WO2012037760A1

WO2012037760A1 - Method, server and system for improving alarm processing efficiency

Info

Publication number: WO2012037760A1
Application number: PCT/CN2010/079942
Authority: WO
Inventors: 郭慧峰; 张伟
Original assignee: 中兴通讯股份有限公司
Priority date: 2010-09-20
Filing date: 2010-12-17
Publication date: 2012-03-29
Also published as: CN101958804A; CN101958804B

Abstract

A method, server and system for improving alarm processing efficiency are disclosed by the present invention, wherein the method includes: alarm information carrying network element identifiers is received at certain time, and on the basis of the order of reporting the alarm information, the alarm information is orderly put into an alarm queue to be processed; the alarm information includes alarm generation information or alarm clear information; the alarm information in the alarm queue to be processed is batch-processed with parallel-thread on the basis of the network element identifiers carried in the alarm information and on the basis of a preset rule. With the present invention, the slave alarm server uses a preset time delay to collect the alarm information, and then transmits the alarm information to the master alarm server in batch, thus the batch-processing ability of the master alarm server is enhanced; meanwhile, in the master alarm server, multi-threading batch-processing is realized on the basis of the preset rule to greatly improve the parallelism for alarm processing, thus the alarm processing efficiency of network management is improved.

Description

Method, server and system for improving alarm processing efficiency

Technical field

The present invention relates to the field of network management technologies, and in particular, to a method, a server, and a system for improving the efficiency of alarm processing in a network management system. Background technique

Currently, the core function of alarm management in the transport network management is to monitor, report, and store alarms from network element hardware devices or path nodes in real time. When the NE device detects a fault, it reports the alarm information to the NMS. When the fault is rectified, the alarm disappears to the NMS. The NMS needs to process and display the alarm information in a timely manner.

The alarm management function in the network management system must ensure the accuracy and timeliness of the alarm information processing. For the alarms of the same network element, the alarm generation and alarm disappearance have timing requirements. The network management system must first process the alarm generation information and then process the alarm disappearance information to ensure The accuracy of the alarm information processing of the network management system.

The traditional method for processing alarm information is serial processing. This method can meet the processing requirements when the alarm amount is small. However, as the network management scale continues to expand, there are thousands of devices in the network management, and the amount of alarm data to be processed. It is very large, and the new generation of network management requires that the current alarm data and historical alarm data are required to enter the network management database, thereby increasing the proportion of interaction with the database in the alarm processing flow, and the traditional serial processing method is bound to cause delay in alarm processing. Increased, the alarm processing efficiency is reduced, and the user's high performance requirements for the network management system cannot be met. Summary of the invention

The main purpose of the present invention is to provide a method, a server and a system for improving the efficiency of alarm processing, which are intended to improve the processing efficiency of alarms in the network management system.

The present invention provides a method for improving the efficiency of alarm processing, including:

The alarm information carrying the identifier of the network element is periodically received, and the alarm information is sequentially placed in the alarm queue to be processed according to the sequence of reporting the alarm information; Performing parallel thread batch processing on the alarm information in the to-be-processed alarm queue according to the network element identifier and the predetermined rule carried in the alarm information.

Preferably, the step of performing parallel thread batch processing on the alarm information in the to-be-processed alarm queue according to the network element identifier and the predetermined rule carried in the alarm information includes:

When there is an idle working thread among the preset n working threads, accessing the first alarm information in the current pending alarm queue;

Dividing the identifier of the network element carried by the alarm by n to obtain the remainder i;

Determine whether the worker thread i is idle;

If the worker thread i is idle, the alarm information of all the remaining alarms in the current alarm queue is taken out to form an alarm processing task, which is dispatched to the worker thread i; and returns to the preset n working threads. The step of accessing the first alarm information in the current pending alarm queue when there is an idle working thread;

If the worker thread i is busy, the next alarm information in the current pending alarm queue is accessed, and the network element identifier carried in the alarm information is divided by n to obtain the remainder i.

Preferably, the step of performing parallel thread batch processing on the alarm information in the to-be-processed alarm queue according to the network element identifier and the predetermined rule carried in the alarm information further includes:

When the preset n working threads are not idle, the execution step is suspended after the predetermined time. When there are idle working threads among the preset n working threads, the first alarm information in the current pending alarm queue is accessed.

Preferably, the step of performing parallel thread batch processing on the alarm information in the to-be-processed alarm queue according to the network element identifier and the predetermined rule carried in the alarm information further includes: the alarm processing of the dispatched by the worker thread i The task is handled by itself;

The step of the i-number worker thread processing the scheduled alarm processing task includes: batch processing alarm generation information of all corresponding network elements before the first alarm disappearance information of each network element in the current alarm processing task;

Deleting the processed alarm generation information from the current alarm processing task, and generating the processed alarm generation message into the network management database; Batching the alarm disappearance information of all corresponding NEs before the first alarm generation information of each network element in the current alarm processing task;

The processed alarm disappearance information is deleted from the current alarm processing task, and the processed alarm disappearance message is entered into the network management database; and all corresponding to the first alarm disappearance information of each network element in the batch current alarm processing task is returned. The step of generating alarms by the alarm of the NE.

The present invention provides a server for improving the efficiency of alarm processing, including:

The receiving module is configured to: periodically receive the alarm information carrying the network element identifier, and place the alarm information into the alarm queue to be processed according to the sequence of reporting the alarm information; the alarm batch processing module is set as follows: The network element identifier and the predetermined rule carried in the alarm information perform parallel thread batch processing on the alarm information in the to-be-processed alarm queue.

Preferably, the alarm batch module includes:

An access unit, configured to: access the first alarm information in the current pending alarm queue when there is an idle working thread among the preset n working threads; and access the current pending alarm when the worker thread i is busy The next alarm information in the queue;

a computing unit, configured to: divide the network element identifier carried in the alarm information by n, to obtain a remainder i;

a determining unit, configured to: determine whether the worker thread i is idle;

The thread scheduling processing unit is configured to: when the worker thread i is idle, the alarm information of all the remaining alarm queues in the current alarm queue is taken out to form an alarm processing task, and is scheduled to be processed by the worker thread i.

Preferably, the thread scheduling processing unit is further configured to: suspend the predetermined time when none of the preset n working threads are idle.

Preferably, the thread scheduling processing unit includes:

The batch processing unit is configured to: batch process alarm generation information of all corresponding network elements before the first alarm disappearing information of each network element in the current alarm processing task; and batch processing the first network element in the current alarm processing task The alarm disappearing information of all corresponding network elements before the alarm generates information; The sub-unit is deleted, and is set to: delete the processed alarm generation information from the current alarm processing task; and delete the processed alarm disappearance information from the current alarm processing task.

The present invention also provides a system for improving the efficiency of alarm processing, including a slave alarm server and a primary alarm server;

The slave alarm server is configured to: receive the alarm information that is sent by the network element and carry the identifier of the network element, perform data conversion and classification processing on the alarm information, and report the alarm information to the primary alarm server periodically;

The primary alarm server is configured to: periodically receive the alarm information, and put the alarm information into the alarm queue to be processed according to the sequence of reporting the alarm information, according to the network element identifier and the predetermined rule carried by the alarm information. Parallel thread batch processing is performed on the alarm information in the alarm queue.

Preferably, the primary alarm server is a server as described above.

The present invention provides a method, a server, and a system for improving the efficiency of alarm processing. The master-slave alarm server collects alarm information by using a predetermined time delay (usually preset to 1 second), and then sends the alarm information to the primary alarm server in batches. The batch processing capability of the main alarm server is enhanced. At the same time, the multi-thread parallel processing is implemented according to the predetermined rule in the main alarm server, which greatly improves the parallelism of the alarm processing and improves the processing efficiency of the network management alarm. BRIEF abstract

1 is a schematic flow chart of an embodiment of a method for improving alarm processing efficiency according to the present invention;

2 is a schematic flowchart of a parallel thread batch processing for processing alarm information in an alarm queue according to an identifier and a predetermined rule according to an embodiment of the present invention. FIG. 3 is a method for improving alarm processing efficiency according to the present invention. A specific process diagram of processing the alarm processing task by the worker thread i in an embodiment;

4 is a schematic structural diagram of an embodiment of a server for improving alarm processing efficiency according to the present invention; FIG. 5 is a schematic structural diagram of an alarm batch processing module in an embodiment of a server for improving alarm processing efficiency according to the present invention; FIG. 6 is a schematic structural diagram of a thread scheduling processing unit in an embodiment of a server for improving alarm processing efficiency according to the present invention; FIG. 7 is a schematic structural diagram of an embodiment of a system for improving alarm processing efficiency according to the present invention.

In order to make the technical solutions of the present invention clearer and clearer, the following will be further described in conjunction with the accompanying drawings. Preferred embodiment of the invention

The solution of the embodiment of the present invention mainly performs parallel thread batch processing on the alarm information of >3⁄4 on the network element according to the preset working thread and the identifier of each network element. The present invention can be applied to a network system that needs to process alarm information, such as a network management system. The following is a description of the technical solution of the present invention by applying the present invention to a network management system, but is not limited to the case of being applied to a network management system.

As shown in FIG. 1 , an embodiment of the present invention provides a method for improving the efficiency of alarm processing, including: Step S101: periodically receiving an alarm message carrying a network element identifier, and sequentially, according to the sequence of reporting the alarm information, Put in the pending alarm queue;

The method operating environment of the embodiment includes a slave alarm server and a master alarm server that performs parallel thread processing on the alarm information, wherein the slave alarm server can receive the alarm device through the network management adapter (referred to as the network element in this embodiment, hereinafter referred to as The alarm information reported by the network element is used by the network management adapter to convert the alarm data reported by the network element into a packet that can be recognized by the network management system, and then report it to the secondary alarm server.

After that, the alarm server classifies each alarm information in the order in which the alarm information is reported in the order of the alarm information. The alarm information is divided into alarm generation information and alarm disappearance information. When the classification is performed, the alarm type server is affixed with a processing type label for each alarm information. For example, the alarm information ME{l}ANew is ALARM_NEW, which indicates that the alarm generation information of the network element 1 is reported, ME{l}BClear It is ALARM-CLEAR, which indicates that the alarm disappearing information of NE 1 is reported. The number in parentheses indicates the unique identifier of the NE in the NMS, and the letters A, B, and C represent the identifier of the alarm information. Then every once in a while (this time can be preset, usually 1 second), will be collected from the alarm server The alarm information is reported to the primary alarm server in batches.

In this embodiment, the alarm information is collected from the alarm server for a period of time delay, and the purpose is to enhance the ability of the main alarm server to process the alarm information in batches.

In this embodiment, the primary alarm server is configured with a pending alarm queue, which is used to receive the alarm information to be processed, which is reported in batches from the alarm server. After receiving the alarm information carrying the identifier of the network element, the primary alarm server puts the alarm information into the alarm queue to be processed in the order in which the alarm information is reported.

Step S102: Perform parallel thread batch processing on the alarm information to be processed in the alarm queue according to the network element identifier and the predetermined rule carried in the alarm information.

In this embodiment, for the alarm information processing, the alarm information generated or disappeared by the same network element has a timing requirement, but the alarm information of different network elements has no timing requirement, so only the alarm information of the same network element is guaranteed. When the same worker thread is processed, the timing requirements can be met, and the alarm information of different network elements can be multi-threaded to improve the alarm processing efficiency.

Therefore, in the main alarm server, the n-th worker thread (the worker thread No. 0, the worker thread No. 1 ... the worker thread n-1) can be pre-configured to execute the alarm processing task in parallel, and set one. The special scheduling thread executes module A, which is used to construct an alarm processing task from the buffer area, and dispatches each alarm processing task to n working threads for processing.

In this embodiment, the predetermined rule is to determine, by using the correspondence between the network element identifier and the number n of the set working threads, which one of the n working threads is to process the alarm information of the network element, and the alarm information belongs to the network. The categorization is performed by using the unique identifier of the network element. The foregoing predetermined rule may be: if the remainder of the network element identifier divided by n is represented by i, the alarm information of the network element is fixed to work at i. Thread processing. The process of batch processing alarm information of the primary alarm server is shown in Figure 2.

As shown in FIG. 2, step S102 specifically includes:

Step S1021, determining whether the number of alarm information in the alarm queue to be processed is 0, and if yes, proceeding to step S1029; otherwise, proceeding to step S1022;

Step S1022, determining whether there are idle working threads among the preset n working threads, and if so, Then proceeds to step S1023; otherwise, proceeds to step S1029;

Step S1023: Access the first alarm information in the current pending alarm queue.

Step S1024: Dividing the identifier of the network element carried in the alarm information by n to obtain a remainder i; i is any integer of 0~η-1.

Step S1025, determining whether the worker thread of the i line is idle, if yes, proceeding to step S1026; otherwise, proceeding to step S1027;

Step S1026: The alarm information of all the remaining numbers in the current pending queue is taken out to form an alarm processing task, and is scheduled to be processed by the i worker thread; and the process returns to step S1021;

Step S1027, determining whether there is a next alarm message, if yes, proceeding to step S1028; otherwise, proceeding to step S1029;

Step S1028, accessing the next alarm information in the current pending alarm queue, and proceeding to step S1024;

In step S1029, the predetermined time is suspended, and the process returns to step S1021.

As shown in FIG. 3, the step of processing the alarm processing task by the worker thread i in step S1026 specifically includes:

Step S10261, determining whether the number of alarm information in the current alarm processing task is 0, and if so, ending the process; otherwise, proceeding to step S10262;

Step S10262, batch processing alarm generation information of all corresponding network elements before the first alarm disappearing information of each network element in the current alarm processing task;

In this step, all the alarm generation information of the network element corresponding to the first alarm disappearing message in each network element is taken out, and all the extracted alarm generation information is formed into a batch for processing.

Step S10263: Delete the processed alarm generation information from the current alarm processing task, and enter the processed alarm generation information into the network management database;

Step S10264: Batch process the alarm disappearance information of all the corresponding network elements before the first alarm generation information of each network element in the current alarm processing task; In this step, all the alarm disappearance information of the network element before the first alarm generation information corresponding to each network element in the current alarm processing task is taken out, and all the removed alarm disappearance information is formed into a batch for processing. .

The current alarm processing task in this step includes all current unprocessed alarm information in the worker thread.

Step S10265: Delete the processed alarm disappearance information from the current alarm processing task, and enter the processed alarm disappearance information into the network management database; and return to step S10262.

In the above steps S10261 to S10265, when the working thread processes the alarm information in an alarm processing task, since the alarm generation and disappearance have timing requirements, the alarm generation information must be processed first and then the alarm disappearance information is processed, so the alarm processing is performed. The alarm information in the task must be processed serially. To improve the batch processing capability of the alarm information, the alarm generation information and the alarm disappearance information are processed in batches, and the processing is performed according to the alarm generation -> alarm disappearance -> alarm generation -> alarm disappearance loop execution until all alarm information It is processed. In the case of ensuring the correctness of the alarm data processing, the batch processing of the alarm information should include alarm data as much as possible.

The workflow of the method of this embodiment is described in detail below by way of an example:

Assume that the current network management manages four devices (that is, the network element in this embodiment), which are ME{1}, ME{2}, ME{3}, and ME{4}, respectively, where the numbers in parentheses indicate the network management of the device. ID, which is the unique identifier of the device in the network management.

Three main working threads are set to process the alarm information in parallel in the main alarm server, which are Thread0.

Threadl and Thread2. A special scheduling thread execution module A is set in the main alarm server.

According to the predetermined rule, the network management ID of ME{1} is 1 and 1%3 is equal to 1. Therefore, the alarm information of ME{1} is processed by the thread Thread1, and the network management ID of ME{2} is 2, 2%3 is equal to 2. Therefore, the alarm information of ME{2} is processed in the thread Thread2, and so on, the alarm information of ME{1} and ME{4} can be processed in the thread Thread1, and the alarm information of ME{2} is in Thread2. No. worker thread processing, ME{3} alarm information is processed in the ThreadO worker thread.

In this example, if the current alarm is reported as:

10:35:20:175 At the moment, ME{1} reported ME{l}ANew, ME{l}BClear, ME{l}CNew three alarm messages;

At 10:35:20:275, ME{2} reported two alarm messages: ME{2}ANew, Me{2}BNew;

At 10:35:20:375, ME{3} reported the ME{3}AClear-strip alarm information;

At 10:35:20:475, ME{4} reported ME{4}ANew, ME{4}BNew,

ME{4}CClear3 alarm information.

Among them, 10:35:20:175 means a certain moment, which means 10:35:20, 175 milliseconds.

ME{l}Anew indicates that the alarm information belongs to the ME{1} device, A is the identifier of the alarm information, New indicates that the new alarm generation information, ME{l}BClear indicates that the alarm information belongs to the ME{1} device, and B is The identifier of the alarm information, Clear indicates that the alarm disappears. The names of other alarm messages are interpreted accordingly.

After the adapter is converted, the above alarm information is reported to the slave alarm server. The order of reporting is in the order in which the alarms are generated: ME{l}ANew, ME{l}BClear, ME{l}CNew, ME{2}ANew, Me{2}BNew, ME{3}AClear, ME{4 }ANew, ME{4}BNew, ME{4}CClear.

Then, the alarm information is sorted in turn from the alarm server, and the processing type label label is affixed. The alarm server sends a batch alarm information to the main alarm server every 1 second, assuming that the slave alarm server is at 10:35:20 last time. The batch alarm information sent at time:000, the time delay of 1 second is delayed to 10:35:21:000, and the newly reported alarm information is reported to the main alarm server from the alarm server, that is, ME{l}ANew, ME{l}BClear, ME{l}CNew, ME{2}ANew, Me{2}BNew, ME{3}AClear, ME{4}ANew, ME{4}BNew, ME{4}CClear - send Give the primary alarm server.

The alarm server sends the alarm information to the alarm queue to be processed according to the sequence of the alarm information. Therefore, the current alarm queue to be processed is:

{ME{l}Anew, ME{l}BClear, ME{l}CNew, ME{2}Anew, Me{2}BNew,

ME{3}AClear, ME{4}Anew, ME{4}BNew, ME{4}CClear }.

The scheduling thread executing module A, which is responsible for task scheduling, performs task scheduling on the to-be-processed alarm information to be processed in the alarm queue. The scheduling process is as follows: 1. Traversing all the alarm information in the queue, and discovering that the ME{l}Anew alarm information belongs to the ME{1} network element. According to the predetermined rule, 1%3 is equal to 1, and the alarm should be processed in the Thread1 working thread, and if If the Thread1 is idle, the alarm information of all the NMS IDs in the queue divided by 3 is taken out to form an alarm processing task, which is scheduled to be processed in Thread1. The alarm information included in the alarm processing task includes {ME{l}. Anew, ME{l}BClear, ME{l}CNew, ME{4}Anew, ME{4}BNew, and ME{4}CClear}. The remaining alarm information in the primary alarm server includes { ME{2}Anew, Me{2}BNew, and ME{3}AClear}.

2. The scheduling thread executes module A and then accesses the first alarm information in the remaining alarm information. Similarly, the ME{2}Anew alarm belongs to the ME{2} network element. According to the predetermined rule, 2%3 is equal to 2, and the alarm is generated. The information should be processed in the Thread2 worker thread. If Thread2 is idle, the alarm information of all the network management IDs in the queue divided by 3 is taken out to form an alarm processing task, which is dispatched to Thread2 for processing. The alarm information contained in the processing task includes {ME{2}Anew, ME{2}BNew}. The remaining alarm information in the primary alarm server includes { ME{3}AClear}

3. The scheduling thread execution module A accesses the first alarm information in the remaining alarm information, and obtains: ME{3} The AClear alarm information belongs to the ME{3} network element, and according to the predetermined rule, 3%3 is equal to 0, and the alarm information is It should be processed in the ThreadO working thread, and if ThreadO is idle at this time, the alarm information of all network element IDs divided by 3 is taken out to form an alarm processing task, which is dispatched to ThreadO for processing. The alarm information contained in it includes {ME{3}AClear}. At this time, the remaining alarm information in the primary alarm server is empty.

4. At this time, the alarm queue to be processed is empty, then pause for 1 second, and then continue to traverse the alarm information for scheduling.

The final scheduling execution results are shown in Table 1 below:

{ME{l}Anew,

ME{l}BClear,

ME{l}CNew, {ME{2}Anew,

{ME{3}AClear } Service ME{4}Anew, Me{2}BNew}

ME{4}BNew,

ME{4}CClear }

Table 1

As can be seen from the above table 1, the above-mentioned alarm processing mode can achieve the maximum parallel processing effect on the batch alarm information.

The following takes the worker thread ThreadO as an example to describe in detail the process of the worker thread ThreadO performing alarm processing:

The alarm processing tasks received by the ThreadO worker thread include the following alarm messages: {ME{l}Anew, ME{l}BClear, ME{l}CNew, ME{4}Anew, ME{4}BNew, ME{4}CClear }.

1. The batch alarm generates a message. For all the NEs in the alarm processing task, the alarm generation information before the first alarm corresponding to the NE disappears is taken out for batch processing. There are two network elements in the current task, ME{1} and ME{4}. The first alarm disappearing information of ME{1} is ME{l}BClear, which will be ME{1} before ME{l}BClear The alarm generation information is taken out for processing, that is, ME{l}Anew can be taken out, and the first alarm disappearing information of ME{4} is ME{4}CClear, and ME{4} before ME{4}CClear The alarm generation information is taken out, and ME{4}Anew and ME{4}BNew are taken out, and finally an execution queue {ME{l}Anew, ME{4}Anew and ME{4}BNew} is formed and batched. After the alarm is processed, the alarm information in the execution queue is deleted in the alarm processing task, and the alarm data is sent to the network management database.

2. The batch alarm disappears. At this time, the alarm information in the alarm processing task remains { ME{l}BClear, ME{l}CNew, ME{4}CClear}. The alarm disappearance information before the first alarm generation information corresponding to all the NEs in the current alarm processing task is taken out and batch processing is performed. The first alarm generation information of ME{1} is ME{l}CNew, and the alarm disappearance information of ME{1} before ME{1}CNew is taken out, that is, ME{l}BClear is taken out; ME{4} has If no alarm is generated, all the alarm disappearance information of the network element is taken out, and ME{4}CClear is taken out. Finally, an execution queue { ME{l}BClear, ME{4}CClear } is formed, after batch processing, The alarm information in the execution queue is deleted in the alarm processing task, and the processed alarm data is entered into the network management database.

3. Batch alarms are generated. At this time, the alarm information remaining in the alarm processing task is { ME{l}CNew} „ The alarm generation information before the first alarm disappearance information corresponding to all the NEs in the current alarm processing task is taken out and batch processed. ME {1} If there is no alarm disappearing information, all the alarm generation information of the network element is taken out, that is, ME{l}CNew is taken out, and finally an execution queue {ME{1}CNew} is formed, and after processing, the alarm is generated. The alarm information in the execution queue is deleted from the processing task, and the processed alarm data is entered into the network management database.

4. When the alarm information in the alarm processing task is empty, the task execution ends. If it is not empty, the batch alarm disappearance information is continued.

Therefore, this alarm processing task serially executes the following batches of alarm information:

{ ME{l}Anew, ME{4}Anew, ME{4}BNew }, go to the alarm generation information batch process, and finally put the alarm data into the network management database.

{ ME{l}BClear, ME{4}CClear }, go to the alarm disappearing information batch processing flow, and finally put the alarm data into the network management database.

{ ME{l}CNew } , go to the alarm to generate the information batch processing flow, and finally put the alarm data into the network management database.

It can be seen from the above alarm information processing process that: in the case of ensuring the correctness of the alarm data processing, each alarm processing batch contains as many alarms as possible to the maximum extent, and the batch processing of the alarm information is maximized. ability.

Compared with the traditional alarm processing method, this embodiment has the following advantages:

1. Introduce the "master-slave alarm server", collect the alarms in the "slave alarm server" with a certain time delay (usually 1 second), and then send them to the "master alarm server" in batches, and enhance the batch of the "master alarm server". Processing capacity.

2. In the "main alarm server", multi-thread processing is implemented according to certain rules, thereby greatly improving the parallelism of alarm processing and improving the processing efficiency of network management alarms.

As shown in FIG. 4, an embodiment of the present invention provides a server, a package for improving alarm processing efficiency. The receiving module 401 and the alarm batch processing module 402; wherein:

The receiving module 401 is configured to periodically receive the alarm information carrying the network element identifier, and sequentially put the alarm information into the alarm queue to be processed according to the sequence of reporting the alarm information;

The alarm information includes alarm generation information or alarm disappearance information.

The alarm batch processing module 402 is configured to perform parallel thread batch processing on the alarm information to be processed in the alarm queue according to the network element identifier and the predetermined rule carried in the alarm information.

In this embodiment, the receiving module 401 periodically receives the reported batch alarm information, and the batch alarm information is reported by the alarm network element to the network management adapter, and after the data is converted by the network management adapter, the alarm information is sent to the slave alarm server for classification. It is reported to the main alarm server receiving module 401 in batches from the alarm server timing (every predetermined time, usually 1 second).

In this embodiment, the primary alarm server is configured with a pending alarm queue, which is used to receive the alarm information to be processed, which is reported in batches from the alarm server. After receiving the alarm information carrying the network element identifier, the primary alarm server receiving module 401 sequentially puts the alarm information into the alarm queue to be processed according to the sequence of reporting the alarm information.

Therefore, in the main alarm server alarm batch processing module 402, the n working threads (the working thread No. 0, the working thread No. 1 ... the working thread n-1) can be pre-set to perform alarm processing in parallel. The task, and set a special scheduling thread execution module A, is used to construct an alarm processing task from the buffer area, and dispatch each alarm processing task to n working threads for processing.

In this embodiment, the predetermined rule is to determine, by using the correspondence between the network element identifier and the number n of the set working threads, which one of the n working threads is to process the alarm information of the network element, and the alarm information belongs to the network. The element is classified, and is specifically classified by the unique identifier of the network element. The foregoing predetermined rule may be: if the remainder of the network element identifier divided by n is represented by i, the advertisement of the network element is set. The police information is fixed in the i worker thread.

As shown in FIG. 5, the alarm batch processing module 402 includes: an access unit 4021, a computing unit 4022, a determining unit 4023, and a thread scheduling processing unit 4024;

The access unit 4021 is configured to: when an idle working thread is included in the preset n working threads, access the first alarm information in the current pending alarm queue; and when the worker thread i is busy, access the current pending alarm queue. The next alarm message;

The calculating unit 4022 is configured to divide the network element identifier carried by the alarm information by n to obtain a remainder i;

The determining unit 4023 is configured to determine whether the worker thread i is idle;

The thread scheduling processing unit 4024 is configured to: when the worker thread i is idle, remove the alarm information of all the remaining alarms in the current alarm queue to form an alarm processing task, and schedule the processing to the i-th worker thread for processing.

Further, the thread scheduling processing unit 4024 is further configured to suspend the predetermined time when none of the preset n working threads are idle.

As shown in FIG. 6, the thread scheduling processing unit 4024 includes:

The batch processing unit 40241 is configured to batch the alarm generation information of all the corresponding network elements before the first alarm disappearing information of each network element in the current alarm processing task; and batch processing the first alarm of each network element in the current alarm processing task The alarm disappearing information of all the corresponding network elements before the information is generated; the deleting subunit 40242 is configured to delete the processed alarm generating information from the current alarm processing task; and delete the processed alarm disappearing information from the current alarm processing task. .

As shown in FIG. 7, an embodiment of the present invention provides a system for improving alarm processing efficiency, including: a slave alarm server 701 and a master alarm server 702;

The alarm server 701 is configured to receive the alarm information that is sent by the network element and carries the identifier of the network element, and performs data conversion and classification processing on the alarm information, and then reports the alarm information to the main alarm server 702. The main alarm server 702 is configured to receive the alarm periodically. Information, according to the alarm information The sequence of the information reporting is sequentially placed in the alarm queue to be processed, and the alarm information in the alarm queue is subjected to parallel thread batch processing according to the network element identifier and the predetermined rule carried in the alarm information.

In this embodiment, the primary alarm server 702 is the server described in the above embodiment.

The embodiment of the present invention collects alarm information by using a predetermined time delay (usually preset to 1 second) by the master-slave alarm server, and then sends the alarm information to the main alarm server in batches, thereby enhancing the batch processing capability of the main alarm server; In the main alarm server, the multi-thread parallel processing is implemented according to the predetermined rule, which greatly improves the parallelism of the alarm processing and improves the processing efficiency of the network management alarm.

While the preferred embodiments of the present invention have been disclosed for purposes of illustration, those skilled in the art will recognize that various modifications, additions and substitutions are possible, and the scope of the invention should not be limited to the embodiments described above.

One of ordinary skill in the art will appreciate that all or a portion of the above steps may be accomplished by a program instructing the associated hardware, such as a read-only memory, a magnetic disk, or an optical disk. Alternatively, all or part of the steps of the above embodiments may also be implemented using one or more integrated circuits. Correspondingly, each module/unit in the above embodiment may be implemented in the form of hardware or in the form of a software function module. The invention is not limited to any specific form of combination of hardware and software.

Industrial applicability

The present invention provides a method, a server, and a system for improving the efficiency of alarm processing. The master-slave alarm server collects alarm information by using a predetermined time delay (usually preset to 1 second), and then sends the alarm information to the primary alarm server in batches. The batch processing capability of the main alarm server is enhanced. At the same time, the multi-thread parallel processing is implemented according to the predetermined rule in the main alarm server, which greatly improves the parallelism of the alarm processing and improves the processing efficiency of the network management alarm.

Claims

Claim

A method for improving the efficiency of alarm processing, comprising:

The alarm information carrying the identifier of the network element is periodically received, and the alarm information is sequentially placed in the alarm queue to be processed according to the sequence of reporting the alarm information;

Performing parallel thread batch processing on the alarm information in the to-be-processed alarm queue according to the network element identifier and the predetermined rule carried in the alarm information.

2. The method according to claim 1, wherein

The step of performing parallel thread batch processing on the alarm information in the to-be-processed alarm queue according to the network element identifier and the predetermined rule carried in the alarm information includes:

Dividing the identifier of the network element carried in the alarm information by n to obtain a remainder i;

Determine whether the worker thread i is idle;

The method according to claim 2, wherein the step of performing parallel thread batch processing on the alarm information in the alarm queue to be processed according to the network element identifier and the predetermined rule carried in the alarm information further includes:

When the preset n working threads are not idle, the step of accessing the first alarm information in the current pending alarm queue when the idle working thread is included in the preset n working threads after the predetermined timeout is suspended .

The method according to claim 2, wherein the network element label carried according to the alarm information The step of performing parallel thread batch processing on the alarm information in the alarm queue to be processed by the identifier and the predetermined rule further includes: the task worker i is processed by the scheduled alarm processing task; the worker thread i is scheduled to be scheduled. The step of processing the alarm processing task by itself includes: batch processing alarm generation information of all corresponding network elements before the first alarm disappearing information of each network element in the current alarm processing task;

The processed alarm generation information is deleted from the current alarm processing task, and the processed alarm generated message is sent into the network management database;

Batching the alarm disappearance information of all corresponding NEs before the first alarm generation information of each network element in the current alarm processing task;

The processed alarm disappearance information is deleted from the current alarm processing task, and the processed alarm disappearance message is sent to the network management database; and all corresponding network elements before the first alarm disappearance information of each network element in the current alarm processing task are returned. The steps of the alarm generating information.

5. A server for improving the efficiency of alarm processing, comprising:

The receiving module is configured to: periodically receive the alarm information carrying the network element identifier, and place the alarm information into the alarm queue to be processed according to the sequence of reporting the alarm information; and the alarm batch processing module, which is set as: Performing parallel thread batch processing on the alarm information in the to-be-processed alarm queue according to the network element identifier and the predetermined rule carried in the alarm information.

The server according to claim 5, wherein the alarm batch processing module comprises: an access unit, configured to: access the current pending alarm queue when there are idle working threads among the preset n working threads The first alarm information; and when the worker thread i is busy, accessing the next alarm information in the current pending alarm queue;

The thread scheduling processing unit is configured to: when the worker thread i is idle, take out all the alarm information of the current pending alarm queue with the remainder i, and form an alarm processing task, scheduling to i No. worker thread for processing.

The server according to claim 6, wherein the thread scheduling processing unit is further configured to: suspend the predetermined time when none of the preset n working threads is idle.

The server according to claim 6 or 7, wherein the thread scheduling processing unit comprises:

The batch processing unit is configured to: batch process alarm generation information of all corresponding network elements before the first alarm disappearing information of each network element in the current alarm processing task; and batch processing the first network element in the current alarm processing task The alarm disappearing information of all corresponding network elements before the alarm generates information;

The sub-unit is deleted, and is set to: delete the processed alarm generation information from the current alarm processing task; and delete the processed alarm disappearance information from the current alarm processing task.

A system for improving the efficiency of alarm processing, comprising: a slave alarm server and a master alarm server; the slave alarm server is configured to: receive, by the network element, alarm information that carries the identifier of the network element, and the alarm The information is periodically reported to the primary alarm server after the data is converted and classified.

The primary alarm server is configured to: periodically receive the alarm information, and put the alarm information into the alarm queue to be processed according to the sequence of reporting the alarm information, according to the identifier and the predetermined rule of the network element carried by the alarm information. Parallel thread batch processing is performed on the alarm information in the alarm queue.

10. The system according to claim 9, wherein the primary alarm server is a claim

The server of any one of 5 to 8.