WO2017197982A1

WO2017197982A1 - Packet processing method, device and system, and computer storage medium

Info

Publication number: WO2017197982A1
Application number: PCT/CN2017/078382
Authority: WO
Inventors: 周佳; 陈丹艳; 钱勇
Original assignee: 中兴通讯股份有限公司
Priority date: 2016-05-16
Filing date: 2017-03-28
Publication date: 2017-11-23
Also published as: CN107395381A

Abstract

Embodiments of the present invention provide a packet processing method, device and system, the method comprising: setting a copy mark in a packet descriptor corresponding to a packet; then executing instructions to process the packet; storing current state information of packet processing according to the copy mark when a specified instruction is executed; and continuing executing the instructions while storing to process the packet. The embodiments of the present invention also provide a computer storage medium.

Description

Message processing method, device and system, and computer storage medium

Technical field

The present invention relates to network switching technologies, and in particular, to a message processing method, apparatus, system, and computer storage medium.

Background technique

With the development of the Internet, the services carried by network devices are more and more complicated, and the requirements for network fault location and service recovery speed are getting higher and higher. How to quickly analyze and solve various problems encountered in network devices It is especially important. Most network devices currently support packet extraction for fault location. The main methods are as follows:

Method 1: An external tester is needed to perform packet extraction and analysis on the faulty service port of the device. However, in many cases, where the network device is running, the capture device such as the tester cannot be provided in time.

Method 2: According to the characteristic values of the packet, such as MAC (Media Access Control), IP (Internet Protocol, protocol inter-network interconnection), UDP_PORT (User Datagram Protocol PORT, user datagram protocol port), etc. Layer 3 information is extracted by setting forwarding entries or configuring ACL (Access Control List) rules. However, these eigenvalues of the packets that cannot be predicted by some abnormal scenarios can only be obtained after all the processes of the packets are processed. The real-time processing of the packets in the network processor cannot be obtained. After the information is reversed, it affects the efficiency of fault location.

Method 3: By adding a breakpoint in the execution program, when a message is processed to a corresponding breakpoint location, all the programs are terminated and the message is extracted, but this method is limited to the test, and cannot be It is applied in the actual live network, otherwise it will cause serious consequences such as business interruption.

Summary of the invention

The embodiments of the present invention provide a packet processing method, device, system, and computer storage medium, and are expected to solve the problem of the external technology of the test device in the failure location of the packet processing, and the low efficiency.

The present invention provides a packet processing method, including:

Setting a copy mark in the message descriptor corresponding to the message, and determining a specified instruction corresponding to the copy mark;

Execute each instruction to process the message;

When the specified instruction is executed, the current state information of the message processing is saved according to the message descriptor corresponding to the message.

The embodiment of the invention further provides a message processing device, including:

a mark setting module configured to set a copy mark in a message descriptor corresponding to the message, and determine a specified instruction corresponding to the copy mark;

An instruction execution module configured to execute each instruction to process the message;

The saving module is configured to save the current state information of the packet processing according to the message descriptor corresponding to the packet when the specified instruction is executed.

The embodiment of the invention further provides a computer storage medium, wherein the computer storage medium stores computer executable instructions, and the computer executable instructions are used to execute the foregoing message processing method.

The present invention provides a message processing method, apparatus, system and computer storage medium, which sets a copy mark in a message descriptor corresponding to a message, and then executes an instruction to process the message, when executing the specified instruction The current state information of the message processing is saved according to the copy mark, which effectively reduces the dependence on other external test equipment and ensures the efficiency of message processing.

DRAWINGS

1 is a flowchart of a packet processing method according to an embodiment of the present invention;

2 is a schematic block diagram of a message processing apparatus according to an embodiment of the present invention;

FIG. 3 is a flowchart of a packet processing method according to an embodiment of the present invention;

FIG. 4 is a schematic block diagram of a message processing system according to an embodiment of the present invention.

detailed description

In the technical solution provided by the embodiment of the present invention, a copy mark is set at a position of a specified instruction of a message, and when the instruction is executed, the current state information of the message processing is cached, and then the current state information of the cache is performed. Analysis and analysis results are obtained, which improves the efficiency of fault location, and also reduces the dependence on other fault location equipment and the impact on the normal execution of the message. The embodiments of the present invention are further described in the following with reference to the accompanying drawings.

First embodiment

This embodiment provides a packet processing method. Referring to FIG. 1, the method includes:

S101. Set a copy mark in a message descriptor corresponding to the message, and determine a specified instruction corresponding to the copy mark;

S102. Execute each instruction to process the packet.

S103. When executing the specified instruction, save the current state information of the packet processing according to the message descriptor corresponding to the message.

The processing of the message is realized by executing the corresponding respective instructions; the processing of the message is the process of executing each instruction. For example, message parsing, forwarding entry lookup, message edit modification, and other instructions are all part of the message processing process. Set in the message descriptor corresponding to the message Copying the mark means that the current state information of the processing of the message needs to be saved, and the timing of saving is to specify the position of an instruction in each executed instruction, and the current state of the processing can be performed when the instruction is executed. The information is saved. Specifying an instruction is arbitrary, can be specified by humans, or can be specified by the characteristics of the instruction. If there is an instruction to extract the message, it is more reasonable to set the copy mark at this instruction. The copy mark and the specified instruction do not necessarily have to be related before setting the copy mark in the message descriptor. However, after the copy mark is set in the message descriptor, the corresponding instruction corresponding to the copy mark is also determined. The relationship may be determined by setting, or may be determined according to the characteristics of the processed message.

The copy tag is set in the message descriptor. The message descriptor is an overview of the content of the message, or a silhouette. The size of the message descriptor is much smaller than the message size, but it can The content of the message is roughly reflected, and information such as the processing logic of the message can be obtained from the message descriptor. The message is processed in the network processor, and the message descriptor is carried when each message is processed in the network processor, and the copy mark in the message descriptor is used as a copy of the message. Logo.

Each instruction in the message can be executed sequentially; the so-called sequential execution is to execute each instruction according to a preset execution logic, and the preset logic can be a system preset or an artificial setting. The number of instructions of a message is often many, and the purpose of executing each instruction is not the same. In the present embodiment, the copy mark is often set at the instruction as the message extraction. It should be noted that the specified message fetching instruction is the instruction logic responsible for the normal processing of the message, and is not specifically used to extract the message. In the process of executing each instruction, the previously executed instruction often affects the executed instruction. On the other hand, there is generally no relationship between the unexecuted instruction and the status of the message. The difference between the preceding and the last is not necessarily the relationship of the order execution that is initially set. In the process of executing the instruction, a jump may occur, that is, the execution in the preset order should be performed later. The instruction may also be a previously executed instruction because of a jump, that is, the order of the previous execution and the subsequent execution depends on the actual execution.

After the specified instruction sets the copy mark, the instruction specified by the copy mark is executed during the execution of the instruction. When the specified instruction is executed, the status of the message at this time is analyzed. In order to analyze the status of the message at this time, the current status information of the message processing needs to be saved. In fact, the status information of the message processing is only related to the executed instruction. The unexecuted command does not affect the status of the message processing. The current status information may include the processing information of the packet header according to the analysis requirements. At least one of the register information and the message buffer may be multiple, and may also have other unmentioned information to serve the state analysis of the message. The processing information of the packet header reflects the change of the message content after the message processing, including the addition, deletion and replacement of the message; the register information is the state information of the register for saving the message, such as the capacity, the access rate and the like. The message cache is the cached content of the packet in the process of processing the packet, which is generally the cached content of the entire packet, and can detect whether the cached device is faulty, whether the cached packet content has a problem or the like.

The copy flag can also be set before executing the specified instruction. The set operation corresponds to the reset operation, and the copy mark is considered to be active if set. When the copy mark is set at the position of the specified instruction of the message, the copy mark can be activated, and if it is not activated, the current status information can be saved. The save in this embodiment can be cached, and the cache refers to the buffer of data exchange. The cache is characterized by fast running speed and much faster than memory, which is more efficient than direct storage. The feature of the cache is that it is only used to cache the status information of the currently specified extraction, that is, the current status information is cached this time. If a new copy mark is specified again, the new status information will be overwritten. Status information cached in the cache.

Since the processing of the current state information is not required for the processing of each packet, further, the processing of the same packet does not necessarily have to save the current state information once. Therefore, the current state of the packet needs to be saved. When the information is saved, then the copy request from the user is received, and then the copy mark is set; at this point, it is considered that the message needs to be performed. Save, and then save the current state information of the message processing when the specified instruction is executed. In addition, when the copy mark is set in the message descriptor, the copy mark can be made to correspond to a plurality of specified instructions; the execution of each execution is performed sequentially, then, before executing to a specified instruction, if it is not required If the current state information is saved, then it is not necessary to set the copy flag. In other words, when the current state information needs to be saved, the copy flag is set. This processing method is flexible and can be achieved very well. Effect.

After saving the current state information, the saved current state information can be obtained, and then the current state information is analyzed; the analysis process is based on the saved message header processing information, register information, message buffer, and the like. The current state information is diagnosed; specifically, the current state information can be presented through the human-machine interaction interface, and the human-computer interaction interface can be obtained by logging in to the corresponding network device through the network management platform telnet (remote terminal protocol). The content of the diagnosis may be whether the current extracted message is abnormal, whether the processing logic has a problem, whether the current state of the register and the storage space is correct, and the like.

In addition, before executing each instruction, each instruction can be compiled, and the serial number corresponding to each instruction is generated after compilation; among all the instructions of the message, the serial number and other instructions corresponding to each instruction are different, so that Directly and quickly locate the specified instruction without error. After compiling each instruction to generate a corresponding serial number, since the copy mark and the specified instruction are corresponding, the corresponding serial number can be determined accordingly.

After determining the serial number of the specified instruction corresponding to the copy mark, the execution of the instruction may be determined according to the serial number, that is, determining whether to execute the specified instruction according to the serial number, thereby further determining whether to save the current status information.

When the current state information of the message processing is saved, the processing of the message, that is, the execution of the instruction is still continuing, and the current state information of the saved message processing does not affect the execution of the remaining instructions, and the subsequent message. The processing and the saving of the current state information may be performed synchronously, so that the efficiency of message processing can be ensured to the greatest extent, and the purpose of fault detection can be achieved at the same time.

The micro-engine in the network processor is a direct unit for processing the message. The processing of each instruction of the message is usually performed in the micro-engine; the micro-engine is responsible for modifying and editing various messages, and forwarding the service entry. Access and processing operations, the processing of the message is completed through a series of instruction operations.

In this embodiment, by setting a copy mark in the position of the specified instruction in the currently processed message, when the instruction is executed, the current state information of the message is cached, and the fault location personnel can according to the cached message state information. The analysis is performed to obtain the analysis result, so that the fault locating personnel can analyze the processing status of the message. Obviously, in this embodiment, the fault location efficiency can be effectively improved without adding additional equipment, not knowing the packet feature value, and not affecting normal service forwarding.

Second embodiment

This embodiment provides a packet processing apparatus. Referring to FIG. 2, the method includes:

The flag setting module 11 is configured to set a copy mark in the message descriptor corresponding to the message, and determine a specified instruction corresponding to the copy mark;

The instruction execution module 12 is configured to execute each instruction to process the message;

The saving module 13 is configured to save the current state information processed according to the message descriptor message corresponding to the message when the specified instruction is executed.

The processing of the message is realized by executing the corresponding respective instructions; the processing of the message is the process of executing each instruction. For example, message parsing, forwarding entry lookup, message edit modification, and other instructions are all part of the message processing process. Setting the copy mark in the message descriptor corresponding to the message means that the current state information of the processing of the message needs to be saved, and the timing of saving is to specify the position of an instruction in each executed instruction, and execute When the instruction is reached, the current status information of the processing can be saved. Specifying an instruction is arbitrary. It can be specified by human or by the characteristics of the instruction. If there is an instruction to extract the message, it is more reasonable to set the copy mark at this instruction. The copy mark and the specified instruction do not necessarily have to be related before setting the copy mark in the message descriptor. However, after the copy mark is set in the message descriptor, the corresponding instruction corresponding to the copy mark is also determined. The relationship may be determined by setting, or may be determined according to the characteristics of the processed message.

The copy tag is set in the message descriptor. The message descriptor is an overview of the content of the message, or a silhouette. The size of the message descriptor is much smaller than the message, but it can be roughly reflected. The contents of the message, including the processing logic of the message, can be obtained from the message descriptor. The message is processed in the network processor, and the message descriptor is carried when each message is processed in the network processor, and the copy mark in the message descriptor is used as a copy of the message. Logo.

After the specified instruction sets the copy mark, the instruction specified by the copy mark is executed during the execution of the instruction. When the specified instruction is executed, it indicates that the status of the message is to be analyzed at this time. In order to analyze the status of the message at this time, the saving module 13 needs to process the message. The current status information is saved. In fact, the status information of the message processing is only related to the executed instruction. The unexecuted command does not affect the status of the message processing. The current status information may include the processing information of the packet header according to the analysis requirements. At least one of the register information and the message buffer may be multiple, and may also have other unmentioned information to serve the state analysis of the message. The processing information of the packet header reflects the change of the message content after the message processing, including the addition, deletion and replacement of the message; the register information is the state information of the register for saving the message, such as the capacity, the access rate and the like. The message cache is the cached content of the packet in the process of processing the packet, which is generally the cached content of the entire packet, and can detect whether the cached device is faulty, whether the cached packet content has a problem or the like.

Optionally, the setting module 15 is further configured to: set the copy flag before saving the current state information of the message processing. The set operation corresponds to the reset operation, and the copy mark is considered to be active if set. When the copy mark is set at the position of the specified instruction of the message, the copy mark can be activated, and if it is not activated, the current status information can be saved. The save in this embodiment can be cached, and the cache refers to the buffer of data exchange. The cache is characterized by fast running speed and much faster than memory, which is more efficient than direct storage. The feature of the cache is that it is only used to cache the status information of the currently specified extraction, that is, the current status information is cached this time. If a new copy mark is specified again, the new status information will be overwritten. Status information cached in the cache.

Since the processing of the current state information is not required for the processing of each packet, further, the processing of the same packet does not necessarily have to save the current state information once. Therefore, the current state of the packet needs to be saved. When the information is saved, then the request receiving module 14 is further configured to receive the copy request from the user and then set the copy flag; at this time, it is considered that the message needs to be saved, and then the specified instruction is executed. The current status information of the message processing is saved. In addition, when the copy mark is set in the message descriptor, the copy mark can be made to correspond to a plurality of specified instructions; the execution of each execution is performed sequentially, then, If you do not need to save the current state information for this time before executing a specified instruction, you do not need to set the copy flag. In other words, when you need to save the current state information, you can copy the flag. With the set, the processing method is flexible and can achieve good results.

After the current state information is saved, the saved current state information can be obtained, and then the current state information is analyzed to obtain an analysis result. The analysis process is based on the current state information of the saved message header processing information, register information, message buffer, etc.; specifically, the current state information can be presented through the human-computer interaction interface, and the human-computer interaction interface can pass The network management platform telnet is logged in to the corresponding network device. The content of the diagnosis may be whether the current extracted message is abnormal, whether the processing logic has a problem, whether the current state of the register and the storage space is correct, and the like.

In addition, the compiling module 16 may be further configured to compile each instruction before executing each instruction, and generate a serial number corresponding to each instruction after compiling; among all the instructions of the message, the serial number corresponding to each instruction and Other instructions are different, so you can quickly and directly locate the specified instruction without error. Optionally, the sequence determining module 17 is further configured to: after compiling each instruction to generate a corresponding serial number, determine a serial number corresponding to the specified instruction, and determine whether to execute the specified instruction according to the serial number; After the serial number of the corresponding specified instruction, the execution of the instruction may be determined according to the serial number, that is, whether the specified instruction is executed according to the serial number, thereby further determining whether to save the current status information.

When the current state information of the message processing is saved, the processing of the message, that is, the execution of the instruction is still continuing, that is, the instruction execution module 12 is further configured to continue while saving the current state information of the message processing. Executing instructions to process the message; saving the current state information of the message processing does not affect the execution of the remaining instructions, and the subsequent processing of the message and the saving of the current state information may be performed synchronously, so that the maximum degree is The efficiency of packet processing is guaranteed, and the purpose of fault detection can be achieved at the same time.

Third embodiment

This embodiment provides a packet processing method. Referring to FIG. 3, the method includes:

Step S201: The message enters the network processor for processing.

It should be noted that the processing of the message in the network processor is completed through a series of instruction operations, such as message parsing, forwarding entry search, message editing and modification, etc., and these instructions are in the micro-engine of the network processor. The middle is executed sequentially, and the instructions are compiled by the compiler to generate an internal serial number, and the serial number of each instruction is unique.

Step S202: Setting a copy mark in the message descriptor may include: finding a position of the message extraction instruction at the compiled serial number according to the location of the message extraction instruction specified by the user, and reporting the message in the message descriptor The text copy flag is set.

Specifically, each instruction is compiled by the compiler to generate a unique serial number. When a processing instruction that needs to extract a message is specified, the instruction is usually a message extraction instruction, and the corresponding serial number is determined by the instruction; When the message is executed to the instruction, it is executed to the corresponding When the serial number is set, the copy flag in the message descriptor corresponding to the message is set. Set is understood to be active, ie the copy mark in the message descriptor is rendered active, corresponding to the reset.

Step S203: determining whether the copy flag in the message descriptor is set, if set, copying and saving the current state information of the message processing in a physical cache for subsequent analysis and processing, and the original message is still Continue processing normally in the microengine; if not set, only the original message is processed.

It should be noted that the copying and saving of the current state information of the message processing, such as the processing information of the packet header, the state of the register, and the copying and saving of the message buffering information, do not affect the forwarding of the original packet, that is, the original message will not be The business process has any influence. When copying and saving the current state information of the message processing, the processing of the message, that is, the execution of the instruction continues, and the two do not affect each other.

Step S204: Copying and saving current state information may include: when fault location and analysis are needed, obtaining current state information of the copy and save from the physical cache, and presenting the current state information through the human-machine interaction interface.

Specifically, the physical cache is usually a small piece of special storage space allocated in memory on the micro engine, and is only used to cache current state information of the currently specified extraction. If a new message extraction location is specified again, the new current state information will be overwritten with the current state information stored in the previous cache. The human-computer interaction interface can be obtained by remotely logging in to the corresponding network device through the network management platform telnet.

Step S205: Acquire saved current state information.

Step S206: analyzing the current state information, including: the fault locating personnel performs diagnosis according to the obtained packet processing information, determines whether the currently extracted packet is abnormal, whether the microengine processing logic has a problem, whether the current state of the register and the storage space is correct. Wait.

Fourth embodiment

This embodiment provides a packet processing system. Referring to FIG. 4, the method includes:

The analyzing device 20 and the message processing device 10; wherein the analyzing device 20 is configured to analyze the current state information held by the message processing device 10 and obtain an analysis result; the message processing device 10 is similar to the above embodiment , no longer repeat them here.

Fifth embodiment:

The embodiment of the present invention further provides a computer storage medium, where the computer storage medium stores computer executable instructions, and the computer executable instructions are used to execute the message processing method provided by any one of the foregoing embodiments; 1. At least one of the methods shown in Figures 2 and 4.

The computer storage medium may be various types of storage media, specifically: an optical disk, a magnetic tape, a hard disk, a random storage medium, a read-only storage medium, or a flash memory; and the like may be a non-transitory storage medium.

The above is a further detailed description of the present invention in connection with the specific embodiments, and the specific embodiments of the present invention are not limited to the description. Modifications made in accordance with the principles of the invention are understood to fall within the scope of the invention.

Claims

A packet processing method includes:

Setting a copy mark in the message descriptor corresponding to the message, and determining a specified instruction corresponding to the copy mark;

Execute each instruction to process the message;

When the specified instruction is executed, the current state information of the message processing is saved according to the message descriptor corresponding to the message.
The packet processing method according to claim 1, wherein when the current state information of the packet processing is saved, the method further includes: continuing to execute the instruction, and processing the packet.
The message processing method according to claim 1, further comprising: setting the copy flag before the executing to the specified instruction.
The message processing method according to claim 3, wherein before the setting of the copy flag, the method further comprises: receiving a copy request from the user.
The message processing method according to any one of claims 1 to 4, further comprising: compiling each instruction to generate a sequence number corresponding to each instruction before executing each instruction.
The message processing method according to claim 5, wherein the determining the specified instruction corresponding to the copy mark further comprises: determining a sequence number corresponding to the specified instruction; and the executing to the specified instruction comprises: The serial number determines whether to execute the specified instruction.
The message processing method according to any one of claims 1 to 4, further comprising: analyzing the saved current state information after the saving of the current state information of the message processing, Analysis results.
A message processing device, comprising:

a mark setting module configured to set a copy mark in a message descriptor corresponding to the message, and determine a specified instruction corresponding to the copy mark;

An instruction execution module configured to execute each instruction to process the message;

The saving module is configured to save the current state information of the packet processing according to the message descriptor corresponding to the packet when the specified instruction is executed.
The message processing device according to claim 8, wherein the instruction execution module is further configured to: when the current state information processed by the message is saved, continue to execute an instruction to process the message;
The message processing apparatus of claim 8 further comprising a set module configured to set said copy flag prior to said executing to a specified instruction.
The message processing apparatus according to claim 10, further comprising a request receiving module configured to receive a copy request from the user before said copying flag is set.
A message processing apparatus according to any of claims 8-11, further comprising a compiling module configured to compile each instruction prior to executing each instruction to generate a sequence number corresponding to each instruction.
The message processing apparatus according to claim 12, further comprising a sequence determining module configured to determine a sequence number corresponding to said specified instruction, and to determine whether to execute said specified instruction based on said serial number.
A message processing system, comprising: an analysis device and a message processing device according to any one of claims 8 to 13; the analysis device configured to analyze the saved current state information and obtain Analysis results.
A computer storage medium having stored therein computer executable instructions for performing the message processing method of any one of claims 1 to 7.