CN1340928A

CN1340928A - Stand-by method and device of communication system

Info

Publication number: CN1340928A
Application number: CN 00125043
Authority: CN
Inventors: 丁贵祥; 杜元光
Original assignee: ZTE Corp
Current assignee: Haimen Science And Technology Development General Corp
Priority date: 2000-09-02
Filing date: 2000-09-02
Publication date: 2002-03-20
Anticipated expiration: 2020-09-02
Also published as: CN1198406C

Abstract

A high-reliability communication system backup method and equipment, which can detect link or system faults in time through software, and can locate the source of the fault and make a judgment. The fault judgment mechanism is through the monitoring of the server and the client. The program judges the communication failure of the system to achieve. It can realize the full backup communication link from the client to the server, ensuring the complete backup of the system, and can quickly restore the normal state when the communication system fails. According to the design, after the system failure occurs, the system can generally be restored in a short time Operation, so as to meet the high reliability requirements of the current telecommunications industry for communication equipment.

Description

Communication system backup method and equipment

The present invention relates to computer communication technology, specifically, relate to a kind of communication system backup method and the equipment of realizing telecommunications industry high reliability local area network (LAN) communication link redundancy backup.

From reliability, the backup of the communication link of network is to improve one of method of network reliability, reduction network failure rate.The thought of backup communication is circuit and the equipment that redundancy is provided at the link of network communication, to guarantee normally carrying out of network communication.Except the communication link backup, also have path redundancy, equipment redundancy etc.The design of these aspects has had many methods to realize.As the paper " realization of Tolerate and redundance network adapter " (1999 on " computer engineering and application " magazine, 11, the 105-107 page or leaf) a kind of implementation method of redundant network adapter has been proposed, " a kind of design of high available fault-tolerant server " (1998 of " marine electronic engineering ", 1,30-37) and following patent documentation some similar solutions are also proposed.

The open CN1136250A " method for fault-tolerant communication under strictly real-time conditions " of Chinese invention patent proposes a kind of bus redundancy backup mode, and it mainly is to realize link backup to local area network (LAN) by the backup to communication link.But this mode realizes the control of main/slave link by the CPU communication controller of special use, has increased the overhead of system, and this system can not provide the full backup from client to the server communication simultaneously.The open CN1192094A of Chinese invention patent " expandable telecommunications system redundancy, exchange capacity expansion and fault isolation arrangements " proposes a kind of backup method to telecommunication switching system, but do not relate to backup, do not propose solution yet for the backup that requires to realize unattended client to its maintenance system.The patent No. is that the United States Patent (USP) that 05227330A, name are called " backup of communication line " (BACKUP SYSTEM FORCOMMUNICATION LINE) has proposed a kind of backup mode at connection, also not have the solution of proposition at whole system.

The method of above prior art all is to a kind of the proposing a solution in the redundancy of path redundancy or server apparatus redundancy or communication apparatus itself.In addition, in the reality, client also needs to possess high reliability sometimes, specifically, when communication failure takes place, wishes to detect rapidly fault and automatic switchover, thereby realizes unattended operation.So just require whole system all to possess high reliability.Still do not have at present to find to have document that this realization is proposed a feasible reliable solution.

The objective of the invention is to propose a kind of communication system backup method and equipment, utilize this method and apparatus, can overcome the redundancy backup dead angle that prior art exists, communication link standby system at the high reliability real-time system, propose a solution, can realize that the software detecting switches the real-time full backup of server, client and communication link, but equal quick recovery system function when any breaks down among the three, thereby make the reliability of backup reach further raising.

The object of the present invention is achieved like this, construct a kind of communication system backup method of high reliability, may further comprise the steps: whether the network measuring module detection by host server once connected with client computer, then be provided with in this way the first overtime T1 less (=2.5*T), otherwise be provided with the first overtime T1 big (=20*T); If through the current first overtime T1, host server can't connect with client, then carry out server and switch for active and standby part; The second overtime T2 (=2.5*T) in the time, detect by the active and standby detection module in the backup server and whether to set up " heartbeat " and is connected, if fail to connect then active and standby part of switching of server with main frame; Detect by the client network detection module and whether once to connect with server, as do not connect then be provided with the overtime the 3rd overtime T3 big (=20*T), otherwise the less (=2*T＜T1) of the 3rd overtime T3 is set; If can not connect with server, then carry out client computer and switch for active and standby part through the current the 3rd overtime T3; The less the 4th overtime T4 (=2*T) in the time, detect by active and standby detection module in the backup client and whether to set up " heartbeat " and is connected, if fail to set up then carry out active and standby part of switching of client computer with main frame; No matter server or client initiatively require active and standby switching as long as detect main frame, and the sign that then will connect is changed to " not connecting " mark.

In backup method provided by the invention, the first overtime T1 smaller value equals the described second overtime T2, the described the 3rd overtime T3 smaller value is less than the described first overtime T1 smaller value, the described the 3rd overtime T3 higher value equals the described first overtime T1 higher value, the described the 3rd overtime T3 smaller value is less than the described first overtime smaller value, and the described the 3rd overtime T3 smaller value is much smaller than finishing active and standby switching required time.Simultaneously, the described first overtime T1 higher value is much larger than the described first overtime T1 smaller value, and the described the 3rd overtime T3 higher value is much larger than the described the 3rd overtime T3 smaller value.

In backup method provided by the invention, the active and standby switching of described server may further comprise the steps: if the link between server and the client is for the circuit of a client then select line with another client for use, if the link between server and the client is master-master or be equipped with-be equipped with end line then select for use main-be equipped with or be equipped with-main end line, if the link between server and the client be main-be equipped with or be equipped with-main line then select master-master for use or fully-reserve line 2.

Another object of the present invention is to realize like this, construct a kind of communication system alternate device, comprise client computer and backup client machine, receive on hub and the backup hub by network adapter respectively, also comprise server and backup server, two network adapter are arranged respectively on described server and the backup server, be connected on hub and the backup hub by each line respectively, between the described active and standby part of client computer bidirectional communication link is arranged, described server and described backup server 108 are connected respectively to a shared data bank.

An aspect according to equipment provided by the invention, described main with comprising the server host administration module in the server, server activestandby state detection module, the first network measuring module and the second network measuring module, in described standby server, comprise server backup machine administration module, server activestandby state detection module, the first network measuring module and the second network measuring module, wherein, the described first network measuring module, be used to detect the main server of using, standby server is connected with the primary client, the described second network measuring module, be respectively applied for and detect the main server of using, standby server is connected with the primary client, and described Host Administration module is carried out the link joint detection by described server activestandby state detection module the indication first or second network measuring module; Described Host Administration module is carried out the link joint detection by described server activestandby state detection module the indication first or second network measuring module.Wherein, client-server administration module, client activestandby state detection module and network measuring module are arranged in client computer, client-server administration module, client activestandby state detection module and network measuring module are arranged in the backup client machine.

According to another aspect of equipment provided by the invention, whether described Host Administration module detects by the network measuring module and connects with client computer, then be provided with in this way the first overtime T1 less (=2.5*T), otherwise be provided with the first overtime T1 big (=20*T); If through the current first overtime T1, host server can't connect with client, then carry out server and switch for active and standby part.Wherein, described Host Administration module the second overtime T2 (=2.5*T) in the time, detect by active and standby detection module and whether to set up " heartbeat " and is connected, if fail to connect then active and standby part of switching of server with main frame; Wherein, client-server administration module, client activestandby state detection module and network measuring module are arranged respectively in client computer and client-server, client-server administration module, client activestandby state detection module and network measuring module are arranged in the backup client machine.Wherein, whether described client manager module detects by the network measuring module and connects with server, as do not connect then be provided with the overtime the 3rd overtime T3 big (=20*T), otherwise the less (=2*T＜T1) of the 3rd overtime T3 is set; If can not connect with server, then carry out client computer and switch for active and standby part through the current the 3rd overtime T3; Wherein, described client manager module by the network measuring module the less the 4th overtime T4 (=2*T) in the time, detect by active and standby detection module and whether to set up " heartbeat " and is connected, if fail to set up then carry out active and standby part of switching of client computer with main frame.

Implement high reliability communication system backup method provided by the invention and equipment, at first, can in time find the link or the system failure by software mode, and but fault location is initiated the source, and make judgement, its fault judgement mechanism is to judge that by the monitoring facilities that resides in server and client the system communication fault realizes.Secondly, method and apparatus of the present invention can be realized the full backup communication link from the client to the server.Guaranteed the backup completely of system.At last, utilize method and apparatus of the present invention, when communication system breaks down, can recover normal condition rapidly, according to designing after the system failure takes place, system generally can resume operation with interior at 1 minute, thereby satisfies the requirement of present telecommunications industry to the high reliability of communication equipment.

In conjunction with the accompanying drawings and embodiments, further specify characteristics of the present invention.In the accompanying drawing:

Fig. 1 is that the hardware of alternate device provided by the invention is realized connection layout.

Fig. 2 is based on the software realization module map that Fig. 1 hardware is realized the inventive method.

Fig. 3 is two NIC driver design frameworks on the server in the system of the present invention.

Fig. 4 is the disconnected handling process in first networking.

Fig. 5 is the server failing process chart.

Fig. 6 is a client collapse process chart.

Fig. 7 is the process chart that two grid lines of server all break down.

In the hardware implementation structure schematic diagram of the realization the inventive method shown in Fig. 1, be provided with client computer 101 and backup client machine 102, they are received on hub 105 and the backup hub 106 by network adapter 103,104 respectively.Equally, two network adapter 109,110 and 111,112 are arranged respectively on server 107 and the backup server 108, also receive on hub 105 and the backup hub 106 by network respectively.These two hubs constitute two networks respectively and backup each other.Herein, two networks are irrelevant mutually, separate.And the network segment of the IP of corresponding network card adapter is identical.But server 107 and backup server 108 are shared a database 113.Dotted line 114,115,116,117,118,119 expression connection message paths among the figure, wherein: the message path of dotted line 114 expression master servers and main client computer; The first handshake information path of dotted line 115 expression master servers and backup server; The message path of dotted line 116 expression master servers and backup client machine; The message path of dotted line 117 expression backup servers and main client computer; The second handshake information path of dotted line 118 expression master servers and backup server; The message path of dotted line 119 expression backup servers and backup client machine, the communication dual link between active and standby part of client computer of bidirectional signal line 120 expressions.

Realize in the block diagram at the software of the inventive method shown in Fig. 2, in leading, comprise server host administration module 201 with server 107, server activestandby state detection module 202, the first network measuring module 203 and the second network measuring module 204, in standby server 108, comprise server backup machine administration module 205, server activestandby state detection module 206, the first network measuring module 207 and the second network measuring module 208, client-server administration module 209 is arranged in client computer 101, client activestandby state detection module 210 and network measuring module 211 have client-server administration module 212 in client computer 102, client activestandby state detection module 213 and network measuring module 214.Also illustrate among Fig. 2, the master is wiredly connected to main with hub 105, backup hub 106 respectively with server 107, standby server 108, and the master is connected to the main client computer 101 of using with hub 105, and backup hub 106 is connected to Alternative Client 102.

In conjunction with the explanation of above equipment tissue, further specify the backup method of communication system of the present invention.Herein, server host 107 softwares are received and dispatched test frame mutually by network TCP/IP and client 101 realization programs, attention is on TCP/IP connect to realize, the mixed-media network modules mixed-media of server 107 is initiatively to carry out the network connection, and the mixed-media network modules mixed-media of client computer 101 is that connect port etc. is to be connected.Originally, the standby host 108 of the standby host of client 102 and server all is idlely promptly not provide application service, but detects the running status of main frames at any time by activestandby state detection module 213.The active and standby machine 101 of client is connected by general communication link 120 with 102, as by two serial ports, sets up two serial ports links and connects.Server end is set up two communication links by line 114, line 116, master with hub 105, line 115, line 116, backup hub 106 and is connected (also can adopt two serial ports to connect).

Server 107 ends detect and being connected of client by network measuring module 203 and 204.Two detection modules are all arranged, as 203,204 and 207,208 on each server.

Network measuring module

207 and 208 on the guest machine is not worked.The active and standby machine of client has a

network measuring module

211 and 214 respectively, is used to detect and being connected of server, and the detection module 214 on the guest machine is not worked.

201 pairs of two network management modules 203 of server host administration module and 204 activestandby state detection modules 202 carry out integrated management.Determine whether carrying out active and standby part of switching, alarm etc. according to following criterion, the time wherein is provided with, and parameter can be changed according to actual needs.Following column of figure just illustrates.(related T, T1, T2, T3, all legends below T4 is suitable for), below, T=20s basic time is set, backup machine switches deadline＜0.2*T from beginning to switch to, and then carries out active and standby part according to following criterion and switches:

If the network measuring module of host server as yet not with the client setting that connects, so overtimely be set to T1=20*T, in case network measuring module and client connect, the overtime T1=2.5*T that is is set so.If can not connect with client through T1, then switch.This moment, main frame became standby host, and standby host becomes main frame.

The standby host discovering server surpasses the T2=2.5*T time, and active and standby detection module can't be set up " heartbeat " and be connected with main frame, then automatically switch into main frame.

The client network detection module does not connect with server as yet, and the overtime T3=20*T of being is set, otherwise T3=2*T＜T1.If can not connect with server through T1, then switch.This moment, main frame became standby host, and standby host becomes main frame.

The standby host client terminal to discover is through the overtime T4=2*T time, and active and standby detection module can't be set up " heartbeat " and be connected with main frame, then automatically switch into main frame.

No matter server or client, main frame initiatively requires to switch, and the sign that then will once connect is changed to the mark that had not connected.

Be provided with do not connect as yet time-out time quite the reason of big (20T) be to prevent active and standby machine server, client is switched back and forth repeatedly.

In the structure of two network interface card driver modules, two NIC drivers adopt two threads that two network interface cards are detected respectively on the server shown in Fig. 3, and network interface path condition detection module manages two network interface cards.

Fig. 4 is when operate as normal, and article one network path 114 is disconnected, and among the figure, 4 " X " of appearance represents the fault point.All are normal for other circuit and backup machine.

Client is switched after the time through T3.Be 0.2*T switching time, finishes 102 switchings of client standby host from link is disconnected to switching, and is total to 2.2*T＜T1, and server host 107 can not switch.After client 102 is finished switching, by network measuring module 204, network measuring module 214 and the backup hub 106 of Fig. 2, make server 107 rebulid and be connected, adapter work, alarm simultaneously with client.

At Fig. 5 host server being shown occurs under the deadlock situation of (representing to crash with " X " in the host server square frame among Fig. 5), the main frame 101 of client will once with server linkage flag be changed to as yet not linkage flag earlier at T3 after the time, automatically switch to alternate client then, alternate client 102 becomes the wait state that enters the 20*T time behind the main frame.The server activestandby state detection module 206 (Fig. 2) of backup server 108 (Fig. 1) was found and can't be connected with main frame in the T2 time, then automatically switch and become server host, network measuring module 208, network measuring module 214 and backup hub 106 that this moment, server guest machine 108 and client guest machine 102 passed through among Fig. 2,

line

219 and 221 rebulids and is connected.

Occur crashing at the client host shown in Fig. 6 101 under (among the figure, crashing with " X " in square frame expression) situation, client guest machine 102 switches after the time through T4.Be 0.2*T switching time, finishes the switching of client standby host from link is disconnected to switching, and is total to 2.2*T＜T1, and server host 107 can not switch.After client is finished switching, the network measuring module 204 by Fig. 2, network measuring module 214, and backup hub 106,

line

217 and 221, server 107 rebulids with client and is connected, and adapter work is alarmed simultaneously.

Fig. 7 illustrates two processing procedures that grid line all breaks down of server host 107.As shown in Figure 7, the main frame 101 of client is as after finding to connect with server 107 through the T3 time, earlier will once be changed to " not being connected as yet " mark with server 107 linkage flags, automatically switch to alternate client 102 then, alternate client 102 becomes the wait state that enters the 20*T time behind the main frame.If server is the mutual detection of setting up state by two networks, the software 206 of the activestandby state of server guest machine 108 detection was so subsequently found and can't be connected with main frame in the T2 time, if (server is to carry out state-detection by serial ports or other communication links, finding after time to be connected through T1 initiatively by the switching of serial ports notice standby host server requirement with client) then automaticallying switch becomes main frame, this moment, server guest machine 108 and client guest machine 102 were by the network measuring module 208 among Fig. 2, network measuring module 214 and backup hub 106,

line

219 and 221 rebulids and is connected.

From as seen last, no matter be a network path break down (Fig. 4), still host server crash (Fig. 5), or crashing (Fig. 6) appears in client computer or two network paths break down (Fig. 7), utilize backup method provided by the invention, can realize quickly and easily that all active and standby part switches.

Claims

1. A communication system backup method, characterized in that, comprising the following steps:

Whether the network detection module (203) of host server (107) detects once with client computer and establishes connection, if so then setting the first overtime T1 is smaller (=2.5*T), otherwise the first overtime T1 is set larger (=20*T) T); if the host server cannot establish a connection with the client through the current first overtime T1, then perform server master-backup switching;

In the second overtime T2 (=2.5*T) time, detect whether to set up " heartbeat " connection with main frame by the master-standby detection module (206) in the backup server (108), if fail to establish connection then server master-backup switch;

Whether the client (101) network detection module (211) detects whether to ever set up connection with the server, if the connection is not established, the third overtime T3 is set to be larger (=20*T), otherwise the third overtime T3 is set to be smaller (= 2*T<T1); if the connection cannot be established with the server after the current third timeout T3, then perform the master-backup switch of the client;

Within the smaller fourth overtime T4 (=2*T) time, whether to establish a "heartbeat" connection with the host by the master-standby detection module (213) in the backup client (102), if it fails to set up, then perform client master backup switching;

Regardless of the server or the client, as long as it is detected that the host actively requires active-standby switchover, the connection establishment flag will be set to the "not connected" flag.

2. The method according to claim 1, wherein the smaller value of the first timeout T1 is equal to the second timeout T2, and the smaller value of the third timeout T3 is smaller than the smaller value of the first timeout T1 , the larger value of the third timeout T3 is equal to the larger value of the first timeout T1, the smaller value of the third timeout T3 is smaller than the smaller value of the first timeout, and the smaller value of the third timeout T3 is far from It is less than the time required to complete the active/standby switchover.

3. The method according to claim 1, wherein the larger value of the first timeout T1 is much larger than the smaller value of the first timeout T1, and the larger value of the third timeout T3 is much larger than the third timeout T3. The timeout T3 is a smaller value.

4. The method according to claim 1, characterized in that said server master-standby switchover comprises the following steps: if the link between the server and the client is a line (216), then select the line (218), if the server and the client If the link between the terminals is the line (218), then select the line (216), if the link between the server and the client is the line (217), then select the line (219), if the link between the server and the client Then select circuit (217) for circuit (219).

5. A communications system backup device, characterized in that it comprises a client (101) and a backup client (102), which are respectively connected to the hub 105 and the backup hub (106) through network adapters (103), (104), Also comprise server (107) and backup server (108), described server (107) and backup server (108) have two network adapters (109)-(110) and (111)-(112) respectively, through Connections (216), (217) and (218), 219 are connected to the hub 105 and the backup hub (106), there is a bidirectional communication link (120) between the main backup clients, and the server (107) and the backup server (108) are respectively connected to a shared database (113).

6. The device according to claim 5, characterized in that, the active server (107) includes a server host management module (201), a server active/standby state detection module (202), a first network detection module (203) ) and the second network detection module 204, including server backup machine management module (205), server active/standby state detection module (206), first network detection module (207) and second network detection module (207) in the standby server (108) module (208), wherein, the first network detection module (203), (207) is used to detect the connection between the active server, the backup server and the main client (101), and the second network detection module (204) , (208) are respectively used to detect the connection between the active server, the standby server and the main client (102), and the server host management module (201) instructs the first or second The second network detection module performs link connection detection; the server backup machine management module (205) instructs the first or second network detection module to perform link connection detection through the server active/standby state detection module (206).

7. The device according to claim 5, characterized in that there are a client server management module (209), a client active/standby state detection module 210 and a network detection module (211) in the client (101), and in the backup client The machine (102) includes a client server management module (212), a client active/standby state detection module (213) and a network detection module (214).

8. The device according to claim 6, characterized in that the server host management module (201) detects whether a connection is established with the client through the network detection module (203), and if so, sets the first timeout T1 to be small (=2.5 *T), otherwise the first overtime T1 is set to be larger (=20*T); if the host server cannot establish a connection with the client through the current first overtime T1, then the main server backup switch is performed.

9. The device according to claim 6, characterized in that, the server backup machine management module (205) detects whether the master/standby detection module 206 establishes " Heartbeat" connection, if the connection fails to be established, the server will switch between the primary and backup servers.

10. The device according to claim 7, characterized in that the client management module (209) detects whether a connection is established with the server through the network detection module (211), and if the connection is not established, the third timeout T3 is set to be larger (=20*T), otherwise the third timeout T3 is set to be smaller (=2*T<T1); if the connection with the server cannot be established after the current third timeout T3, then the main backup switch of the client is performed.

11. The device according to claim 7, characterized in that, the client management module (212) passes the master/standby detection through the network detection module (213) within a smaller fourth timeout T4 (=2*T) time The module (213) detects whether a "heartbeat" connection is established with the host, and if it fails to be established, the main backup switch of the client is performed.