CN117499258A - Service data network transmission management method and DPU - Google Patents

Abstract

The application provides a service data network transmission management method and a DPU, wherein the method is executed in the DPU and comprises the following steps: according to the service type of each data packet to be transmitted by the network, searching the QoS strategy data of each data packet in the comparison relation data between the current QoS strategy and the service type, and outputting each data packet to the target network for transmission based on the QoS strategy data of each data packet; and monitoring the target network and each data packet in real time, and judging whether to dynamically adjust QoS strategy data of the data packet according to the corresponding monitoring result. The method and the device can realize accurate identification of different service flows and QoS policy allocation so as to provide high-quality network service, improve the flexibility of service data network transmission management, and be suitable for various and high-variability service flow transmission scenes.

Description

Service data network transmission management method and DPU

Technical Field

The present disclosure relates to the field of traffic management processing technologies, and in particular, to a service data network transmission management method and a DPU.

Background

In network communications, quality of service QoS (Quality of Service) refers to classifying and managing different types of network traffic to ensure reliability of network services, reasonable utilization of bandwidth, and stability of quality. Currently, traffic management in a data center is an important task. To ensure quality of network service (QoS), data centers need to ensure priority and bandwidth for different traffic flows.

Currently, CPU-based traffic management schemes typically use static QoS policies to manage traffic. However, CPU-based traffic management schemes cannot effectively and accurately distinguish and optimize various traffic flows. And when processing highly variable and diverse traffic flows, the static QoS policy lacks flexibility and adaptability, and cannot adapt to the changes of different traffic types and network conditions quickly, which may cause the problems of performance degradation and QoS non-assurance. In addition, when the conventional flow management scheme based on the CPU processes high-speed flow of a large-scale data center, the calculation and processing capacity of the CPU are limited, so that the real-time processing requirement on a large number of data packets cannot be met, and delay increase and throughput decrease are caused.

That is, the CPU-based traffic management scheme has the disadvantages of insufficient flexibility, inability to accommodate diverse traffic flows, inability to handle highly variable traffic and performance bottlenecks using static QoS policies. These drawbacks limit the precise management and optimization of different traffic types and network conditions, resulting in performance and QoS instability.

Disclosure of Invention

In view of this, embodiments of the present application provide a traffic data network transmission management method and a DPU to obviate or mitigate one or more disadvantages in the prior art.

One aspect of the present application provides a service data network transmission management method, performed in a DPU, the service data network transmission management method including:

according to the service type of each data packet to be transmitted by the network, searching the QoS strategy data of each data packet in the comparison relation data between the current QoS strategy and the service type, and outputting each data packet to a target network for transmission based on the QoS strategy data of each data packet;

and monitoring the real-time state of the target network and each data packet, and judging whether to dynamically adjust QoS strategy data of the data packet according to a corresponding monitoring result.

In some embodiments of the present application, before searching for QoS policy data of each data packet in the comparison relationship data between the current QoS policy and the service type according to the service type to which each data packet to be transmitted by the network belongs, the method further includes:

receiving a data packet to be transmitted by a network;

analyzing the data packet to obtain header information of the data packet;

based on a preset service type identification mode, identifying and obtaining the service type of the data packet according to the head information of the data packet.

In some embodiments of the present application, the service type identification manner includes: feature matching, machine learning or deep packet inspection;

correspondingly, the identifying, based on the preset service type identifying manner, the service type to which the data packet belongs according to the header information of the data packet includes:

if the service type identification mode is the characteristic matching, extracting service characteristics from the header information of the data packet, and searching the service type of the data packet in comparison relation data between the preset service characteristics and the service types based on the service characteristics;

If the service type identification mode is the machine learning mode, inputting the head information of the data packet into a preset machine learning model for identifying the service type, so that the service type identification model correspondingly outputs the service type to which the data packet belongs;

and if the service type identification mode is the deep data packet detection, performing the deep data packet detection on the head information of the data packet to identify the service type to which the data packet belongs.

In some embodiments of the present application, before searching for QoS policy data of each data packet in the comparison relationship data between the current QoS policy and the service type according to the service type to which each data packet to be transmitted by the network belongs, the method further includes:

receiving control relation data between a preset QoS strategy and a service type and storing the control relation data to a local;

correspondingly, the service data network transmission management method further comprises the following steps:

and self-adaptively adjusting the control relation data between the QoS strategy and the service type according to the network state of the target network, and/or adjusting the control relation data between the QoS strategy and the service type based on the user-defined QoS strategy data if the user-defined QoS strategy data is received.

In some embodiments of the present application, the QoS policy data includes: priority and bandwidth allocation weights;

correspondingly, according to the service type to which each data packet to be transmitted by the network belongs, respectively searching the QoS policy data of each data packet in the comparison relationship data between the current QoS policy and the service type, and respectively outputting each data packet to a target network for transmission based on the QoS policy data of each data packet, including:

according to the service type of each data packet to be transmitted by the network, searching the corresponding priority and bandwidth allocation weight of each data packet in the comparison relation data between the current QoS strategy and the service type;

sequencing the data packets according to the order of the priority from high to low;

the current bandwidth resources of the target network are respectively distributed to the data packets according to the bandwidth distribution weights of the data packets;

and outputting the sequenced data packets to the target network in sequence for transmission based on the bandwidth resources corresponding to the data packets.

In some embodiments of the present application, the priority includes: a high priority, a medium priority and a low priority which are divided from high to low, wherein the medium priority comprises a plurality of medium priorities which are divided from high to low, and each medium priority comprises a plurality of sub-priorities which are divided from high to low;

correspondingly, the sorting the data packets according to the order of the priority from high to low includes:

and respectively placing the data packets into the queues corresponding to the high priority, the medium priority and the low priority according to the priorities corresponding to the data packets, wherein the data packets placed into the queues corresponding to the medium priority are ordered according to the order of the sub-priorities from high to low.

In some embodiments of the present application, the allocating, according to the bandwidth allocation weights of the data packets, current bandwidth resources of the target network to the data packets includes:

acquiring the current total bandwidth resource of the target network;

and dividing the percentages of the total bandwidth resources according to the respective bandwidth allocation weights of the data packets, and allocating the current bandwidth resources of the target network to the data packets based on the corresponding percentage division results.

In some embodiments of the present application, the outputting, based on the bandwidth resources corresponding to each of the data packets, the ordered data packets to the target network for transmission includes:

and based on the bandwidth resources corresponding to the data packets, sequentially outputting the data packets in the queues corresponding to the high priority, the medium priority and the low priority to the target network in a first-in first-out mode for transmission.

In some embodiments of the present application, the performing real-time status monitoring on the target network and each data packet, and determining whether to dynamically adjust QoS policy data of the data packet according to a corresponding monitoring result includes:

the target network and each data packet are monitored in real time, and whether the data packet in any queue exceeds a preset quantity threshold value, whether the current network state of the target network meets preset congestion conditions or whether broadband resources of any data packet are smaller than a preset resource threshold value corresponding to the data packet or not is judged according to a corresponding monitoring result;

if the data packets in any queue exceed the preset quantity threshold, extracting the data packets from the queue, and modifying the priority corresponding to the extracted data packets so that the data packets with the modified priorities are added into the corresponding queues again;

If the current network state of the target network meets a preset congestion condition, dynamically adjusting the priority and broadband allocation weight of the data packet which is not output to the target network, and/or sending a network congestion prompt message to a sending end of the data packet which is not output to the target network;

and if any broadband resource of the data packet is smaller than the preset resource threshold corresponding to the data packet, dynamically adjusting the broadband allocation weight of the data packet so as to reallocate the broadband resource of the data packet.

Another aspect of the present application provides a DPU, in which a service data network transmission management module is disposed;

the service data network transmission management module is used for the service data network transmission management method;

the service data network transmission management module is in communication connection with the data center so as to acquire a data packet currently sent to the data center from the data center.

The service data network transmission management method is executed in a DPU, according to the service type of each data packet to be transmitted in a network, respective QoS strategy data of each data packet is searched in comparison relation data between a current QoS strategy and the service type, and each data packet is output to a target network for transmission based on the respective QoS strategy data of each data packet; and monitoring the real-time state of the target network and each data packet, and judging whether to dynamically adjust QoS strategy data of the data packet according to a corresponding monitoring result. By introducing the DPU, the QoS capability of the data center can be improved, the accurate identification of different service flows and QoS strategy allocation can be realized, and high-quality network service can be provided; the target network and each data packet are used for carrying out real-time state monitoring and dynamic QoS strategy adjustment, so that the flexibility of service data network transmission management can be improved, and the method is applicable to various and highly-changeable service flow transmission scenes; compared with the traditional scheme based on the CPU, the method has the advantages of obvious high performance, low delay, high parallel processing capacity, flexibility, programmability and the like.

Additional advantages, objects, and features of the application will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the application. The objectives and other advantages of the application may be realized and attained by the structure particularly pointed out in the written description and drawings.

It will be appreciated by those skilled in the art that the objects and advantages that can be achieved with the present application are not limited to the above-detailed description, and that the above and other objects that can be achieved with the present application will be more clearly understood from the following detailed description.

Drawings

The accompanying drawings are included to provide a further understanding of the application, and are incorporated in and constitute a part of this application. The components in the figures are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the application. Corresponding parts in the drawings may be exaggerated, i.e. made larger relative to other parts in an exemplary device actually manufactured according to the present application, for convenience in showing and describing some parts of the present application. In the drawings:

fig. 1 is a schematic flow chart of a first method for managing transmission of a service data network according to an embodiment of the present application.

Fig. 2 is a second flowchart of a method for traffic data network transmission management according to an embodiment of the present application.

Fig. 3 is a third flowchart of a method for managing transmission of a service data network according to an embodiment of the present application.

Fig. 4 is a schematic diagram of execution logic of a DPU-based QoS policy implementation method in an application example of the present application.

Fig. 5 is a schematic structural diagram of a service data network transmission management module according to an embodiment of the present application.

Detailed Description

For the purposes of making the objects, technical solutions and advantages of the present application more apparent, the present application will be described in further detail with reference to the embodiments and the accompanying drawings. The exemplary embodiments of the present application and their descriptions are used herein to explain the present application, but are not intended to be limiting of the present application.

It should be noted here that, in order to avoid obscuring the present application due to unnecessary details, only structures and/or processing steps closely related to the solution according to the present application are shown in the drawings, while other details not greatly related to the present application are omitted.

It should be emphasized that the term "comprises/comprising" when used herein is taken to specify the presence of stated features, elements, steps or components, but does not preclude the presence or addition of one or more other features, elements, steps or components.

It is also noted herein that the term "coupled" may refer to not only a direct connection, but also an indirect connection in which an intermediate is present, unless otherwise specified.

Hereinafter, embodiments of the present application will be described with reference to the drawings. In the drawings, the same reference numerals represent the same or similar components, or the same or similar steps.

Conventional CPU-based traffic management schemes use static QoS policies with the following drawbacks:

1. lack of flexibility: the static QoS strategy is predefined and fixed in design, and cannot be dynamically adjusted according to real-time network conditions and service requirements. This results in an inability to quickly adapt and adjust QoS policies in the face of different traffic types and network traffic variations, resulting in reduced performance and no QoS guarantees.

2. Service traffic that cannot accommodate diversity: static QoS policies are typically based on generalized assumptions and rules that do not adequately account for and cope with the differences and special needs between different traffic types. Due to the diversity of traffic flows, static QoS policies may not provide personalized optimization and service guarantees for each traffic type.

3. High variability traffic cannot be handled: static QoS policies cannot respond in time to rapid changes and fluctuations in network traffic. When faced with bursty high traffic or congestion conditions, static QoS policies cannot adjust priority and bandwidth allocation in real-time, resulting in performance degradation and QoS instability.

4. Performance bottlenecks: because the flow management scheme based on the CPU is limited by the calculation and processing capacity of the CPU, the real-time processing requirement of the high-speed flow of the large-scale data center cannot be met. This leads to problems of increased delay, reduced throughput and performance bottlenecks.

Therefore, in order to solve the above-mentioned problems, a Data Processing Unit (DPU) is introduced as a key technology for service traffic management, and embodiments of the present application provide a service data network transmission management method, a service data network transmission management module for executing the service data network transmission management method, and a DPU including the service data network transmission management module, which can implement accurate identification and QoS policy allocation for different service traffic to provide high-quality network service, and can improve flexibility of service data network transmission management, so as to be applicable to service traffic transmission scenarios with diversity and high variability.

The following examples are provided to illustrate the invention in more detail.

Based on this, the embodiment of the present application provides a service data network transmission management method that may be implemented in a DPU, referring to fig. 1, where the service data network transmission management method specifically includes the following contents:

Step 100: according to the service type of each data packet to be transmitted by the network, searching the QoS strategy data of each data packet in the comparison relation data between the current QoS strategy and the service type, and outputting each data packet to the target network for transmission based on the QoS strategy data of each data packet.

In one or more embodiments of the present application, DPU (Data Processing Unit) is a data processing unit, i.e., a dedicated hardware accelerator, with high performance and low latency characteristics, capable of efficiently handling and managing different types of traffic. The DPU has the advantage of high parallel processing capacity and special hardware acceleration, and can realize efficient flow identification, classification and processing. Compared with the traditional CPU-based scheme, the DPU can capture and analyze the header information of the data packet more quickly, and accurately judge the service type through the feature extraction and matching algorithm. Meanwhile, the DPU has parallel processing capability, and can process a plurality of data packets simultaneously, so that the processing efficiency is improved.

In addition, the DPU has flexibility and programmability, and can define and adjust QoS strategies according to specific requirements of a data center. By predefining and storing QoS policies of different traffic types, the DPU can select and apply corresponding policies according to the traffic types of the data packets. Meanwhile, the DPU supports dynamic adjustment, and timely adjustment of priority and bandwidth allocation can be carried out according to real-time network states and service flow conditions so as to ensure execution of QoS policies and optimization of network performance.

In step 100, the DPU acquires data packets sent to the data center, where the data packets may be multiple data packets acquired by the DPU at the same time, or may be acquired individually in real time, and may be specifically set according to the actual application situation. It is understood that the data packet may refer to a compressed packet of service data.

In one or more embodiments of the present application, the collation relation data between QoS policies and traffic types refers to data for storing correspondence relations between respective QoS policy data and respective traffic types.

The QoS policy data may include at least one of policy data such as priority, bandwidth allocation weight, delay requirement, and packet loss rate requirement.

Step 200: and monitoring the real-time state of the target network and each data packet, and judging whether to dynamically adjust QoS strategy data of the data packet according to a corresponding monitoring result.

In step 200, the DPU needs to dynamically monitor the network state and the traffic conditions. The DPU can adjust the priority and bandwidth allocation in due time according to the real-time network condition and service requirement. For example, the bandwidth occupation ratio of different service types is adjusted according to the network load condition and the priority of the service types.

As can be seen from the above description, the service data network transmission management method provided by the embodiment of the present application can improve the QoS capability of a data center by introducing a DPU, and can implement accurate identification and QoS policy allocation for different service flows, so as to provide high-quality network services; the target network and each data packet are used for carrying out real-time state monitoring and dynamic QoS strategy adjustment, so that the flexibility of service data network transmission management can be improved, and the method is applicable to various and highly-changeable service flow transmission scenes; compared with the traditional scheme based on the CPU, the method has the advantages of obvious high performance, low delay, high parallel processing capacity, flexibility, programmability and the like.

In order to further improve the validity and reliability of acquiring the service type to which each data packet belongs, in the service data network transmission management method provided in the embodiment of the present application, referring to fig. 2, before step 100 in the service data network transmission management method, the following contents are specifically included:

step 010: and receiving the data packet to be transmitted by the network.

In step 010, once the packet arrives at the data center, the DPU will acquire the packet to begin the identification and classification operations.

Step 020: and analyzing the data packet to obtain the header information of the data packet.

In step 020, the DPU captures and parses header information of the packet, which may include source address, destination address, transport protocol type, packet size, request header, etc.

Step 030: based on a preset service type identification mode, identifying and obtaining the service type of the data packet according to the head information of the data packet.

In step 030, by analyzing this information, the DPU can determine the traffic type to which the packet belongs. For example, some specific source and destination addresses may be associated with a specific service, or some specific transport protocol may be used for a certain service, even the size of the data packet and the request header may imply the type of service.

That is, embodiments of the present application may utilize a data plane development suite DPDK (Data Plane Development Kit) library to enable capture and parsing of data packets using a data packet capture and analysis library supported by a DPU. All available ethernet devices are then started, packets are then received from these devices, and header information of these packets is parsed out. And finally, adding a code of the user to identify and classify the data packet according to the analyzed header information. The DPU outputs the classification result, which may be a marked traffic type or other form of representation.

In order to further improve accuracy, efficiency and effectiveness of identifying the service type to which the data packet belongs according to the header information of the data packet, in the service data network transmission management method provided in the embodiment of the present application, the service type identification manner includes: feature matching, machine learning or deep packet inspection; referring to fig. 3, the step 030 in the service data network transmission management method specifically includes the following:

step 031: and if the service type identification mode is the characteristic matching, extracting service characteristics from the header information of the data packet, and searching the service type of the data packet in comparison relation data between the preset service characteristics and the service types based on the service characteristics.

Specifically, the DPU may extract features related to the service type, i.e., service features, by analyzing header information obtained by parsing. For example, specific identifiers in the source and destination addresses may be extracted to determine whether to associate with a particular service; it can be checked whether the transport protocol type matches a certain known traffic protocol; the traffic type may be inferred from the size of the data packet and the request header information.

The DPU then matches the extracted features with predefined traffic features. The predefined traffic characteristics may include a mapping table of source and destination addresses, known traffic protocols and related characteristics, association rules of traffic types with packet sizes and request headers, etc.

In order to further improve the degree of automation and intelligence of identifying the service type to which the data packet belongs according to the header information of the data packet, referring to fig. 3, step 030 in the service data network transmission management method may further include the following:

step 032: if the service type identification mode is the machine learning mode, the head information of the data packet is input into a preset machine learning model for identifying the service type, so that the service type identification model correspondingly outputs the service type to which the data packet belongs.

In particular, by training the model, the DPU is enabled to automatically learn and identify traffic types based on the characteristics of the data packets without requiring predefined mappings or rules. The method can provide more accurate and adaptive traffic classification and adapt to the continuously changing service types and traffic patterns. The machine learning model for identifying the service type can select a classification model such as a decision tree model, a support vector machine model, a logistic regression model and the like.

In order to further improve accuracy of identifying the service type to which the data packet belongs according to the header information of the data packet, referring to fig. 3, step 030 in the service data network transmission management method may further include the following:

step 033: and if the service type identification mode is the deep data packet detection, performing the deep data packet detection on the head information of the data packet to identify the service type to which the data packet belongs.

Specifically, deep learning and deep packet inspection techniques are employed to perform deep analysis of the content of the packets to identify and classify traffic. Through deep data packet detection, the DPU can acquire more detail information, and the accuracy of identification and classification is improved.

Deep packet inspection, among other things, is a technique for inspecting and analyzing the contents of packets in real time as they flow through a network, thereby enabling real-time analysis and decision making. It is used for a variety of purposes including security, traffic management, data leakage, policy violations, malware, and quality of service. Deep packet inspection allows packet data to be checked at the application layer of the network, rather than just header information, which provides more information about the packet content. It involves looking at the actual payload or content of the data packet, including the data being transmitted and the application that generated it.

In order to further improve the efficiency, flexibility and reliability of searching the QoS policy data of each data packet in the comparison relationship data between the current QoS policy and the service type, referring to fig. 2 or fig. 3, the following are specifically included before step 100 in the service data network transmission management method:

step 040: and receiving the comparison relation data between the preset QoS strategy and the service type and storing the comparison relation data to the local.

Specifically, qoS policies are predefined and stored in the DPU. Each traffic type corresponds to one or more QoS policies. These policies may be set according to the nature of the traffic type and the requirements of the data center. For example, for traffic requiring low latency, its priority may be set higher; for a large amount of data transmission traffic, the bandwidth usage weight thereof may be set to be large. When the DPU identifies and classifies a data packet, a corresponding QoS policy is selected and applied according to the traffic type of the data packet.

In the DPU, qoS policies of different traffic types are predefined and stored. Each traffic type corresponds to one or more QoS policies. QoS policies may include, but are not limited to, the following: priority setting, bandwidth allocation, delay requirements, packet loss rate requirements, etc. These policies are set according to the nature of the traffic type and the requirements of the data center.

Correspondingly, in order to further improve the flexibility and the applicability of the service data network transmission management, a step 050 is specifically included before the step 100 or after the step 100 in the service data network transmission management method, referring to fig. 2 or fig. 3, taking the step 050 performed before the step 100 as an example, the step 050 specifically includes the following contents:

step 050: and self-adaptively adjusting the control relation data between the QoS strategy and the service type according to the network state of the target network, and/or adjusting the control relation data between the QoS strategy and the service type based on the user-defined QoS strategy data if the user-defined QoS strategy data is received.

Specifically, by adopting the self-adaptive QoS strategy setting method, qoS parameters can be adjusted according to the real-time network state and service requirements. By monitoring network load, delay, bandwidth utilization, etc., the DPU can dynamically adjust priority and bandwidth allocation to accommodate different network environments and traffic demands.

And meanwhile, the user can be allowed to customize the QoS strategy according to the self requirement. The DPU provides a user interface or configuration interface to enable a user to flexibly set parameters such as priority, bandwidth allocation, delay requirements, etc. The method can meet the personalized requirements of different users or businesses.

In order to further improve efficiency, flexibility and reliability of service data network transmission management, in the service data network transmission management method provided in the embodiment of the present application, the QoS policy data may at least include priority and bandwidth allocation weight; correspondingly, referring to fig. 2, step 100 in the service data network transmission management method specifically includes the following:

step 110: and according to the service type of each data packet to be transmitted by the network, searching the corresponding priority and bandwidth allocation weight of each data packet in the comparison relation data between the current QoS strategy and the service type.

Specifically, when the DPU identifies and classifies a packet, the service type to which the packet belongs is first determined according to the traffic identification and classification scheme described above. For example, if feature matching is successful, the data packet may be marked as a particular traffic type, such as a video stream, an audio stream, a Web request, and so on.

When the DPU identifies and classifies the data packet, the service type of the data packet is determined according to the traffic identification and classification method. The DPU then selects the corresponding QoS policies based on the traffic type of the data packet. Which QoS policy to choose may be determined based on predefined mappings or rules. For example, the matching may be performed by a mapping table of traffic types and QoS policies, or it may be determined which QoS policy needs to be applied according to characteristics of the traffic types. So that the DPU may subsequently apply the selected QoS policy to the data packets or related traffic. Depending on the QoS policy selected, the DPU may adjust the priority of the packets, bandwidth allocation weights, or other relevant parameters.

To distinguish between different types of network traffic, qoS introduces some classification labels, of which EF, AF and BE are the three most common labels.

EF (Expedited Forwarding) is a high priority label, and is mainly used for network traffic with high real-time performance, such as VoIP (Voice over IP) and video communication. The EF marks indicate that the packet has the highest priority, the transmission speed should be as fast as possible, and the delay and jitter should be small to ensure real-time and stability.

AF (Assured Forwarding) is a medium priority label, mainly for network traffic that requires a certain priority but does not require strict guarantees, such as network games, file transfer, etc. The AF flag classifies the packet with 4 different priorities, each with 3 different subclasses for a total of 12 priorities. Specifically, the 4 performance categories in the AF flag are AF1, AF2, AF3, and AF4, respectively.

BE (Best Effort) is a low priority tag, mainly for non-real time data streams, such as e-mail, file downloads, etc. The BE flag indicates that the packet has no special priority handling requirements, the network switch will handle other priority packets, and the BE flag packet will handle lower priority.

In a word, three marks of EF, AF and BE are used for classifying the priorities of different types of network traffic so as to ensure that various types of data can BE reasonably processed and guaranteed during transmission.

For each service type, the corresponding bandwidth allocation weight can be calculated according to the priority and the weight calculation formula.

EF (Expedited Forwarding): since EF is a high priority traffic type, its weight can be set to the highest value, e.g. 1.0.

AF (Assured Forwarding): according to the 4 performance classes of AF (AF 1, AF2, AF3 and AF 4), each class has 3 different subclasses, totaling 12 priorities. Each sub-class is assigned an appropriate weight, e.g., 0.8, 0.6, 0.4, etc., according to traffic and performance requirements.

BE (Best Effort): since BE is a low priority traffic type, its weight can BE set to a minimum value, e.g. 0.2.

Step 120: and sequencing the data packets according to the order of the priority from high to low.

Specifically, the specific manner in which the DPU performs priority adjustment on the data packet may be: the packets are ordered according to priority rules defined in the QoS policy, with the high priority packets being ordered in front.

And, step 130: and respectively distributing current bandwidth resources of a target network to each data packet according to the bandwidth distribution weight of each data packet.

In particular, the DPU may assign weights based on bandwidth, and assign available bandwidth to each traffic flow in a weighted proportion. Bandwidth allocation may be performed using weighted fair queuing (Weighted Fair Queuing) or other suitable algorithms.

Weighted fair queuing (Weighted Fair Queuing): and allocating corresponding bandwidth for each service type according to the weight proportion. For example, if EF weights 1.0, AF weights 0.8, BE weights 0.2, then EF traffic will get 50% bandwidth, AF traffic will get 40% bandwidth, BE traffic will get 10% bandwidth.

Step 140: and outputting the sequenced data packets to the target network in sequence for transmission based on the bandwidth resources corresponding to the data packets.

Specifically, the specific manner in which the DPU manages and transmits traffic according to the QoS policy settings may be: according to the result of priority and bandwidth allocation, the DPU processes the data packets according to the set priority order and allocates an appropriate bandwidth to each traffic.

That is, the DPU processes the packets in the order of priority and transmits the packets in accordance with the bandwidth allocation result. Wherein high priority data packets are preferentially processed and transmitted, and each service flow occupies a corresponding proportion of bandwidth for transmission according to its bandwidth allocation weight.

In steps 110 to 140, the DPU performs priority adjustment and bandwidth allocation according to the QoS policy corresponding to each packet. Specifically, the DPU first prioritizes the packets. High priority packets are prioritized and sent. At the same time, the DPU also controls the bandwidth usage of each traffic flow. This is typically achieved by means of weights, the bandwidth usage weight of each traffic flow determining the proportion of its available bandwidth. In this process, the DPU needs to dynamically monitor the network status and traffic conditions in order to make timely adjustments to the priority and bandwidth allocation, i.e., the content of step 200 described above.

In order to further improve efficiency, flexibility and reliability of service data network transmission, in the service data network transmission management method provided in the embodiment of the present application, the priority includes: high-priority EF, medium-priority AF and low-priority BE divided from high to low, wherein the medium-priority AF comprises a plurality of medium priorities divided from high to low, such as AF1, AF2, AF3 and AF4, and each medium priority comprises a plurality of sub-priorities divided from high to low; correspondingly, referring to fig. 3, step 120 in the service data network transmission management method specifically includes the following:

Step 121: and respectively placing the data packets into the queues corresponding to the high priority, the medium priority and the low priority according to the priorities corresponding to the data packets, wherein the data packets placed into the queues corresponding to the medium priority are ordered according to the order of the sub-priorities from high to low.

For example, queuing algorithms are employed for different kinds of traffic for priority scheduling of three different traffic (EF, AF, BE):

a) Three queues are defined: the system comprises an EF queue, an AF queue and a BE queue, wherein the AF queue comprises an AF1 queue to an AF4 queue and is used for storing data packets with corresponding priorities.

B) The data packet enters the queue: and placing the data packets into corresponding queues according to the marks of the data packets. The EF marked data packet is placed in an EF queue, the AF marked data packet is placed in an AF queue, and the BE marked data packet is placed in a BE queue.

The queuing algorithm can reasonably schedule and process different types of services according to the priority of the data packet, ensure the instantaneity and stability of the EF data packet, provide a certain priority service of the AF data packet and process the low priority requirement of the BE data packet. The algorithm can be adjusted and optimized according to specific requirements to meet the priority requirements of different services and the change of network conditions.

In order to further improve the efficiency, flexibility and reliability of the service data network transmission, in the method for managing service data network transmission provided in the embodiment of the present application, referring to fig. 3, step 130 in the method for managing service data network transmission specifically includes the following:

step 131: and acquiring the current total bandwidth resource of the target network.

Step 132: and dividing the percentages of the total bandwidth resources according to the respective bandwidth allocation weights of the data packets, and allocating the current bandwidth resources of the target network to the data packets based on the corresponding percentage division results.

Specifically, the pseudo code corresponding to step 132 is shown in Table 1.

TABLE 1

In the above pseudo code, a coefficient variable is added for adjusting the proportion of the weights. This coefficient can be set according to the actual requirements to achieve a more flexible bandwidth allocation.

In table 1, ef_weight represents the expedited forwarding queue weight, indicating that the packet has the highest priority and needs to be transmitted quickly. AF stands for guaranteed forwarding queue, representing medium priority packets for network traffic that requires a certain priority but does not require strict guarantees. There are 4 queues in the AF queue, wherein AF1_weight indicates that there is a guaranteed forwarding queue 1 weight, AF2_weight indicates that there is a guaranteed forwarding queue 2 weight, AF3_weight indicates that there is a guaranteed forwarding queue 3 weight, and AF4_weight indicates that there is a guaranteed forwarding queue 4 weight.

BE is a best effort queue representing low priority packets, primarily for non-real time data flows. Be_weight indicates best effort queue weights. BE queue is an AF from AF queue value _{n_weight} Wherein n ranges from 1 to 4.

Total_Bandwidth represents the Total Bandwidth, which represents the Total transmission capacity of the entire network or a particular portion. This is calculated from the actual situation and is related to the performance of the device after leaving the factory.

coefficient is a constant coefficient used to adjust the weight of computing the bandwidth of each class of service. By adjusting this coefficient, the bandwidth allocation ratio of each service class can be affected.

Ef_bandwidth represents the value of the accelerated forwarding queue bandwidth; AF1 bandwidth represents guaranteed forwarding queue 1 bandwidth value; AF2 bandwidth represents guaranteed forwarding queue 2 bandwidth value; AF3_ bandwidth represents the guaranteed forwarding queue 3 bandwidth value; AF4 bandwidth represents the guaranteed forwarding queue 4 bandwidth value; be_bandwidth represents a best effort queue bandwidth value.

In order to further improve the efficiency, flexibility and reliability of the service data network transmission, in the method for managing service data network transmission provided in the embodiment of the present application, referring to fig. 3, step 140 in the method for managing service data network transmission specifically includes the following contents:

Step 141: and based on the bandwidth resources corresponding to the data packets, sequentially outputting the data packets in the queues corresponding to the high priority, the medium priority and the low priority to the target network in a first-in first-out mode for transmission.

In one example, the specific process of processing the data packets in the order of priority may be: and taking out a data packet from the EF queue for processing and transmission. If the EF queue is empty, a data packet is fetched from the AF queue for processing and transmission. If the AF queue is also empty, a packet is fetched from the BE queue for processing and transmission.

Wherein, the data packets are processed in a first-in first-out (FIFO) mode in each queue. I.e., in each queue, packets that first enter the queue are processed and transmitted first.

In order to further improve the efficiency, flexibility and reliability of the service data network transmission, in the service data network transmission management method provided in the embodiment of the present application, referring to fig. 2 or fig. 3, step 200 in the service data network transmission management method specifically includes the following contents:

step 210: and monitoring the target network and each data packet in real time, and judging whether the data packets in any queue exceed a preset quantity threshold value or not, whether the current network state of the target network meets a preset congestion condition or whether broadband resources of any data packet are smaller than a preset resource threshold value corresponding to the data packet or not according to a corresponding monitoring result.

Step 220: if the data packets in any queue exceed the preset quantity threshold, the data packets are extracted from the queue, and the priority corresponding to the extracted data packets is modified so that the data packets with the modified priorities are added into the corresponding queues again.

Step 230: if the current network state of the target network meets the preset congestion condition, dynamically adjusting the priority and broadband allocation weight of the data packet which is not output to the target network currently, and/or sending a network congestion prompt message to the sending end of the data packet which is not output to the target network currently.

Step 240: and if any broadband resource of the data packet is smaller than the preset resource threshold corresponding to the data packet, dynamically adjusting the broadband allocation weight of the data packet so as to reallocate the broadband resource of the data packet.

Specifically, the DPU dynamically monitors the network state and the condition of traffic flows. The DPU can adjust the priority and bandwidth allocation in due time according to the real-time network condition and service requirement. For example, if the network is congested or some traffic requires more bandwidth, the DPU may recalculate the weights and perform bandwidth reassignment.

That is, the priority scheduling policy may be dynamically adjusted based on real-time network conditions and traffic demands. For example, if there are too many packets in the EF queue, the weight of the priority schedule may be adjusted so that the packets in the EF queue can be processed and transmitted more quickly. The number of data packets and the network condition of each queue are monitored in real time. According to the monitoring result, feedback information can be provided to the sending end, for example, a congestion signal is sent to the sending end through a congestion control mechanism, so as to control the sending rate of the data packet.

Meanwhile, the DPU has the capability of fault recovery and exception handling. When a fault or abnormal situation occurs, such as a network interruption or packet loss situation, the DPU may take corresponding measures, such as retransmitting the lost data packet, adjusting priority or bandwidth allocation, etc., so as to ensure continuity and reliability of the service.

In order to further explain the service data network transmission management method provided by the above embodiment, the present application further provides an application example of a QoS policy implementation method based on a DPU, which can guarantee priorities and bandwidths of different service flows. Specifically, the DPU will first identify and classify incoming traffic and then set priority and allocate bandwidth according to preset QoS policies. And finally, the DPU transmits and manages the service traffic according to the adjusted priority and bandwidth.

Referring to fig. 4, after capturing a data packet from a data stream, a dpu sequentially performs header information analysis, feature extraction, feature matching and service type judgment, and then performs service type matching according to the obtained service type, and in this process, monitors network state and service flow in real time; and then according to the service type obtained by matching, selecting QoS strategy selection corresponding to the service type in a strategy warehouse formed by defining strategy rules and RBAC admission in advance according to the service type, distinguishing service strategy execution points, adjusting the priority of the data packet according to a data packet classification model, adding the data packet into each queue (such as EF queue, AF1 queue, AF2 queue and BE queue) in the QoS strategy queue according to the priority, and then carrying out bandwidth allocation weight calculation, bandwidth allocation, data packet processing and sending according to the priority, and carrying out fault recovery and abnormal processing when faults are detected.

The specific application example of the DPU-based QoS strategy implementation method comprises the following contents:

(1) Based on the result of priority and bandwidth allocation, the DPU performs transmission and management of traffic. Specifically, the data packets are sent sequentially by priority. Each service flow also occupies a corresponding proportion of the bandwidth according to its bandwidth usage weight. Throughout the process, the DPU needs to keep monitoring of network status and traffic flows to ensure proper enforcement of QoS policies and to respond quickly to any bursty network conditions.

(2) Packet reception and ordering: the DPU receives the data packets and sorts the data packets according to priority. The high priority packets are ranked in front and the low priority packets are ranked in rear.

(3) Bandwidth allocation weight calculation: the DPU calculates the bandwidth allocation weight for each traffic flow according to the bandwidth allocation rules defined in the QoS policy. These weights determine the proportion of each traffic flow that occupies the available bandwidth.

(4) Packet transmission and bandwidth management: the DPU transmits the data packets in a priority order and manages bandwidth usage according to bandwidth allocation weights. High priority packets are sent preferentially, while each traffic occupies a corresponding proportion of the bandwidth according to its bandwidth allocation weight. This ensures that higher priority traffic gets more bandwidth resources.

(5) Network status and traffic monitoring: the DPU continuously monitors the status of network states and traffic flows. By monitoring indexes such as network delay, bandwidth utilization rate and the like in real time and analyzing service flow, the DPU can know actual conditions and service demands of the network.

(6) QoS policy enforcement and response: the DPU ensures the correct enforcement of QoS policies and responds quickly to any bursty network conditions. If network congestion or other abnormal conditions occur, the DPU can dynamically adjust priority and bandwidth allocation according to the network state and traffic flow conditions monitored in real time so as to maintain QoS performance.

Fault recovery and exception handling: the DPU has fault recovery and exception handling capabilities. When a fault or abnormal situation occurs, such as a network interruption or packet loss situation, the DPU may take corresponding measures, such as retransmitting the lost data packet, adjusting priority or bandwidth allocation, etc., so as to ensure continuity and reliability of the service.

The flow describes the DPU-based prioritization and bandwidth allocation implementation steps including packet reception, qoS policy selection, packet prioritization, bandwidth allocation weight calculation, bandwidth allocation, monitoring and adjustment, and packet processing and transmission.

In summary, the DPU-based traffic management scheme has significant advantages over conventional CPU-based schemes, including high performance, low latency, high parallel processing capability, flexibility, and programmability. By introducing the DPU, the QoS capability of the data center can be improved, and the accurate identification, priority adjustment and bandwidth allocation of different service flows can be realized so as to provide high-quality network service.

That is, the service data network transmission management method provided in the embodiment of the present application adopts a queuing algorithm and a priority scheduling mechanism for different service priorities by designing a traffic identification and classification method based on a DPU, a QoS policy setting and implementation manner, a priority adjustment and bandwidth allocation operation, a traffic transmission and management manner, and a queuing algorithm and priority scheduling. By defining different queues and processing the data packets according to the priority order, effective management and scheduling of EF, AF and BE services are realized. The core innovation in this section includes the definition of queues and the mechanism by which packets enter the queues, as well as the flow of scheduling packets according to priority.

By using the technical scheme of the application, the service flow of the data center can be managed more effectively, and the service quality of different services is ensured. Meanwhile, the flow management scheme based on the DPU can more flexibly cope with the change of the service flow, so that the overall performance and efficiency of the data center are improved.

The present application further provides a service data network transmission management module for executing all or part of the service data network transmission management method, referring to fig. 5, where the service data network transmission management module provided in the DPU specifically includes the following contents:

the QoS policy obtaining module 10 is configured to search, according to a service type to which each data packet to be transmitted in the network belongs, qoS policy data of each data packet in comparison relationship data between a current QoS policy and the service type, and output each data packet to a target network for transmission based on the QoS policy data of each data packet;

and the QoS policy dynamic adjustment module 20 is used for monitoring the real-time state of the target network and each data packet, and judging whether to dynamically adjust the QoS policy data of the data packet according to the corresponding monitoring result. .

The embodiment of the service data network transmission management module provided in the present application may be specifically used to execute the processing flow of the embodiment of the service data network transmission management method in the above embodiment, and the functions thereof are not described herein again, and may refer to the detailed description of the embodiment of the service data network transmission management method.

The part of the service data network transmission management module for carrying out service data network transmission management can be completed in the client device. Specifically, the selection may be made according to the processing capability of the client device, and restrictions of the use scenario of the user. The present application is not limited in this regard. If all operations are performed in the client device, the client device may further include a processor for performing specific processing of traffic data network transmission management.

The client device may have a communication module (i.e. a communication unit) and may be connected to a remote server in a communication manner, so as to implement data transmission with the server. The server may include a server on the side of the task scheduling center, and in other implementations may include a server of an intermediate platform, such as a server of a third party server platform having a communication link with the task scheduling center server. The server may include a single computer device, a server cluster formed by a plurality of servers, or a server structure of a distributed device.

Any suitable network protocol may be used for communication between the server and the client device, including those not yet developed at the filing date of this application. The network protocols may include, for example, TCP/IP protocol, UDP/IP protocol, HTTP protocol, HTTPS protocol, etc. Of course, the network protocol may also include, for example, RPC protocol (Remote Procedure Call Protocol ), REST protocol (Representational State Transfer, representational state transfer protocol), etc. used above the above-described protocol.

As can be seen from the above description, the service data network transmission management module provided in the embodiments of the present application can improve QoS capability of a data center by introducing a DPU, and can implement accurate identification and QoS policy allocation for different service flows, so as to provide high-quality network services; the target network and each data packet are used for carrying out real-time state monitoring and dynamic QoS strategy adjustment, so that the flexibility of service data network transmission management can be improved, and the method is applicable to various and highly-changeable service flow transmission scenes; compared with the traditional scheme based on the CPU, the method has the advantages of obvious high performance, low delay, high parallel processing capacity, flexibility, programmability and the like.

The embodiment of the application also provides a DPU, which can comprise a service data network transmission management module;

the service data network transmission management module is configured to execute the service data network transmission management method provided in the foregoing embodiment;

the service data network transmission management module is in communication connection with the data center, and can be arranged in the data center to acquire a data packet currently sent to the data center from the data center.

The embodiments of the present application also provide a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the foregoing traffic data network transmission management method. The computer readable storage medium may be a tangible storage medium such as Random Access Memory (RAM), memory, read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, floppy disks, hard disk, a removable memory disk, a CD-ROM, or any other form of storage medium known in the art.

Those of ordinary skill in the art will appreciate that the various illustrative components, systems, and methods described in connection with the embodiments disclosed herein can be implemented as hardware, software, or a combination of both. The particular implementation is hardware or software dependent on the specific application of the solution and the design constraints. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application. When implemented in hardware, it may be, for example, an electronic circuit, an Application Specific Integrated Circuit (ASIC), suitable firmware, a plug-in, a function card, or the like. When implemented in software, the elements of the present application are the programs or code segments used to perform the required tasks. The program or code segments may be stored in a machine readable medium or transmitted over transmission media or communication links by a data signal carried in a carrier wave.

It should be clear that the present application is not limited to the particular arrangements and processes described above and illustrated in the drawings. For the sake of brevity, a detailed description of known methods is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present application are not limited to the specific steps described and illustrated, and those skilled in the art can make various changes, modifications, and additions, or change the order between steps, after appreciating the spirit of the present application.

The features described and/or illustrated in this application for one embodiment may be used in the same way or in a similar way in one or more other embodiments and/or in combination with or instead of the features of the other embodiments.

The foregoing description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and variations may be made to the embodiment of the present application by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims (10)

1. A traffic data network transmission management method, performed in a DPU, the traffic data network transmission management method comprising:

according to the service type of each data packet to be transmitted by the network, searching the QoS strategy data of each data packet in the comparison relation data between the current QoS strategy and the service type, and outputting each data packet to a target network for transmission based on the QoS strategy data of each data packet;

and monitoring the real-time state of the target network and each data packet, and judging whether to dynamically adjust QoS strategy data of the data packet according to a corresponding monitoring result.

2. The method for traffic data network transmission management according to claim 1, wherein before searching the QoS policy data of each data packet in the comparison relationship data between the current QoS policy and the traffic type according to the traffic type to which each data packet to be transmitted belongs, respectively, the method further comprises:

receiving a data packet to be transmitted by a network;

analyzing the data packet to obtain header information of the data packet;

Based on a preset service type identification mode, identifying and obtaining the service type of the data packet according to the head information of the data packet.

3. The traffic data network transmission management method according to claim 2, wherein the traffic type identification means comprises: feature matching, machine learning or deep packet inspection;

correspondingly, the identifying, based on the preset service type identifying manner, the service type to which the data packet belongs according to the header information of the data packet includes:

if the service type identification mode is the characteristic matching, extracting service characteristics from the header information of the data packet, and searching the service type of the data packet in comparison relation data between the preset service characteristics and the service types based on the service characteristics;

if the service type identification mode is the machine learning mode, inputting the head information of the data packet into a preset machine learning model for identifying the service type, so that the service type identification model correspondingly outputs the service type to which the data packet belongs;

and if the service type identification mode is the deep data packet detection, performing the deep data packet detection on the head information of the data packet to identify the service type to which the data packet belongs.

4. The method for traffic data network transmission management according to claim 1, wherein before searching the QoS policy data of each data packet in the comparison relationship data between the current QoS policy and the traffic type according to the traffic type to which each data packet to be transmitted belongs, respectively, the method further comprises:

receiving control relation data between a preset QoS strategy and a service type and storing the control relation data to a local;

correspondingly, the service data network transmission management method further comprises the following steps:

and self-adaptively adjusting the control relation data between the QoS strategy and the service type according to the network state of the target network, and/or adjusting the control relation data between the QoS strategy and the service type based on the user-defined QoS strategy data if the user-defined QoS strategy data is received.

5. The traffic data network transmission management method according to claim 1, wherein the QoS policy data includes: priority and bandwidth allocation weights;

correspondingly, according to the service type to which each data packet to be transmitted by the network belongs, respectively searching the QoS policy data of each data packet in the comparison relationship data between the current QoS policy and the service type, and respectively outputting each data packet to a target network for transmission based on the QoS policy data of each data packet, including:

According to the service type of each data packet to be transmitted by the network, searching the corresponding priority and bandwidth allocation weight of each data packet in the comparison relation data between the current QoS strategy and the service type;

sequencing the data packets according to the order of the priority from high to low;

the current bandwidth resources of the target network are respectively distributed to the data packets according to the bandwidth distribution weights of the data packets;

and outputting the sequenced data packets to the target network in sequence for transmission based on the bandwidth resources corresponding to the data packets.

6. The traffic data network transmission management method according to claim 5, wherein the priority comprises: a high priority, a medium priority and a low priority which are divided from high to low, wherein the medium priority comprises a plurality of medium priorities which are divided from high to low, and each medium priority comprises a plurality of sub-priorities which are divided from high to low;

correspondingly, the sorting the data packets according to the order of the priority from high to low includes:

And respectively placing the data packets into the queues corresponding to the high priority, the medium priority and the low priority according to the priorities corresponding to the data packets, wherein the data packets placed into the queues corresponding to the medium priority are ordered according to the order of the sub-priorities from high to low.

7. The traffic data network transmission management method according to claim 5, wherein the allocating current bandwidth resources of the target network to each of the data packets according to the respective bandwidth allocation weights of each of the data packets comprises:

acquiring the current total bandwidth resource of the target network;

and dividing the percentages of the total bandwidth resources according to the respective bandwidth allocation weights of the data packets, and allocating the current bandwidth resources of the target network to the data packets based on the corresponding percentage division results.

8. The traffic data network transmission management method according to claim 6, wherein the outputting the sequenced data packets to the target network for transmission based on the bandwidth resources corresponding to the data packets, respectively, includes:

And based on the bandwidth resources corresponding to the data packets, sequentially outputting the data packets in the queues corresponding to the high priority, the medium priority and the low priority to the target network in a first-in first-out mode for transmission.

9. The traffic data network transmission management method according to claim 6, wherein the performing real-time status monitoring on the target network and each of the data packets and determining whether to dynamically adjust QoS policy data of the data packets according to the corresponding monitoring result includes:

the target network and each data packet are monitored in real time, and whether the data packet in any queue exceeds a preset quantity threshold value, whether the current network state of the target network meets preset congestion conditions or whether broadband resources of any data packet are smaller than a preset resource threshold value corresponding to the data packet or not is judged according to a corresponding monitoring result;

if the data packets in any queue exceed the preset quantity threshold, extracting the data packets from the queue, and modifying the priority corresponding to the extracted data packets so that the data packets with the modified priorities are added into the corresponding queues again;

If the current network state of the target network meets a preset congestion condition, dynamically adjusting the priority and broadband allocation weight of the data packet which is not output to the target network, and/or sending a network congestion prompt message to a sending end of the data packet which is not output to the target network;

and if any broadband resource of the data packet is smaller than the preset resource threshold corresponding to the data packet, dynamically adjusting the broadband allocation weight of the data packet so as to reallocate the broadband resource of the data packet.

10. The DPU is characterized in that a service data network transmission management module is arranged in the DPU;

the service data network transmission management module is configured to perform the service data network transmission management method according to any one of claims 1 to 9;

the service data network transmission management module is in communication connection with the data center so as to acquire a data packet currently sent to the data center from the data center.