double-head arrow, that form by ISL1 and ISL4 cascade.

In addition, for switch 6, it belongs to situation about describing among the follow-up main link recovery control method embodiment two, is specially, and switch 6 primary link is at its upstream received the PIRD message, thereby transmits and arrive switch 7 to the E_Port of other Up; And switch 7 also belongs to situation about describing among the follow-up main link recovery control method embodiment two, primary link is received the PIRD message but it does not have the E_Port of other Up again at its upstream, thereby to the E_Port transmission SW_RJT message of receiving the PIRD message, this SW_RJT message can arrive switch 5 through switch 6.At this moment, because switch 5 established new upstream primary link ISL4, thereby receive that the SW_RJT message can not initiate BF and reshuffle.

Main link recovery control method embodiment two

Fig. 5 is the exemplary flow schematic diagram among the main link recovery control method embodiment two of the present invention.As shown in Figure 5, be that the upstream primary link E_Port of non-host exchange, intermediary switch receives that the PIRD message is an example with intermediary switch, the main link recovery control method in the present embodiment comprises the steps:

All normal E_Port send the PIRD message to it for step 501, downstream switch;

The E_Port that step 502, intermediary switch are positioned at its upstream primary link receives the PIRD message that downstream switch sends;

Also there are the E_Port of other Up in step 503, intermediary switch except that this E_Port that receives the PIRD message, thereby transmit the PIRD message to other normal all E_Port;

In the practical application,, then respond SW_RJT message and process ends to its E_Port that receives the PIRD message if there is not the E_Port of other Up in intermediary switch except that this E_Port that receives the PIRD message;

Step 504, intermediary switch will send the SW_ACC message of the response that the E_Port of the Up of PIRD receives, and be forwarded to from downstream switch to receive the E_Port of PIRD message, also will be recorded as the downstream primary link from the upstream primary link change that downstream switch receives this E_Port place of PIRD message;

In this step, the SW_ACC message of receiving is the SW_ACC message responded of the intermediary switch in the flow process as shown in Figure 3 normally; After intermediary switch in the flow process is responded the SW_ACC message so as shown in Figure 3, also might need other all receive the intermediary switch of PIRD message all primary link E_Port forwarding PIRD messages by self from the downstream primary link, have found new primary link so that the upstream switches of inefficacy primary link is known the downstream switch of this inefficacy primary link; Correspondingly, the upstream switches of inefficacy primary link has E_Port to receive the PIRD message that intermediary switch sends in 2F_S_TOV, then responds the SW_ACC message to this E_Port, and can not trigger BF and reshuffle;

Step 505, downstream switch connects the E_Port of intermediary switch, receives the SW_ACC message in the F_S_TOV after the upstream primary link lost efficacy, and this E_Port place link is recorded as new upstream primary link.

So far, this flow process finishes.

In the above-mentioned flow process, intermediary switch has received the PIRD message from the upstream primary link, make downstream switch in the inefficacy primary link by by after changing downstream primary link of the original upstream of this intermediary switch primary link, this intermediary switch and by the non-primary link upstream primary link after changing between this intermediary switch and the inefficacy link middle and upper reaches switch, again be communicated with inefficacy link middle and upper reaches switch, thus recovered and host exchange between complete primary link.

Fig. 6 a～6b is the example schematic among the main link recovery control method embodiment two of the present invention.

Referring to Fig. 6 a, there are switch 1～6, switch 1 to be host exchange, wherein,

Link ISL5 between switch 5 and the switch 6 is the downstream primary link of switch 5, the upstream primary link of switch 6, and promptly the primary link between switch 6 and the switch 1 is the cascade of ISL2, ISL5, ISL6;

Link ISL7 between switch 4 and the switch 6 is non-primary link.

For switch 6, when its after receiving the PIRD message on the primary link ISL5 of upstream, can transmit and arrive switch 4 to the E_Port of other Up;

Switch 6 is new downstream primary link to switch 5 response SW_ACC messages and with its original upstream primary link ISL6 change record again;

Switch 5 is received the SW_ACC message, then the ISL6 change for the downstream primary link originally can be recorded as new upstream primary link.

Like this, between switch 5 and the switch 1 former just like shown in the dashed double, by the complete primary link that ISL2 and ISL5 cascade form, just be replaced and revert to complete primary link shown in the solid line double-head arrow, that form by ISL2 and ISL7 and ISL6 cascade.

Referring to Fig. 6 b, there are switch 1～7, switch 1 to be host exchange, wherein,

Link ISL4 between switch 2 and the switch 5 is non-primary link;

Suppose that ISL2 lost efficacy, then switch 1 is intermediary switch for the upstream switches of inefficacy primary link, switch 4 for the downstream switch of inefficacy primary link, all the other switches.

So, switch 4 can send the PIRD message to the E_Port of its all Up, and switch 5 can be received this PIRD message.

For switch 5, when its after receiving the PIRD message on the primary link ISL5 of upstream, can transmit and arrive switch 2 and switch 6 to the E_Port of other Up;

For switch 2, it belongs to the situation among the aforementioned backbone road recovery control method embodiment one, as previously mentioned, switch 2 is owing to have upstream primary link ISL1, thereby can to respond the SW_ACC messages and write down ISL4 to switch 5 be new downstream primary link, and switch 2 also is forwarded to host exchange again with the PIRD message, is switch 1;

Switch 1 will be responded the SW_ACC message as host exchange to switch 2 after receiving the PIRD message;

Switch 5 receives that it is new upstream primary link that the SW_ACC message then can write down ISL4, responds SW_ACC, also the ISL5 change for the upstream primary link originally is recorded as the downstream primary link to switch 4 then.

Switch 4 is received SW_ACC in F_S_TOV after, can write down ISL5 is new upstream primary link;

And, because switch 1 also belongs to intermediary switch and receives the PIRD message from its downstream primary link ISL1, thereby switch 1 is also transmitted the PIRD message to other all primary link E_Port, so that in 2F_S_TOV, receive the PIRD message, do not reshuffle thereby can not trigger BF as the switch 4 of inefficacy primary link upstream switches.

Like this, between switch 4 and the switch 1 former just like shown in the dashed double, by the direct-connected complete primary link that forms of ISL2, just be replaced and revert to complete primary link shown in the solid line double-head arrow, that form by ISL1 and ISL4 and ISL5 cascade.

In addition, for switch 6, its upstream primary link is received the PIRD message, thereby transmits and arrive switch 7 to the E_Port of other Up; And switch 7 at its upstream primary link receive the PIRD message but it does not have the E_Port of other Up, thereby send the SW_RJT message to the E_Port that receives the PIRD message, this SW_RJT message can arrive switch 4 through switch 6, switch 5.At this moment, because switch 5 established new upstream primary link ISL5, thereby receive that the SW_RJT message can not initiate BF and reshuffle.

Main link recovery control method embodiment three

Fig. 7 is the exemplary flow schematic diagram among the main link recovery control method embodiment three of the present invention.As shown in Figure 7, be that host exchange is an example with the intermediary switch, the main link recovery control method in the present embodiment comprises the steps:

All normal E_Port send the PIRD message to it for step 701, downstream switch;

Step 702, an E_Port of intermediary switch receives the PIRD message that downstream switch sends;

Step 703 is responded the SW_ACC message, also this E_Port place link is recorded as new downstream primary link to its E_Port that receives the PIRD message as the intermediary switch of host exchange;

After this step, because host exchange also is an intermediary switch, thereby comprise host exchange other all receive the PIRD message from the downstream primary link all intermediary switch all need to transmit PIRD messages by all the primary link E_Port of self found new primary link so that the upstream switches of inefficacy primary link is known the downstream switch of this inefficacy primary link;

Correspondingly, the upstream switches of inefficacy primary link has E_Port to receive the PIRD message that intermediary switch sends in 2F_S_TOV, then responds the SW_ACC message to this E_Port, and can not trigger BF and reshuffle;

Step 704, downstream switch connect the E_Port of intermediary switch, receive in the F_S_TOV after the upstream primary link lost efficacy that switch replys the SW_ACC message, and this E_Port place link is recorded as new upstream primary link.

So far, this flow process finishes.

In the above-mentioned flow process, because intermediary switch is host exchange, thereby the downstream switch in the feasible inefficacy primary link skips the original upstream switches in the inefficacy primary link and leads to host exchange, thus recovered and host exchange between complete primary link.Certainly, above-mentioned flow process mainly is to be the explanation that the situation of host exchange is carried out around intermediary switch, as being direct-connected between the intermediary switch of host exchange and the downstream switch, also might be any one or more combined situation among the aforementioned backbone road recovery control method embodiment one, two.

Fig. 8 is the example schematic among the main link recovery control method embodiment three of the present invention.Referring to Fig. 8, there are switch 1～6, switch 1 to be host exchange, wherein,

Link ISL4 between switch 1 and the switch 5 is non-primary link;

So, switch 5 can send the PIRD message to the E_Port of its all Up, and switch 1, switch 6 can be received this PIRD message.

For switch 1, it is as intermediary switch, and can directly respond the SW_ACC message and write down ISL4 to switch 5 be new downstream primary link;

Switch 5 is received SW_ACC in F_S_TOV after, can write down ISL4 is new upstream primary link;

And, because switch 1 also belongs to intermediary switch and receives the PIRD message from its downstream primary link ISL4, thereby switch 1 is also transmitted the PIRD message to other all primary link E_Port, so that in 2F_S_TOV, receive the PIRD message, do not reshuffle thereby can not trigger BF as the switch 4 of inefficacy primary link upstream switches.

Like this, between switch 5 and the switch 4 former just like shown in the dashed double, by the complete primary link that ISL2 and ISL5 cascade form, just be replaced revert to shown in the solid line double-head arrow, by the direct-connected complete primary link that forms of ISL4.

In addition, for switch 6, situation about describing among its aforementioned backbone road recovery control method embodiment two, be specially, switch 6 primary link is at its upstream received the PIRD message but it does not have the E_Port of other Up again, thereby to the E_Port transmission SW_RJT message of receiving the PIRD message, this SW_RJT message can arrive switch 5.At this moment, because switch 5 established new upstream primary link ISL4, thereby receive that the SW_RJT message can not initiate BF and reshuffle.

Above-mentioned four main link recovery control method embodiment are explanations of carrying out at the different situations of intermediary switch respectively.

In the practical application, because the various situations of intermediary switch all may occur, thereby preferable complete scheme be among above-mentioned three main link recovery control method embodiment any two or all three exist simultaneously, be between the intermediary switch of every kind of situation and the downstream switch and/or between intermediary switch and the upstream switches, also can pass through intermediary switch cascade, thereby some steps of each flow process might replace in any way and/or execution simultaneously among the different embodiment in other situations.In this case, may there be the new link that has many can lead to host exchange simultaneously, so at this moment, downstream switch can be upgraded the once upstream primary link of record when whenever receiving a SW_ACC message, also can only upgrade once the upstream primary link of record when receiving the SW_ACC message for the first time and ignore the follow-up SW_ACC message of receiving.

More than, be detailed description to main link recovery control method in the embodiment of the invention.Below, again to realizing in the embodiment of the invention that switch system that primary link recovers and all kinds of switches in this switch system are elaborated.

Switch system in the present embodiment comprises:

Downstream switch, to its all Up, promptly the E_Port by link negotiation sends the PIRD message, because the E_Port of inefficacy upstream primary link is Up no longer, thereby, send the PIRD message from the E_Port of other all Up, can detect other links direct-connected or cascade that except the inefficacy primary link, whether exist towards host exchange;

Specifically, upstream switches has some E_Port and comprises:

Packet sending unit sends the PIRD message to all normal E_Port;

Recover timing unit, after the upstream primary link loses efficacy, picked up counting, and timing finishes when arriving F_S_TOV;

Recover decision unit, if there is E_Port to receive that switch replys the SW_ACC message before finishing recovering the timing unit timing, then this E_Port place link is recorded as new upstream primary link, further, if have more than an E_Port and recovering the timing unit timing when receiving the SW_ACC message before finishing, then preferably, recover the E_Port place link that decision unit further will receive the SW_ACC message at first and be recorded as new upstream primary link; Do not receive the SW_RJT message that upstream switches or intermediary switch are replied if all receive SW_ACC message or all E_Port that upstream switches or intermediary switch are replied at the PIRD message in the F_S_TOV of all E_Port after the upstream primary link lost efficacy, then trigger BF and reshuffle.

Intermediary switch, when the E_Port place link that receives the PIRD message is non-primary link, if there is the upstream primary link in this intermediary switch, represent that then this intermediary switch can be used as the new upstream switches of inefficacy primary link middle and lower reaches switch, thereby to the E_Port response SW_ACC message that receives the PIRD message, and the non-primary link in E_Port place of recorder PIRD message is new downstream primary link, and, also transmit the PIRD message to the upstream of this intermediary switch primary link, and the SW_ACC message responded of the switch that receives this intermediary switch upstream primary link other end, thereby receive the SW_ACC message in downstream switch, perhaps downstream switch and other intermediary switch that is not chosen as host exchange receive the E_Port that record behind the SW_ACC message sends the PIRD message, can realize that downstream switch passes through to set up between this intermediary switch and the host exchange primary link of new cascade; If there is not the upstream primary link in this intermediary switch, represents that then this intermediary switch not can be used as the new upstream switches of inefficacy primary link middle and lower reaches switch, thereby respond the SW_RJT message to the E_Port that receives the PIRD message.

Intermediary switch, at the E_Port place link that receives the PIRD message during for the upstream primary link of this intermediary switch:

If except that this E_Port that receives the PIRD message, also there is the E_Port of other Up, then transmit the PIRD message to other normal all E_Port, to seek other link of alternative inefficacy primary link, and the SW_ACC message of the response that the E_Port of other Up is received is forwarded to the E_Port that receives the PIRD message, also the E_Port place link with the SW_ACC that receives is recorded as the upstream primary link, the upstream primary link change that receives the E_Port place of PIRD message is recorded as the downstream primary link, thereby receive the SW_ACC message in downstream switch, after perhaps downstream switch and other intermediary switch that is not chosen as host exchange receive the SW_ACC message, record sends the E_Port of PIRD message, can realize that downstream switch passes through to set up between this intermediary switch and the host exchange primary link of new cascade; Preferably, after other normal all E_Port transmit the PIRD message, if there is E_Port to receive the SW_ACC message more than other Up, then the SW_ACC message that receives at first can be forwarded to the E_Port that receives the PIRD message, and the E_Port place link that will receive SW_ACC at first is recorded as the upstream primary link, the upstream primary link that will receive the E_Port place of PIRD message is recorded as the downstream primary link; In addition, after other normal all E_Port transmit the PIRD message, if the E_Port of all other Up all receives the SW_RJT message of response, represent that then this intermediary switch not can be used as the new upstream switches of inefficacy primary link middle and lower reaches switch, transmit the SW_RJT message to the E_Port that receives the PIRD message;

Intermediary switch, at the E_Port place link that receives the PIRD message (PIRD message that intermediary switch is received from the downstream primary link during for the downstream primary link of this intermediary switch, normally be received in non-primary link, the upstream primary link is received the PIRD message that the intermediary switch of PIRD message is transmitted to the host exchange direction), can't know that for fear of inefficacy primary link upstream switches downstream switch has found new primary link, and the situation that false triggering BF reshuffles occurs, the PIRD message of receiving is transmitted to all primary link E_Port (comprising its upstream primary link E_Port and other downstreams primary link E_Port), so that inefficacy primary link upstream switches can be received the PIRD message, thereby the primary link upstream switches of avoiding losing efficacy has found in downstream switch and triggers BF under the situation of new primary link and reshuffle.

Specifically, intermediary switch has some E_Port and comprises:

Any E_Port place Linktype that receives the PIRD message is judged in the link identification unit;

Transmit decision unit, if receive the E_Port place link of PIRD message is that non-primary link and this intermediary switch exist the upstream primary link, then responding the SW_ACC message, write down this non-primary link in E_Port place to the E_Port that receives the PIRD message is the downstream primary link, and transmits the PIRD message, receives the SW_ACC message of responding then to the upstream of this intermediary switch primary link E_Port; If receive the upstream primary link of the E_Port place link of PIRD message for this intermediary switch, and except that this E_Port, also there are other normal E_Port, then send the PIRD message to other normal all E_Port, to seek other link of alternative inefficacy primary link, and the SW_ACC message that other normal E_Port receive is forwarded to the E_Port that receives the PIRD message, the E_Port place link that also will receive SW_ACC is recorded as the upstream primary link, the upstream primary link that receives the E_Port place of PIRD message is recorded as the downstream primary link, further, if have more than other normal E_Port and receive the SW_ACC message, then preferably, transmit the SW_ACC message that decision unit further will receive at first and be forwarded to the E_Port that receives the PIRD message, and the E_Port place link that further will receive SW_ACC at first is recorded as the upstream primary link, the upstream primary link that receives the E_Port place of PIRD message is recorded as the downstream primary link; If receive the downstream primary link of the E_Port place link of PIRD message for this intermediary switch, then transmit the PIRD message so that inefficacy primary link upstream switches can be received the PIRD message, receive the SW_ACC message of responding from host exchange then to other all primary link E_Port.

In addition, a kind of special situation is, intermediary switch is a host exchange, so at this moment, the downstream switch of link one end of need not to lose efficacy is surveyed towards the link of this inefficacy link other end upstream switches again, but directly and host exchange set up primary link and get final product, therefore, the intermediary switch that is chosen as host exchange can directly be responded the SW_ACC message (if the E_Port of reception PIRD message is non-primary link to the E_Port that receives the PIRD message, then also need this non-primary link in E_Port place is recorded as new downstream primary link), thereby receive the SW_ACC message in downstream switch, after perhaps downstream switch and other intermediary switch that is not chosen as host exchange received the SW_ACC message, the E_Port of record transmission PIRD message can realize setting up between downstream switch and the host exchange primary link of new direct-connected or cascade; As the intermediary switch of host exchange except responding the SW_ACC message, also transmit the PIRD message so that inefficacy primary link upstream switches can be received the PIRD message, receive the SW_ACC message of responding from other all primary link E_Port then to other all primary link E_Port.

Upstream switches, receive the PIRD message that intermediary switch sends if having in the 2F_S_TOV of E_Port after the downstream primary link lost efficacy, then respond the SW_ACC message so that this SW_ACC message can be responded to the downstream switch of the primary link other end that lost efficacy or respond to this downstream switch and other intermediary switch to this E_Port; Further, be non-primary link if receive this E_Port place link of PIRD message in the 2F_S_TOV, then this E_Port place link is recorded as new downstream primary link; If all do not receive the PIRD message in the 2F_S_TOV of all E_Port after the downstream primary link lost efficacy, then trigger BF and reshuffle.

Specifically, downstream switch has some E_Port and comprises:

Recover timing unit, when the downstream primary link loses efficacy, picked up counting, and when arriving 2F_S_TOV, finish timing;

Recover decision unit, if there is E_Port recovering to receive the PIRD message before the timing unit timing end, then respond the SW_ACC message to this E_Port, and if this E_Port place link be non-primary link, then this E_Port place link is recorded as new downstream primary link; If all do not receive the PIRD message in the 2F_S_TOV of all E_Port after the downstream primary link lost efficacy, then trigger BF and reshuffle.

In the switch system of present embodiment and the various switch, the PIRD message can comprise the command word of the heading and the expression message content of expression SW_ILS message, the command code, the switch name of inefficacy primary link middle and upper reaches switch and the switch name of inefficacy primary link middle and lower reaches switch that comprise sign PIRD message in the command word again, switch name can be selected WWN for use, and intermediary switch can not make an amendment to it when transmitting the PIRD message.

The above is preferred embodiment of the present invention only, is not to be used to limit protection scope of the present invention.Within the spirit and principles in the present invention all, any modification of being done, be equal to and replace and improvement etc., all should be included within protection scope of the present invention.

Claims

1. a main link recovery control method is characterized in that, this control method comprises the steps:

2. control method as claimed in claim 1 is characterized in that,

3. control method as claimed in claim 1 or 2, it is characterized in that, include the switch name of the command code that identifies this message, described inefficacy primary link middle and upper reaches switch and the switch name of described inefficacy primary link middle and lower reaches switch in the PIRD message.

4. control method as claimed in claim 3 is characterized in that, described switch name is global unique identification WWN.

5. switch, the upstream primary link with some switches interconnected pattern fibre channel port E_Port and this switch lost efficacy, and it is characterized in that this switch comprises:

6. switch as claimed in claim 5, it is characterized in that, have more than an E_Port and recovering the timing unit timing when receiving the SW_ACC message before finishing, recover the E_Port place link that decision unit further will receive the SW_ACC message at first and be recorded as new upstream primary link.

7. as claim 5 or 6 described switches, it is characterized in that, include the command code that identifies this message, the switch name of inefficacy upstream primary link other end upstream switches and the switch name of described switch in the PIRD message.

8. switch as claimed in claim 7 is characterized in that, described switch name is global unique identification WWN.

9. switch has some switches interconnected pattern fibre channel port E_Port and this switch and is not chosen as host exchange and is arranged in any other links outside the inefficacy primary link, it is characterized in that this switch comprises:

10. switch as claimed in claim 9 is characterized in that,

11. as claim 9 or 10 described switches, it is characterized in that, include the switch name of the command code that identifies this message, described inefficacy primary link middle and upper reaches switch and the switch name of described inefficacy primary link middle and lower reaches switch in the PIRD message.

12. switch as claimed in claim 11 is characterized in that, described switch name is global unique identification WWN.

13. a switch, the downstream primary link with some switches interconnected pattern fibre channel port E_Port and this switch lost efficacy, and it is characterized in that this switch comprises:

14. switch as claimed in claim 13 is characterized in that, includes the switch name of the command code that identifies this message, described switch and the switch name of inefficacy downstream primary link other end downstream switch in the PIRD message.

15. switch as claimed in claim 14 is characterized in that, described switch name is global unique identification WWN.