WO2016206520A1

WO2016206520A1 - Method and apparatus for implementing flow table traversal service

Info

Publication number: WO2016206520A1
Application number: PCT/CN2016/083703
Authority: WO
Inventors: 路鹏
Original assignee: 中兴通讯股份有限公司
Priority date: 2015-06-26
Filing date: 2016-05-27
Publication date: 2016-12-29
Also published as: CN106330694A

Abstract

Provided is a method for implementing a flow table traversal service. The method comprises: a control module acquires a service parameter of a flow table traversal service, and delivers the service parameter to a coprocessor module; and the coprocessor module controls network processors in the coprocessor module to process the flow table traversal service according to the service parameter.

Description

Method and device for implementing flow table traversal service

Technical field

This document relates to, but is not limited to, the field of data communication technology, and in particular, to a method and device for implementing a flow table traversal service.

Background technique

With the explosive development of the Internet, network traffic has increased dramatically, and various emerging services have emerged in an endless stream. Data communication devices that describe the context of each data packet by establishing a flow table will establish a large number of session entries, usually in the order of ten megabytes or more. How to manage this large number of flow meters efficiently becomes a problem that must be solved for communication equipment.

In the traditional solution, the storage device that directly accesses the storage flow table through the CPU has lower efficiency when there are more entries due to CPU processing capability and PCI (Peripheral Component Interconnect) bus speed.

Therefore, how to provide an implementation method of a flow table traversal service that can improve efficiency is a technical problem to be solved by those skilled in the art.

Summary of the invention

The following is an overview of the topics detailed in this document. This Summary is not intended to limit the scope of the claims.

This paper provides a method and device for implementing a flow table traversal service to improve the processing efficiency of the flow table traversal service.

The method for implementing the flow table traversal service is provided, the method includes: the control module obtains the service parameters of the flow table traversal service, and sends the service parameters to the co-processing module; the co-processing module controls each network processor to process the flow table according to the service parameters. Traverse the business.

Optionally, the method further includes: the control module acquiring a processing result of the coprocessing module.

Optionally, the processing result of the control module acquiring the coprocessing module includes: the coprocessing module sends the processing result to the storage module, and the control module obtains the processing result from the storage module.

Optionally, the method further includes: the control module performs statistics according to the processing result, and displays the statistical result to the user.

Optionally, the coprocessing module controls each network processor to process the flow table according to the service parameter, and the traversal service includes: the scheduling core in the network processor schedules the flow table to each service core according to the service parameter, and the service core processes the scheduled Flow table.

The present invention provides a device for implementing a flow table traversal service, comprising: a control module, configured to acquire a service parameter of a flow table traversal service, and send the service parameter to a co-processing module; the co-processing module is configured to control each network processor therein The flow table is traversed according to the business parameters.

Optionally, the control module is further configured to obtain a processing result of the coprocessing module.

Optionally, the device further includes a storage module;

The coprocessing module is further configured to send the processing result to the storage module;

The control module is further configured to obtain a processing result from the storage module.

Optionally, the control module is further configured to perform statistics according to the processing result, and display the statistical result to the user.

Optionally, the coprocessing module is configured to control, in the following manner, each network processor to process the flow table according to the service parameter traversal service: the scheduling core in the network processor schedules the flow table to each service core according to the service parameter. The business core processes its scheduled flow table.

The embodiment of the invention further provides a computer readable storage medium storing computer executable instructions, which are implemented when executed by a processor.

The embodiment of the invention provides a method for implementing a flow table traversal service, which uses a network processor to perform a flow table traversal service, and various algorithms can be implemented by hardware, and implement complex congestion management, queue scheduling, flow classification, and QoS functions. Under the premise, it can also achieve extremely high search and forwarding performance and achieve "hard forwarding". Compared with pure hardware chips, the network processor fully supports programming, and the programming mode is simple. Once new technologies or requirements appear, it is convenient. The ground is implemented by microcode programming; in addition, the network processor is also scalable, and multiple network processors can be interconnected to form a network processor cluster to support larger and high-speed network processing, which can effectively solve A less efficient problem exists with related art solutions.

Other aspects will be apparent upon reading and understanding the drawings and detailed description.

BRIEF abstract

FIG. 1 is a schematic structural diagram of an apparatus for implementing a flow table traversal service according to a first embodiment of the present invention.

FIG. 2 is a flowchart of a method for implementing a flow table traversal service according to a second embodiment of the present invention.

FIG. 3 is a flowchart of a method for implementing a flow table traversal service according to a third embodiment of the present invention.

Embodiments of the invention

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be noted that, in the case of no conflict, the features in the embodiments and the embodiments in the present application may be arbitrarily combined with each other.

The invention will now be further illustrated by way of specific embodiments in conjunction with the accompanying drawings.

First embodiment:

1 is a schematic structural diagram of an implementation apparatus according to a first embodiment of the present invention. As shown in FIG. 1, in the embodiment, the apparatus 1 for implementing a flow table traversal service includes:

The control module 11 is configured to acquire the service parameters of the flow table traversal service, and send it to the co-processing module 12;

The co-processing module 12 is configured to control each of the network processors to process the flow table to traverse the service according to the service parameters.

In some embodiments, the control module 11 in the above embodiment is further configured to obtain the processing result of the coprocessing module 12.

In some embodiments, as shown in FIG. 1, the implementation apparatus 1 in the above embodiment further includes a storage module 13 that is further configured to send the processing result to the storage module 13, and the control module 11 is further configured to be stored. The module gets 13 processing results.

In some embodiments, the control module 11 in the above embodiment is further configured to perform statistics according to the processing result and display the statistical result to the user.

In some embodiments, the co-processing module 12 in the above embodiment is configured to control each of the network processors in the following manner to process the flow table according to the service parameter traversal service: the scheduling core in the network processor will flow the table according to the service parameters. It is scheduled to each service core, and the service core processes its scheduled flow table.

Second embodiment:

2 is a flowchart of an implementation method according to a second embodiment of the present invention. As shown in FIG. 2, in this embodiment, a method for implementing a flow table traversal service includes the following steps:

S201: The control module obtains the service parameter of the flow table traversal service, and sends the service parameter to the co-processing module;

S202: The co-processing module controls each network processor to process the flow table according to the service parameter to traverse the service.

In some embodiments, the above embodiment further includes: the control module acquiring a processing result of the co-processing module.

In some embodiments, the obtaining, by the control module in the foregoing embodiment, the processing result of the co-processing module includes: the co-processing module sends the processing result to the storage module, and the control module acquires the processing result from the storage module.

In some embodiments, the above embodiment further includes: the control module performs statistics according to the processing result, and displays the statistical result to the user.

In some embodiments, the coprocessing module in the foregoing embodiment controls each network processor to process the flow table according to the service parameter to traverse the service, and the scheduling core in the network processor schedules the flow table to each service core according to the service parameter. The business core processes its scheduled flow table.

The present invention will be further explained in conjunction with specific application examples.

Third embodiment:

FIG. 3 is a flowchart of an implementation method according to a third embodiment of the present invention. As shown in FIG. 3, in the embodiment, the method for implementing a flow table traversal service includes the following steps:

S301: The control module obtains the service parameter and sends it to the coprocessing module.

The user sends a service request, and the control module receives the service request sent by the user, and sends the obtained parameter to the underlying co-processing module in the form of a message, and simultaneously listens to the processing process of the co-processing module.

S302: The coprocessing module processes the flow table traversal service.

The two network processors of the co-processing module respectively receive the statistics messages sent by the control module, parse the parameters and perform corresponding service processing according to the user requirements. Here, the flow table service processing is directly traversed with respect to the traditional CPU. Multi-core parallel processing, one core is responsible for command scheduling, other cores are responsible for traversing the flow table to handle the corresponding business, rather than one kernel traversing all flow tables, but the entire kernel is divided into other business processing kernel processing to improve efficiency, Finally, the result is either written to the peripheral storage device or sent directly to the control module in the form of a message.

In the present application, the flow table traversal process is placed in the co-processing module processing, and the total flow table is set to M. The design uses multiple cores to process the entire flow table traversal process in parallel, one of which is used for command scheduling, and the rest is left. The L cores are used for specific service processing. When the core of the command scheduling receives the service message sent by the upper layer control module, it is initialized first, including setting a global control variable index_cnt. The value of index_cnt can be used to know the service processing. schedule. When index_cnt is equal to L, the other cores are notified to start traversing the flow table. At the same time, to improve efficiency, the core for doing business processing is not a core traversing all the flow tables, but dividing the entire flow table into L blocks, where the average is It is not an absolute average, but an approximate average. The principle is that there is no overlap between the flow table blocks responsible for each core, and all flow tables are covered at the same time. In addition, the business volume of each business core processing is as large as possible to achieve the most efficient overall efficiency. Excellent purpose. As shown in FIG. 3, it is assumed that there are M flow tables, a=M/L, so that the i-th service core can be responsible for the flow table service in the range of (i-1)*a+1 to i*a. The L service cores process the flow table blocks that are respectively responsible for each other. When each service core finishes the flow table block that is responsible for each, the global control variable index_cnt is decremented by one, and when index_cnt is reduced to 0, it represents the entire service processing. After completion, the scheduling core integrates the processing results of all service cores, and directly reports the final result to the control module or to the peripheral storage device. If it is written to the peripheral storage device, it also needs to notify the upper control module service. Processing has been completed.

S303: The control module counts the processing result.

When the control module monitors that the co-processing module has been statistically completed, the results of the two network processors are respectively obtained for re-processing, and the final result is fed back to the upper-layer user. The specific TOP-N statistics are For example, the TOP-N statistic is based on a more complicated embodiment of the flow table traversal described above, and can quickly count traffic ranking information based on ip addresses, protocols, etc., and display the top 10 and top 20, respectively. Top 50.

The TOP-N statistics include the following steps:

The user sends the flow table service to the control module through command line parameter configuration.

When the control module receives the user statistics request, the relevant parameters are parsed, and then the parameters are configured into messages, which are respectively sent to the two network processors NP of the coprocessing module in the form of messages.

The TOP-N statistics message sent by the upper layer is obtained by the core for command scheduling in each NP, and the initialization process is performed. The global variable index_cnt is set to L (the number of service cores), and the entire block flow table is divided into L shares. The service core starts to traverse the flow table blocks that are responsible for each. A hash list is created based on each flow table and the delivered parameters, and all flow tables are counted based on the delivered parameters. Before creating a HASH table, the HASH table specification is specified. For example, the M-th square bar of 2, the corresponding flow table information in the flow table is taken out as a keyword according to the parameters obtained from the message, and the HASH value is obtained by HASH operation, and then, The 0-(M-1) bit of the value H is used as the INDEX table index IDX corresponding to the HASH entry. If there is no matching INDEX entry, a new INDEX table and the corresponding ENTRY table need to be created. If there is a matching INDEX table, the corresponding ENTRY table is mapped by the value in the INDEX table. Here, since there may be a HASH conflict, after the ENTRY table is fetched, the keyword is also searched for an exact match to confirm whether it is actually indexed. The entry, if it is, accumulates the statistics in the table. If not, you need to create a new ENTRY table and save its index value in the previous ENTRY table that conflicts with its hash to establish contact. After each service core processes its own responsible business, it decrements index_cnt by 1. When the core responsible for command scheduling monitors that index_cnt is 0, it represents the completion of the business core statistics service. Because the HASH table is relatively large after the statistics are completed, the statistics cannot be counted. The result is immediately fed back to the upper control module, but all the HASH tables are saved in the peripheral storage device, and the underlying statistical service completion is notified to the upper control module.

When the control module monitors that two NP statistics are completed, the peripheral storage devices that are suspended from the NP are directly read from the respective HASH statistical linked list. Because the statistical results of the two network processors are integrated into one, the HASH chain established by one of the network processors in the peripheral storage device can be read once in a single block, and another network processing is traversed on this basis. The HASH statistics chain of the device, through the same HASH establishment method as the bottom layer or cumulative or new, will eventually be the two network processors The statistical results are integrated into a HASH chain and stored in a dynamically requested memory. Finally, the HASH chain is traversed, and the statistical result of the previous TOP-N is discharged by the binary sorting method to the upper layer user.

In addition, an embodiment of the present invention further provides a computer readable storage medium storing computer executable instructions, which are implemented when executed by a processor.

In summary, the above embodiments of the present invention have at least the following beneficial effects:

The network processor is used to perform the flow table traversal service, and various algorithms can be implemented by hardware. Under the premise of implementing complex congestion management, queue scheduling, flow classification, and QoS functions, the algorithm can also achieve extremely high search and forwarding performance. Hard forwarding"; compared to pure hardware chips, the network processor fully supports programming, the programming mode is simple, once new technologies or requirements appear, it can be easily implemented by microcode programming; in addition, the network processor also has Scalability, multiple network processors can be interconnected to form a network processor cluster to support larger and high-speed network processing, which can improve the processing efficiency of the flow table traversal service.

One of ordinary skill in the art will appreciate that all or a portion of the above steps may be performed by a program to instruct related hardware, such as a processor, which may be stored in a computer readable storage medium, such as a read only memory, disk or optical disk. Wait. 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, for example, by implementing an integrated circuit to implement its corresponding function, or may be implemented in the form of a software function module, for example, executing a program stored in the memory by a processor. Instructions to achieve their corresponding functions. This application is not limited to any specific combination of hardware and software.

It should be noted that various other embodiments and modifications may be made by those skilled in the art without departing from the spirit and scope of the application, Corresponding changes and modifications are intended to fall within the scope of the appended claims.

Industrial applicability

The technical solution provided by the embodiment of the present invention uses a network processor to perform a flow table traversal service, and various algorithms can be implemented by hardware. Under the premise of implementing complex congestion management, queue scheduling, flow classification, and QoS functions, High search and forwarding performance, achieving "hard forwarding"; compared to pure hardware chips, the network processor fully supports programming, and the programming mode is simple. Once new technologies or requirements appear, it can be easily implemented by microcode programming. In addition, the network processor is also scalable, and multiple network processors can be interconnected to form a network processor cluster to support larger and high-speed network processing, which can improve the processing efficiency of the flow table traversal service.

Claims

A method for implementing a flow table traversal service, comprising:

The control module obtains the service parameters of the flow table traversal service, and sends the service parameters to the co-processing module;

The co-processing module controls each of the network processors to process the flow table traversal service according to the service parameters.
The implementation method of claim 1 wherein:

The method further includes: the control module acquiring a processing result of the co-processing module.
The implementation method of claim 2 wherein:

The control module acquires processing results of the coprocessing module, including:

The coprocessing module sends the processing result to a storage module, and the control module acquires the processing result from the storage module.
The implementation method of claim 2 wherein:

The method further includes: the control module performing statistics according to the processing result, and displaying the statistical result to the user.
The implementation method according to any one of claims 1 to 4, wherein:

The coprocessing module controls each of the network processors to process the flow table traversal service according to the service parameter, including:

The scheduling core in the network processor schedules the flow table to each service core according to the service parameter, and the service core processes the flow table for its scheduling.
An implementation device for a flow table traversal service, comprising:

The control module is configured to obtain a service parameter of the flow table traversal service, and send the service parameter to the co-processing module;

The co-processing module is configured to control each network processor therein to process the flow table traversal service according to the service parameter.
The implementation device of claim 6 wherein:

The control module is further configured to obtain a processing result of the coprocessing module.
The implementation device of claim 7 wherein:

The implementation device further includes a storage module;

The coprocessing module is further configured to send the processing result to the storage module;

The control module is further configured to acquire the processing result from the storage module.
The implementation device of claim 7 wherein:

The control module is further configured to perform statistics according to the processing result and display the statistical result to the user.
The implementation device according to any one of claims 6 to 9, wherein:

The coprocessing module is configured to control, within the following manner, each network processor to process the flow table according to the service parameter traversal service: the scheduling core in the network processor schedules the flow table to each service core according to the service parameter, The business core processes its scheduled flow table.