CN100442722C

CN100442722C - Diagnostic method for network fault and its device

Info

Publication number: CN100442722C
Application number: CNB200510053457XA
Authority: CN
Inventors: 胡旻
Original assignee: Huawei Technologies Co Ltd
Current assignee: Huawei Technologies Co Ltd
Priority date: 2005-03-10
Filing date: 2005-03-10
Publication date: 2008-12-10
Anticipated expiration: 2025-03-10
Also published as: CN1832434A

Abstract

The present invention discloses a diagnosis method for network faults. The present invention solves the problems of the existence of inaccurate positioning, long positioning time and high cost in the prior art when network faults are diagnosed. The method has the procedures: the fault report of a user is received and recorded; whether the conversation of the user who reports faults is normal or not is judged, if the conversation is abnormal, user line faults are determined, and otherwise, connectivity test commands are sent out to a network side form a user resident device corresponding to the user; whether the connectivity of a protocol layer is normal or not is judged based on test results, if the connectivity is abnormal, protocol layer connectivity faults are determined, if not, the CPE simulation protocol operation or/and requests are sent to the network side, and whether network protocols are normal is determined according to operation results. The present invention also discloses a network fault diagnosis device simultaneously.

Description

A kind of diagnostic method of network failure and device thereof

Technical field

The present invention relates to broadband network technology, relate in particular to Digital Subscriber Line (DSL) broadband network fault Diagnosis Method and device thereof.

Background technology

The user inserts the DSL broadband network need pass through DSL subscriber's premises equipment (CPE), Digital Subscriber Line access module (DSLAM) (CPE and DSLAM are the DSL access device), BAS Broadband Access Server (BAS), dns server plurality of devices such as (DNS Server), when the user ran into network failure, a plurality of functional departments of needs (charging center, network management center, measuring chamber, ground office and outside line class etc.) coordination was found and is fixed a breakdown.

At present, at DSL network row barrier, telecom operators are general to adopt broadband test plate scheme to carry out the test of circuit at measuring chamber earlier, grabs the line function by the webmaster support and locatees line fault.If not line fault is then checked port status by the personnel of network management center, determined whether fault, perhaps assist to the user terminal location to understand phenomenon of the failure by the outside line personnel, each side links up the location repeatedly then, rule of thumb progressively investigates fault.After fixing a breakdown, need notify the user to attempt whether can normally using broadband network.

There is following shortcoming in above-mentioned Fault Locating Method:

1, line fault proportion in whole broadband network fault is generally less than 20%, and line test need be adopted expensive broadband test head, thus can be restricted because of the cost problem makes its enforcement, especially in medium and small office point.

2, whole row's barrier process is progressively investigated by experience, and not only the time is long and the human cost height, and the location also not exclusively accurately.

3, row barrier process need user cooperates at the scene, causes arranging the barrier cycle like this because human factor and long also influences customer satisfaction.

Summary of the invention

The invention provides a kind of diagnostic method and device thereof of network failure, have the problem that the location is inaccurate, positioning time is long and cost is high when the diagnose network faults to solve in the prior art.

A kind of diagnostic method of network failure, the user terminal in the described network insert external network by customer premises equipment, CPE (CPE), Digital Subscriber Line access module (DSLAM), BAS Broadband Access Server (BAS) successively, comprise the steps:

A, webmaster side joint are received and accuse the fault Shen of recording user; B, judge that whether the user's communication of fault is accused in the Shen normal, if normally then continue step C, otherwise determines the subscriber's line fault; C, initiate the continuity testing order to network side, judge according to test result whether the protocol layer connectedness is normal from the customer premises equipment, CPE (CPE) of described user's correspondence, if then continue step D, otherwise, determine the protocol layer failure of connectivity; D, send the simulation protocol operation or/and request to network side, and determine according to operating result whether procotol is normal from described CPE.

In the steps A, when the subscriber complaint fault, import whether normally information of conversation, in step B, judge directly according to this information whether this user's communication is normal by the voice suggestion user; Perhaps, in step B, judge by making a call whether user's communication is normal to Subscriber Number.

Step C may further comprise the steps:

C1, Digital Subscriber Line access module (DSLAM) port is activated and deactivating operation, if activate and deexcitation normally then continues step C2, otherwise definite local side fault; C2, customer premises equipment, CPE (CPE) is carried out remote reset, then determine this CPE fault, otherwise continue step C3 if can not normally reset; C3, initiate the Ping order to BAS Broadband Access Server (BAS),, then continue step C4 if BAS replys normally from CPE, otherwise failure of connectivity between definite CPE and BAS; C4, the net control device that makes progress to network side to BAS respectively from CPE are initiated the Ping order, if each net control device is replied normally, then continue subsequent step; Otherwise determine that there is failure of connectivity in BAS between improper net control device of replying.

Step D comprises the steps:

D1, call out to the PPP that BAS initiates emulation, judge whether BAS is normal to the response that PPP calls out from CPE, if then continue step D2, otherwise definite BAS protocol layer fault; D2, initiate the domain name mapping request to name server (DNS) from CPE, undesired if dns server is replied, then determine DNS Protocol layer fault.If determine that DNS is normal, also comprise the steps: D3, initiate file transfer protocol (FTP) (FTP) rate test from CPE, according to test result, determine whether to exist local network and/or external network speed low from CPE to normative document host-host protocol (FTP) testing server.

A kind of network fault diagnosis device comprises:

The diagnosis administrative unit is used for setting up the mapping relations between network user and the network of relation resource, receives the also fault Shen announcement of recording user, and it is unusual that the user's communication of fault is accused in the judgement Shen, then determines the subscriber's line fault; The user's communication of judging Shen announcement fault is normal, then issues the continuity testing order, and judges according to the test result that receives whether the protocol layer connectedness is normal, if unusual, then determines the protocol layer failure of connectivity; If normal, then issue the emulation testing order, and determine according to the test result that receives whether procotol is normal;

The diagnostic agent unit is used for the continuity testing order that issues according to described diagnosis administrative unit, initiates the continuity testing order from the customer premises equipment, CPE CPE of user's correspondence to network side, and returns test result to the diagnosis administrative unit; And, send the simulation protocol operation or/and ask to network side, and return test result to the diagnosis administrative unit from described CPE according to the emulation testing order that described diagnosis administrative unit issues.

Described diagnosis management unit pack is drawn together:

Interface module is accused in the fault Shen, is used to receive the Shen relevant with fault and accuses information;

The user resources administration module is used for user profile is tied to the physical port that the user uses;

The diagnosis Configuration Manager is used for the configuration testing parameter;

The network management interface module is used to issue test command;

The test assignment administration module is used for setting up test assignment according to the configuration testing parameter, and collects test result;

The Analysis on Fault Diagnosis module is used for carrying out accident analysis according to test result.

Described diagnosis administrative unit also comprises: Web service visit support module is used to provide the visit of browser mode.

The present invention has following beneficial effect:

1, by the automatic trouble diagnosis function of native system, telecommunications attendant is the fault location section rapidly, improves the barrier speed of looking into.

2, by the transformation of voice suggestion, force users powers on, and the erroneous judgement that reduces automatic trouble diagnosis is disconnected, and has collected the circuit on-off situation.

3, the validation test of carrying out automatically by system after telecommunications attendant arranges barrier, does not need the user to cooperate, and can carry out the whether normal checking of network.

4, by the automatic trouble diagnosis of native system, adopt standard FTP reference server and canonical reference website, the local telecommunication attendant can judge that the fault of access network still is the fault of the different route office directions of external network (metropolitan area network).

5, system manages in conjunction with user resources, directly arranges barrier based on the subscriber complaint telephone number, has reduced the inconvenience that the attendant operates in a plurality of systems.

6, the present invention is the scheme based on terminal management, has solved the fault of network 80%, does not need the broadband test head, therefore is suitable for the diagnostic test of medium and small office point.

Description of drawings

Fig. 1 is a broadband system end-to-end protocol stack schematic diagram;

Fig. 2 is the networking schematic diagram of trouble-shooter of the present invention and place system thereof;

Fig. 3 is the structural representation of the diagnosis administrative unit in the trouble-shooter;

Fig. 4 is the flow chart of diagnosis failure of connectivity among the present invention;

Fig. 5 is the flow chart of diagnosing protocol layer fault among the present invention.

Embodiment

From broadband network business of networking operation link, the whole service path can be divided into PC, PC-customer premises equipment, CPE (CPE), CPE, CPE-Digital Subscriber Line access module (DSLAM), DSLAM, DSLAM-BAS Broadband Access Server (BAS), BAS, BAS-intetnet server and internet server link, all these links all may have problems, the range of management of local telecommunication Virtual network operator is that CPE is between the BAS, the present invention covers PC-CPE, CPE, CPE-DSLAM, DSLAM, DSLAM-BAS, BAS, the failure diagnosis and the location of these links of BAS-internet server, the present invention does not diagnose the fault of PC and internet server itself.Described CPE refers to the terminal equipment used in the DSL network comprise ADSL CPE, VDSLCPE, G.SHDSL CPE etc.

Consult shown in Figure 1, the protocol description of common end peering protocol (PPP) dial user in dialing procedure and in the last network process following (annotate: PPP dialing be the mode of the individual domestic consumer of current major part practicality):

1, PPP dialing procedure protocol description

(1) in the PPP dialing, the user initiates PPP at the PC end and calls out and request authentication, and this authentication request is carried on the ethernet frame (MAC).

(2) on CPE, mac frame is encapsulated in ATM cell, and is packaged on the DSL superframe.

(3) the DSL superframe interaction data by fixed length between CPE and local side DSLAM access device.

(4) DSLAM separates ATM cell get off from the DSL superframe, if the DSLAM of ATM upstream Interface, just and between the BAS authenticating device by the communication of ATM cell mode; If the DSLAM of IP upstream Interface also needs to be reduced to mac frame from ATM cell, and by carrying out communication between mac frame and the BAS authenticating device.

(5) BAS solves PPP dialing authentication request package from mac frame, authenticates, if by authentication, distribute an IP address, is included in the PP response, sends to PC according to the direction opposite with authentication request, and the user just has the IP address like this, can surf the Net.Wherein, PPP calls out and terminates on BAS.

2, go up protocol description in the network process

When (1) surfing the Net, the user all is carried on IP and wraps at all application layers, the session-layer protocol (such as HTTP request, FTP, SMTP etc.) that the PC end uses, because various application are a lot, but all belong to upper-layer protocol, omitted enumerating of the application bag more than the IP bag here.All IP bags are carried on the ethernet frame (MAC).

(5) BAS receives the IP bag, directly is transmitted to the router on the internet, finally arrives the internet server of upper-layer protocol requirement.The Internet server is bundled to IP with the response message of application layer protocol and wraps, and returns to BAS, and BAS is sent to user's PC by rightabout with data.

In order to diagnose DSL broadband network fault low-cost and fast, the present invention inquires corresponding user resources by telephone number in system resource management after the subscriber complaint fault, judge whether subscriber's line is normal earlier; If subscriber's line is normal, then transmitting order to lower levels is initiated the continuity testing order from customer premises equipment, CPE (CPE) to network side automatically, judges according to test result whether the protocol layer connectedness is normal; If protocol layer is connective normal, then initiate the emulation property testing from CPE, promptly initiate the simulation protocol operation or/and network side is mail in request, and determine according to operating result whether procotol is normal at CPE.

Consult shown in Figure 2ly, trouble-shooter of the present invention comprises diagnosis administrative unit and diagnostic agent unit.The diagnosis administrative unit is arranged on the webmaster side, and the diagnostic agent unit can be arranged among the CPE.The diagnosis administrative unit is responsible for setting up the mapping relations between user and the network of relation resource, and issues the series of tests order automatically to the diagnostic agent unit among the CPE; The diagnostic agent unit automatically performs corresponding test operation after receiving test command, and returns test result to the diagnosis administrative unit, and the diagnosis administrative unit is carried out faulty section analysis and protocol layer analysis to test result, and the fault location section is also exported failure cause.

Consult shown in Figure 3ly, diagnosis management unit pack is drawn together:

Web service visit support module: system is based on the Apache platform, by the Web service MAC layer, the visit of IE mode is provided, realize zero-configuration client install (promptly do not need to install any client, use PC operating system with internet explorer get final product).

Interface module is accused in the fault Shen: be used to connect telecom operators OSS (Operation SupportSystem, OSS) single system is accused in the fault Shen, information is accused in the relevant Shen of accepting automatically to accuse the fault Shen, to realize automatic fault diagnosis functions.

User resources administration module:, support to carry out the inquiry of Complaining Subscriber with the concrete physical port locating information binding that broadband user's name, number of the account, telephone number and user use.

Analysis on Fault Diagnosis module: according to the result of user test, carry out faults analysis and judge, the key diagnosis function of (comprising the substep diagnosis) is provided.

The diagnosis Configuration Manager: the user disposes various test parameters as required, forms different configuration templates, uses in the time of for test.

Test assignment administration module: according to the user, adopt the diagnosis configuration to set up test assignment automatically, collect test result, offer the diagnostic analysis module and look into barrier.

Network management interface module: be used to connect network management system, issue Simple Network Management Protocol (SNMP) order to equipment, require to carry out test function by network management system.

When the user can't surf the Net, need accuse by carry out the fault Shen to 112 or 10000 of telecommunication bureau, stay its telephone number.Network management system can be found subscriber's line corresponding equipment port and CPE according to this telephone number.In order to judge whether subscriber's line is normal, fairly simple method is whether to require the user to import in the voice suggestion that accuse the fault Shen to make a phone call normal information, and user's telephone number and this information are recorded in the network management data storehouse.Certainly, can not require that the user imports yet, and determine by system's active call Subscriber Number whether circuit is normal.

In order not need the user to cooperate, open the power supply of CPE after when fault is accused in user's Shen, pointing out subscriber complaint by prompt tone at fault location and when barrier row.

After receiving the subscriber complaint fault ticket, the diagnosis administrative unit can find the user resources that the telephone number correspondence of fault is accused in the Shen from the network management data storehouse, use snmp protocol issue automatically various test commands give diagnostic agent unit on DSLAM and the CPE carry out the Complaining Subscriber network path on various emulation and physical testing, and test result is reported to the diagnosis administrative unit; The diagnosis administrative unit is carried out faulty section and protocol layer analysis according to the result of test, outputs test result and points out corresponding failure cause with graphics mode.

Consult shown in Figure 4ly, whether diagnosis IP layer connectivity exists the process of fault as follows:

Step 10: whether the conversation that finds user input according to user's telephone number normal information, normal if the user determines conversation, then carry out step 11, otherwise determines the subscriber's line fault.

Step 11: the DSLAM port is carried out deexcitation, activation manipulation,,, then carry out step 12 if can normally activate to ensure the eliminating of local side port problem; Otherwise determine that local side breaks down.

After breaking down, local side can get rid of by changing replacement methods such as port.

Step 12: remote reset CPE, if CPE resets normally, then carry out step 13, otherwise determine the CPE fault.

Because some fault may be a CPE software fault problem, therefore, long-range CPE is resetted carrying out needing to carry out before series detects, get rid of generic software fault (because end user's level difference problem, even currently accuse after the prompt tone of platform or manual board require the user to reset in the Shen, system also needs to carry out automatically this operation, declares the basis of barrier to ensure whole flow process).Can adopt the replacement method to arrange barrier for the CPE fault.

Step 13: initiate ping test from CPE, the destination address of Ping is the IP address of BAS, to detect the connectedness from CPE to BAS, if BAS returns normal Ping Test result, shows that then CPE is normal to the connectedness of BAS, carry out step 14; Otherwise determine to have failure of connectivity between CPE and the BAS, carry out step 15.

Step 14: be initiated to the ping test of dns server, egress router, internet server from CPE, if each server all can return normal Ping Test result, show that then CPE arrives the connective normal of each server, otherwise determine that there is failure of connectivity in CPE between the server that returns abnormal Ping Test result.

Step 15: carry out ATM(Asynchronous Transfer Mode) layer F5 loopback, detect the connectedness of BAS to DSLAM successively, DSLAM is to the connectedness of CPE, and the fault that the ATM layer is found adopts the replacement method to arrange barrier.If it is no problem that the ATM layer detects, then determine the IP layer software existing problems of BAS or DSLAM, can be by changing method row barriers such as port or load software.

OAMF5 layer loopback, promptly in ATM cell, insert OAM loopback diagnosis cell, realize the data transmission quality diagnosis of ATM layer: at first be looped back to the DSL port of Complaining Subscriber correspondence, detect the connection situation of ATM layer between DSLAM and the BAS from the ATM (downlink port) of BAS; Then, be looped back to the DSL port of CPE, detect that ATM connects situation (the diagnostic agent unit by CPE provides support) between DSLAM and the CPE from the atm port of DSLAM.

If normal, then need the test protocol layer whether to have fault to the IP layer connectivity.As shown in Figure 5, the test process of protocol layer is as follows:

Step 20: initiate PPP (point-to-point) emulation testing from CPE, promptly initiate the PPP calling (consistent by the process that CPE calls out) of similar PC with the user from CPE, determine according to BAS equipment return results whether BAS is normal to the response that PPP calls out, if normally then carry out step 21, otherwise determine BAS protocol layer fault.

Whether call out the failure cause return according to PPP, can orient is that BAS distributing IP address is unsuccessful, perhaps, and multiple reason (supporting) such as authentication password is incorrect by the diagnostic agent unit on the CPE.

Step 21: carry out name server (DNS) emulation testing, promptly initiate DNS and ask analog subscriber PC online request domain name analysis process by CPE, judge that according to the dns server return results whether normal DNS Protocol reply, if normally then carry out step 22, otherwise determine that there is fault in DNS.

Step 22: carry out file transfer protocol (FTP) (FTP) and test the speed, promptly initiate the FTP request to standard FTP testing server, and calculate the FTP downloading rate according to the size and the download time of file in download from CPE.If exist FTP speed slow, then carry out step 23, if exist agreement not support situation, then point out.

According to determining that from CPE to the test result of standard FTP testing server local network still is that external network speed is slow.

Step 23: repeatedly activate CPE again, test the speed by line activation parameter and FTP, judgement is that activation rate is low or upper-layer protocol speed is low, if activation rate is low, can determine that line scan pickup coil side wears out and crosstalks and cause, if activation rate is normal, and FTP speed is low, then determine it is network side outlet bottleneck problem, need to adjust the bandwidth on this office direction.

Behind all faults row barrier, all need to carry out again PPP emulation testing and FTP and test and guarantee the user network availability.If pinpoint the problems in the validation test, just diagnose again.If validation test is normal, then notify client's fault to get rid of.

Convenient in order to test, can customize following test template:

Customization Ping Test purpose IP template: customize different templates at different DHCP (DHCP) server, dns server, BAS server, the router that inserts internet and operator standard testing website, like this at different users, just can adopt different templates to test, be determined to the break-make situation of different servers or different segment fast.

The destination server template of customization FTP rate test: the slow reason of networking speed has a variety of, concerning operator, mainly be the speed that the network oneself provide is provided, by setting up the ftp server of standard, at the speed quality of this server test, the network that can differentiate operator oneself provides still is the problem of external web services quality.

Customization DNS/PPP simulation modular: owing to need built-in account information when carrying out PPP emulation, so, adopt different templates that different PPP emulation number of the account parameters can be set, convenient definition voluntarily at the operator of different regions.

By realizing and the interface of OSS system, the row of whole system hinders, test and draw diagnosis all automatically performs.

Adopt the present invention, telecom operators can the quick fault location of autorun occur in which physics link of broadband network, which protocol layer.The present invention is based on the scheme of terminal management, will have very big commercial value in the application of medium and small office point.

Obviously, those skilled in the art can carry out various changes and modification to the present invention and not break away from the spirit and scope of the present invention.Like this, if of the present invention these are revised and modification belongs within the scope of claim of the present invention and equivalent technologies thereof, then the present invention also is intended to comprise these changes and modification interior.

Claims

1, a kind of diagnostic method of network failure, the user terminal in the described network insert external network by customer premises equipment, CPE CPE, Digital Subscriber Line access module DSLAM, BAS Broadband Access Server BAS successively; It is characterized in that, comprise the steps:

A, webmaster side joint are received and accuse the fault Shen of recording user;

B, judge that whether the user's communication of fault is accused in the Shen normal, if normally then continue step C, otherwise determines the subscriber's line fault;

C, initiate the continuity testing order to network side, judge according to test result whether the protocol layer connectedness is normal from the customer premises equipment, CPE CPE of described user's correspondence, if then continue step D, otherwise, determine the protocol layer failure of connectivity;

D, send the simulation protocol operation or/and request to network side, and determine according to operating result whether procotol is normal from described CPE.

2, whether normally the method for claim 1 is characterized in that, in the steps A, import conversation information by the voice suggestion user when the subscriber complaint fault, judges directly according to this information whether this user's communication is normal in step B; Perhaps, in step B, judge by making a call whether user's communication is normal to Subscriber Number.

3, method as claimed in claim 2 is characterized in that, in the steps A, also opens the CPE power supply by the voice suggestion user when the subscriber complaint fault.

4, the method for claim 1 is characterized in that, step C may further comprise the steps:

C1, Digital Subscriber Line access module DSLAM port is activated and deactivating operation, if activate and deexcitation normally then continues step C2, otherwise definite local side fault;

C2, customer premises equipment, CPE CPE is carried out remote reset, then determine this CPE fault, otherwise continue step C3 if can not normally reset;

C3, initiate the Ping order to BAS Broadband Access Server BAS,, then continue step C4 if BAS replys normally from CPE, otherwise failure of connectivity between definite CPE and BAS;

C4, the net control device that makes progress to network side to BAS respectively from CPE are initiated the Ping order, if each net control device is replied normally, then continue subsequent step; Otherwise determine that there is failure of connectivity in BAS between improper net control device of replying.

5, method as claimed in claim 4 is characterized in that, determines also to comprise the steps: behind the failure of connectivity between CPE and BAS in step C3

C31, respectively carrying out asynchronous transfer mode ATM layer F5 loopback between CPE and the DSLAM and between DSLAM and the BAS to determine whether the ATM layer exists fault, if the ATM layer is normal, then determine the IP layer software fault of BAS or DSLAM.

6, as each described method of claim 1 to 5, it is characterized in that step D comprises the steps:

D1, call out to the PPP that BAS initiates emulation, judge whether BAS is normal to the response that PPP calls out from CPE, if then continue step D2, otherwise definite BAS protocol layer fault;

D2, initiate the domain name mapping request to domain name server (DNS) from CPE, undesired if dns server is replied, then determine DNS Protocol layer fault.

7, method as claimed in claim 6 is characterized in that, determines that DNS just often also comprises the steps:

D3, initiate file transfer protocol (FTP) FTP rate test,, determine whether to exist local network and/or external network speed low according to test result from CPE to normative document host-host protocol FTP testing server from CPE.

8, method as claimed in claim 7 is characterized in that, if when existing FTP speed slow, also comprises the steps:

D4, repeatedly activate CPE, test the speed by line activation parameter and FTP, determine activation rate low still be that upper-layer protocol speed is low, if activation rate is low, then determine to exist line scan pickup coil side aging and crosstalk; If upper-layer protocol speed is low, determine that then there is bottleneck in the network side outlet.

9, method as claimed in claim 8 is characterized in that, this method also comprises step:

E, fix a breakdown;

Whether F, the PPP that initiates emulation to BAS from CPE call out and initiate file transfer protocol (FTP) FTP rate test from CPE, get rid of with validation fault.

10, a kind of network fault diagnosis device is characterized in that, comprising:

Diagnose administrative unit, be used for setting up the mapping relations between network user and the network of relation resource, receive the also fault Shen announcement of recording user,, then determine the subscriber's line fault if judge that the user's communication of Shen announcement fault is unusual; If judge that the user's communication of Shen announcement fault is normal, then issue the continuity testing order, and judge according to the test result that receives whether the protocol layer connectedness is normal, if unusual, then determine the protocol layer failure of connectivity; If normal, then issue the emulation testing order, and determine according to the test result that receives whether procotol is normal;

11, trouble-shooter as claimed in claim 10 is characterized in that, described diagnosis management unit pack is drawn together:

The network management interface module is used to issue test command;

12, trouble-shooter as claimed in claim 11 is characterized in that, described diagnosis administrative unit also comprises:

Web service visit support module is used to provide the visit of browser mode.

As claim 10,11 or 12 described trouble-shooters, it is characterized in that 13, described diagnostic agent unit is arranged among the customer premises equipment, CPE CPE in the network, described diagnosis administrative unit is arranged in the network of network management system.