A network resource management method, device, equipment and readable storage medium
Dynamically adjusting network slicing configuration and transmission strategy through the network policy generation model, solving the problem of static network slicing bandwidth configuration, improving flexibility and application efficiency, avoiding data loss and congestion, and improving network transmission performance.
Patent Information
- Application Number
- CN202410822351.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-06-25
- Publication Date
- 2024-11-08
- Estimated Expiration
- 2044-06-25
AI Technical Summary
In the prior art, the bandwidth configuration of network slices cannot be dynamically adjusted, resulting in low flexibility, low application efficiency, and easy congestion during data transmission, resulting in loss of important data.
The network slices are divided through the initial network configuration policy, and the network policy generation model is used to generate network configuration update policies based on the bandwidth requirements of the data packet to be transmitted, the network slice bandwidth of the non-occupancy state and the future network state, and dynamically adjust the network slice configuration and data transmission policy. , ensure smooth data transmission.
It improves the flexibility and application efficiency of network slicing, reduces data loss, and combines historical and future data generation transmission strategies to improve the transmission performance of the switch and avoids data congestion.
Smart Images

Figure CN118382108B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of network data transmission, and in particular to a network resource management method, device, equipment and readable storage medium. Background Art
[0002] With the development of fifth generation (5G) and beyond 5G (B5G) technologies, data traffic has surged and services have become more diverse, bringing new challenges to network resource management. Especially when various applications coexist on a shared network infrastructure, bandwidth congestion and unstable connection quality can cause problems. NS (Network Slicing) has emerged as a strategic solution. By dividing the network into different slices, each slice is tailored for a specific service or application.
[0003] However, since the types of data traffic transmitted through each network slice are different, the required bandwidth is also different. In the existing technology, the network slice bandwidth of the switch is directly configured and cannot be changed, resulting in low flexibility and low application efficiency of the network slice.
[0004] When network slicing transmits data, if there is a lot of data, congestion is likely to occur. Congestion will cause some data to be discarded, resulting in the loss of important data.
[0005] Therefore, the present application provides a network resource management method, apparatus, device and readable storage medium. Summary of the invention
[0006] In view of this, the object of the present invention is to provide a network resource management method, device, equipment and readable storage medium to improve the flexibility and application efficiency of network slicing while reducing the loss of important data.
[0007] In a first aspect, a network resource management method is provided, comprising:
[0008] Divide the transmission network of the target switch into multiple network slices based on the initial network configuration policy;
[0009] Obtain the bandwidth demand of the network data packets to be transmitted on the target switch, the bandwidth of the network slices in the unoccupied state, and the future network status;
[0010] Determine whether the initial network configuration strategy needs to be updated based on the bandwidth demand of the network data packets to be transmitted, the bandwidth of the network slices in the non-occupied state, and the future network status;
[0011] If necessary, the bandwidth requirement of the network data packet to be transmitted, the bandwidth of the network slice in the non-occupied state and the future network state are input into the trained network policy generation model to generate a network configuration update strategy; the network configuration update strategy includes a network slice configuration strategy and a data transmission strategy;
[0012] Re-divide network slices based on network slice configuration strategy and determine target network slices;
[0013] The data transmission policy is sent to the target network slice so that the target network slice transmits the network data packets to be transmitted according to the data transmission policy.
[0014] Optionally, obtaining the bandwidth requirement of the network data packet to be transmitted on the target switch includes:
[0015] Obtaining the traffic type and number of network data packets to be transmitted, where the traffic type includes at least one or more combinations of text, voice, video and picture;
[0016] Determine the bandwidth required for each traffic type data packet according to the traffic type and the preset correspondence between the traffic type and the bandwidth;
[0017] The bandwidth requirement of the network data packets to be transmitted is determined based on the bandwidth required for each data packet and the number of data packets.
[0018] Optionally, obtaining the future network state of the network slice in the non-occupied state includes:
[0019] Obtain the transmission rate of the token bucket of each non-occupied network slice and the transmission rate of the network data packet to be transmitted;
[0020] The future network state of the network slice in the non-occupied state is determined based on the rates of the two. If the transmission rate of the token bucket is greater than the transmission rate of the network data packets to be transmitted, the future network state is smooth; otherwise, the future network state is congested.
[0021] Optionally, the bandwidth of the network slice in the non-occupied state includes the bandwidth of a single network slice and the total bandwidth of all network slices in the non-occupied state. Judging whether the initial network configuration strategy needs to be updated according to the bandwidth demand of the network data packet to be transmitted, the bandwidth of the network slice in the non-occupied state and the future network state includes:
[0022] Determine whether the total bandwidth of all network slices in the non-occupied state is greater than or equal to the bandwidth requirement of the network data packet to be transmitted;
[0023] If the total bandwidth is greater than or equal to the bandwidth requirement of the network data packet to be transmitted, determine whether the maximum bandwidth in the network slice in the non-occupied state is less than the bandwidth requirement of the network data packet to be transmitted;
[0024] If the maximum bandwidth is less than the bandwidth requirement of the network data packet to be transmitted, the initial network configuration strategy needs to be updated;
[0025] If the maximum bandwidth is greater than the bandwidth requirement of the network data packet to be transmitted, it is determined whether the future network state in the non-occupied network slice corresponding to the maximum bandwidth is a congested state;
[0026] If so, the initial network configuration policy needs to be updated.
[0027] Optionally, the training process of the network policy generation model includes:
[0028] Obtaining data to be trained; the data to be trained includes bandwidth requirements of historical transmission data, bandwidth of network slices in a current non-occupied state corresponding to the historical transmission data, and future network states corresponding to the historical transmission data;
[0029] The data to be trained, including the bandwidth demand of historical transmission data, the bandwidth of the network slice in the current non-occupied state corresponding to the historical transmission data, and the future network state corresponding to the historical transmission data, are divided into a training set and a test set;
[0030] The training set is input into the network strategy generation model to be trained for forward noise processing;
[0031] Perform reverse denoising iterative training based on the forward denoised data;
[0032] When the preset number of iterative training times is reached, each training result and test set is input into the quality assessment network for evaluation;
[0033] Solve the optimal solution of the quality assessment network according to the optimization theory model;
[0034] The network strategy generation model corresponding to the optimal solution is determined as the final network strategy generation model.
[0035] Optionally, determining the target network slice includes:
[0036] In the re-divided network slices, compare the bandwidth of each network slice;
[0037] The network slice with the largest bandwidth is selected as the target network slice.
[0038] Optionally, the data transmission strategy includes a committed information rate and a peak information rate, and sending the data transmission strategy to the target network slice so that the target network slice transmits the network data packet to be transmitted according to the data transmission strategy includes:
[0039] Get the transmission rate of the network data packet to be transmitted;
[0040] Determine the transmission priority of the network data packets to be transmitted according to the transmission rate, committed information rate and peak information rate;
[0041] The transmission priority is sent to the token bucket corresponding to the target network slice; the token bucket transmits the network data packet according to the transmission priority.
[0042] In a second aspect, a network resource management device is provided, the device comprising:
[0043] A partitioning unit, configured to partition a transmission network of a target switch into a plurality of network slices based on an initial network configuration policy;
[0044] An acquisition unit, used to acquire the bandwidth demand of the network data packet to be transmitted on the target switch, the bandwidth of the network slice in the non-occupied state, and the future network state;
[0045] A determination unit, configured to determine whether it is necessary to update the initial network configuration strategy according to the bandwidth demand of the network data packet to be transmitted, the bandwidth of the network slice in the non-occupied state, and the future network state;
[0046] A generation unit, configured to input the bandwidth requirement of the network data packet to be transmitted, the bandwidth of the network slice in the non-occupied state, and the future network state into the trained network policy generation model to generate a network configuration update strategy when the initial network configuration strategy needs to be updated; the network configuration update strategy includes a network slice configuration strategy and a data transmission strategy;
[0047] A re-dividing unit, used to re-divide the network slice based on the network slice configuration policy and determine the target network slice;
[0048] The sending unit is used to send the data transmission policy to the target network slice, so that the target network slice transmits the network data packet to be transmitted according to the data transmission policy.
[0049] In a third aspect, an electronic device is provided, comprising a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other via the communication bus;
[0050] Memory, used to store computer programs;
[0051] The processor is used to implement any method step of the first aspect when executing the program stored in the memory.
[0052] According to a fourth aspect, a computer-readable storage medium is provided, wherein a computer program is stored in the computer-readable storage medium, and when the computer program is executed by a processor, any method step of the first aspect is implemented.
[0053] The embodiments of the present invention provide a network resource management method, device, equipment and readable storage medium, which performs initial configuration on the switch through the initial network configuration strategy, divides the slices, and dynamically adjusts the initial configuration strategy through the network strategy generation model, which can adjust the strategy in real time according to the needs of the data packets to be transmitted, re-divide the network slices, and improve the flexibility and adaptability of the network slices. In addition, the network strategy generation model also considers the future network state of the network slices in the non-occupied state when generating the network configuration update strategy, so that the data transmission strategy of the target slice can be adjusted in advance based on the future network state to meet the data transmission needs in the future. The model not only relies on real-time data, but also can generate data transmission strategies in combination with future data, which greatly improves the transmission performance of the switch. In addition, the transmission strategy of the target slice is dynamically adjusted through the data transmission strategy, which can not only avoid data congestion, but also avoid the loss of important data.
[0054] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0055] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.
[0056] Figure 1 A flow chart of a network resource management method provided by an embodiment of the present invention is shown;
[0057] Figure 2 A schematic diagram showing a data transmission strategy provided by an embodiment of the present invention is shown;
[0058] Figure 3 A schematic flow chart of data generation of a network strategy generation model provided by an embodiment of the present invention is shown;
[0059] Figure 4 A schematic diagram showing a comparison of dynamic throughputs of slicing and non-slicing traffic provided by an embodiment of the present invention is shown;
[0060] Figure 5 A schematic diagram of throughput evaluation of different flows provided by an embodiment of the present invention is shown;
[0061] Figure 6 A performance schematic diagram of a network policy generation model provided by an embodiment of the present invention is shown;
[0062] Figure 7 A schematic diagram showing the structure of a network resource management device provided by an embodiment of the present invention is shown;
[0063] Figure 8 A schematic diagram of an electronic device is shown. DETAILED DESCRIPTION
[0064] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.
[0065] The embodiment of the present invention provides a network resource management method, such as Figure 1 As shown, the following steps are included:
[0066] Step S101: Divide the transmission network of the target switch into multiple network slices based on the initial network configuration strategy.
[0067] In the embodiment of the present application, the target switch refers to a switch used for network slicing. In one implementation, the target switch may be a P4 (Programming Protocol-independent Packet Processors) switch. The P4 switch allows the behavior of network devices to be dynamically programmed according to network conditions, thereby providing a flexible and vendor-independent way to implement network slicing. This method not only improves the adaptability of the network, but also reduces the reliance on complex orchestration tools.
[0068] The initial network configuration strategy includes two strategies. One strategy is used to divide the network slices, for example, to divide several network slices and configure bandwidth, latency and other parameters for each network slice. The other strategy is used to configure the data transmission strategy for each network slice. Specifically, each network slice transmits data through a token bucket. Each token bucket is configured with a meter instance. The meter can be configured as single color, dual color or multi-color. Each color represents a different traffic processing strategy.
[0069] A single-color meter measures and limits packets based on a single rate, and takes action when the packet transmission rate exceeds a set threshold. A two-color meter provides two rate thresholds, usually one is the Committed Information Rate (CIR) and the other is the Peak Information Rate (PIR). Packets are marked green when the transmission rate is below the CIR, yellow when it is between the CIR and PIR, and red when it is above the PIR. Multi-color meters can have more rate thresholds, providing more detailed traffic control.
[0070] In a possible implementation, Figure 2 As shown, the data transmission strategy can be configured as a dual-rate, three-color mode, and the token bucket can transmit data according to the priority of the data packet. For example, green data packets are high priority, yellow are low priority, and red are discarded. This ensures that high-priority data packets pass first, and low-priority data packets are second. In this way, the P4 switch can flexibly manage network traffic and achieve advanced network performance and security strategies.
[0071] Step S102: Obtain the bandwidth requirement of the network data packet to be transmitted on the target switch, the bandwidth of the network slice in the unoccupied state, and the future network status.
[0072] In the embodiment of the present application, the bandwidth demand of the network data packet to be transmitted refers to the total flow of the network data packet to be transmitted, for example, 100 megabytes. The bandwidth of the network slice in the non-occupied state refers to the bandwidth of the network slice in the idle state without data transmission, and the future network state refers to the network state of the network slice in the non-occupied state when it is used to transmit the network data packet to be transmitted.
[0073] In one example, the amount of bandwidth of a network slice in a non-occupied state can be monitored and acquired in real time through a register array of a switch.
[0074] In the embodiment of the present application, obtaining the bandwidth requirement of the network data packet to be transmitted on the target switch includes:
[0075] Obtaining the traffic type and number of network data packets to be transmitted, where the traffic type includes at least one or more combinations of text, voice, video and picture;
[0076] Determine the bandwidth required for each traffic type data packet according to the traffic type and the preset correspondence between the traffic type and the bandwidth;
[0077] In an embodiment of the present application, the correspondence between traffic type and bandwidth can be obtained from the network service quality manual, which records the bandwidth range required for different types of traffic. For example, text-type traffic generally requires a few bits of bandwidth, while video-type traffic generally requires tens to hundreds of megabytes of bandwidth.
[0078] The bandwidth requirement of the network data packets to be transmitted is determined based on the bandwidth required for each data packet and the number of data packets.
[0079] In one example, for example, the network data packets to be transmitted include 1 text data packet and 2 voice data packets, and the bandwidth required for 1 text data packet is 10 bits, and the bandwidth required for 1 voice data packet is 20 megabits, then the total bandwidth requirement for the network data packets to be transmitted is 40 megabits and 10 bits.
[0080] In this embodiment of the present application, obtaining the future network state of a network slice in a non-occupied state includes:
[0081] Obtain the transmission rate of the token bucket of each non-occupied network slice and the transmission rate of the network data packet to be transmitted;
[0082] The future network state of the network slice in the non-occupied state is determined based on the rates of the two.
[0083] Specifically, the rates of the two are compared. If the transmission rate of the token bucket is greater than the transmission rate of the network data packets to be transmitted, the future network state is smooth, otherwise, the future network state is congested.
[0084] Step S103: Determine whether the initial network configuration strategy needs to be updated based on the bandwidth demand of the network data packet to be transmitted, the bandwidth of the network slice in the unoccupied state, and the future network status.
[0085] Among them, the bandwidth of the network slice in the unoccupied state includes the bandwidth of a single network slice and the total bandwidth of all network slices in the unoccupied state.
[0086] In an embodiment of the present application, judging whether the initial network configuration strategy needs to be updated according to the bandwidth demand of the network data packet to be transmitted, the bandwidth of the network slice in the non-occupied state, and the future network state includes:
[0087] Determine whether the total bandwidth of all network slices in the non-occupied state is greater than or equal to the bandwidth requirement of the network data packet to be transmitted;
[0088] If the total bandwidth is greater than or equal to the bandwidth requirement of the network data packet to be transmitted, determine whether the maximum bandwidth in the network slice in the non-occupied state is less than the bandwidth requirement of the network data packet to be transmitted;
[0089] If the total bandwidth is less than the bandwidth requirement of the network data packet to be transmitted, the transmission of the data packet is stopped, that is, the task of transmitting the data packet is abandoned.
[0090] If the maximum bandwidth is less than the bandwidth requirement of the network data packet to be transmitted, it means that all the network slices in the non-occupied state do not meet the transmission requirements, and the initial network configuration strategy needs to be updated;
[0091] If the maximum bandwidth is greater than the bandwidth requirement of the network data packet to be transmitted, it is determined whether the future network state in the non-occupied network slice corresponding to the maximum bandwidth is a congested state;
[0092] If yes, you need to update the initial network configuration policy. If it is not congested, you can directly transmit data according to the initial network configuration policy. If it is congested, it means that adjustments are needed.
[0093] Step S104: If necessary, the bandwidth requirement of the network data packet to be transmitted, the bandwidth of the unoccupied network slice and the future network status are input into the trained network policy generation model to generate a network configuration update strategy; the network configuration update strategy includes a network slice configuration strategy and a data transmission strategy.
[0094] In the embodiment of the present application, the network policy generation process is performed in the control plane, such as Figure 3 As shown, a training diagram of the network strategy generation model is given, first forward diffusion is performed, and then backward diffusion is performed. The forward diffusion process refers to the process in which data is gradually covered by noise. Subsequently, the model is iteratively optimized through reverse operations, gradually revealing and restoring the essential structure of the data. This process simulates a latent variable system, in which the true form of the data slowly emerges through the denoising step. In a specific embodiment, the training process of the network strategy generation model includes:
[0095] The first step is to obtain the data to be trained; the data to be trained includes the bandwidth demand of historical transmission data, the network slice bandwidth in the current non-occupied state corresponding to the historical transmission data, and the future network state corresponding to the historical transmission data.
[0096] In one example, the training data can be represented as ,in, Indicates the bandwidth demand for historical transmission data, Indicates the current non-occupied network slice bandwidth corresponding to the historical transmission data, Indicates the future network status corresponding to historical transmission data.
[0097] In the second step, the data to be trained, including the bandwidth demand of historical transmission data, the network slice bandwidth in the current non-occupied state corresponding to the historical transmission data, and the future network state corresponding to the historical transmission data, are divided into a training set and a test set.
[0098] The third step is to input the training set into the network strategy generation model to be trained for forward noise processing.
[0099] In this step, Gaussian noise can be introduced into the policy parameters, similar to gradually blurring an image. The gradual increase in noise level gradually obscures the original data, simulating the uncertainty and variability inherent in network traffic patterns.
[0100] The fourth step is to perform reverse denoising iterative training based on the forward denoised data.
[0101] After the data reaches a certain noise saturation, the reverse diffusion process begins. In this stage, the noise is removed in an orderly manner.
[0102] In one example, the policy generated by the network policy generation model can be expressed as ,in, represents the committed information rate, indicates the committed burst size, represents the peak information rate, and these three values constitute the data transmission strategy; Indicates the network slice configuration policy.
[0103] In a specific embodiment, the diffusion model of the reverse denoising is as follows:
[0104] Formula 1;
[0105] in, represents the joint probability distribution of the strategy from the initial denoising to the final generation; represents the covariance matrix; Represents the conditional probability distribution through the Markov chain, from arrive The reverse diffusion process, Indicates the optimal network configuration strategy.
[0106] Can be modeled as a Gaussian distribution Formula 2;
[0107] in, Formula 3;
[0108] Formula 4;
[0109] After conversion calculation of formula 3 and formula 4, we can get
[0110] Formula 5;
[0111] in, represents the mean, represents the amount of noise added at time step t during the forward diffusion process, represents the intensity and direction of the noise added in the i-th step, represents the parameter added in the low i step of back diffusion, represents the identity matrix, represents the network configuration update strategy generated in step i, Represents a series of parameters The cumulative product of until the i-th step.
[0112] In the fifth step, when the preset number of iterative training times is reached, each training result and test set is input into the quality assessment network for evaluation.
[0113] The sixth step is to find the optimal solution of the quality assessment network according to the optimization theoretical model.
[0114] In one example, the optimization theory model is as follows:
[0115] Formula 6;
[0116] in, represents the optimal solution of the quality assessment network, represents the quality assessment network; Represents the loss function of the network policy generation model.
[0117] Formula 6 represents the network configuration strategy for achieving optimal resource management in a specific environment e. The role of Qv is to evaluate the expected utility of each strategy f in a specific environment e, that is, the performance of the strategy in future network operations. This utility is quantified by cumulative rewards, which represents the overall benefit of the network in the future after selecting the strategy, that is, to find the optimal strategy that maximizes the expected cumulative reward by minimizing the loss function.
[0118] The network policy generation model is evaluated through this quality assessment network, quantifying the expected cumulative rewards. This quantification process not only takes into account the current state of the network, but also predicts the benefits of adhering to the selected policy in future operations.
[0119] In the seventh step, the network strategy generation model corresponding to the optimal solution is determined as the final network strategy generation model.
[0120] This network policy generation model not only relies on real-time network data, but also takes into account the future network status. It generates a network configuration update strategy based on the future network status data. It can adjust the two transmission rate values of the token bucket in advance based on the future network status, thereby making the generated data transmission strategy more accurate and meeting the data transmission needs in the future, greatly improving the performance of network slicing.
[0121] Step S105: Re-divide the network slices based on the network slice configuration strategy and determine the target network slices.
[0122] In an embodiment of the present application, determining the target network slice includes:
[0123] In the re-divided network slices, compare the bandwidth of each network slice;
[0124] The network slice with the largest bandwidth is selected as the target network slice.
[0125] By selecting the network slice with the largest bandwidth as the target network slice, the transmission requirements of the data packet can be met to the greatest extent, ensuring the smooth transmission of the data packet. It should be noted that other network slices that can meet the bandwidth for data packet transmission can also be selected as the target network slice.
[0126] Step S106: Send the data transmission policy to the target network slice so that the target network slice transmits the network data packet to be transmitted according to the data transmission policy.
[0127] In an embodiment of the present application, data transmission is performed on a data forwarding plane, a data transmission strategy includes a committed information rate and a peak information rate, and sending the data transmission strategy to a target network slice so that the target network slice transmits the network data packet to be transmitted according to the data transmission strategy includes:
[0128] Get the transmission rate of the network data packet to be transmitted;
[0129] Determine the transmission priority of the network data packets to be transmitted according to the transmission rate, committed information rate and peak information rate;
[0130] The transmission priority is sent to the token bucket corresponding to the target network slice; the token bucket transmits the network data packet according to the priority.
[0131] In the embodiment of the present application, the data transmission strategy is a meter strategy, which is a strategy set for a token bucket so that the token bucket performs data transmission according to the strategy.
[0132] In an example, for example, data packets are divided into two priorities. The ones between the committed information rate and the peak information rate are high priority and are set to green. The ones below the committed information rate are medium priority and are set to yellow. The ones above the peak information rate are low priority and are set to red. For example, there are 5 data packets with high priority, 2 data packets with medium priority, and 1 data packet with low priority. There are 7 tokens in the token bucket. Then, during the first transmission, both the high priority and medium priority data packets can be transmitted through the token bucket, while the low priority data packets are cached. This can effectively avoid data congestion.
[0133] The following is a specific test experiment to test the actual effect of the network resource management method, as follows:
[0134] First, use the network performance testing tool iPerf to simulate traffic. It can generate UDP and TCP traffic to simulate various network traffic scenarios. For UDP and TCP flows, start from a server connected to a P4 switch and end with a directed transmission to the server on the target switch.
[0135] The Flowgrind tool was used to collect the throughput of the round-trip time (RTT) of TCP traffic and UDP traffic from a switch to the target switch. Each test lasted for 80 seconds. The initial 19 seconds were characterized by the baseline state without network slicing implemented, which served as a control for subsequent comparisons. From the 20th second until the 39th second, network slicing was activated, and the initial configuration policy of the network slicing was S(0.7, 0.8), indicating 0.7 Gbps CIR and 0.8 Gbps PIR. Subsequently, the configuration policy of the slice was adjusted to S(0.2, 1) in the interval of [40, 59] seconds on the horizontal axis to simulate the dynamic adjustment of the network configuration policy. Finally, the system was restored to the non-slicing state for comparative analysis with the predefined slicing interval.
[0136] Based on the above settings, the test results are as follows: Figure 4 As shown in Figure 4, the throughput of sliced and non-sliced UDP flows varies over time, capturing the subtle behavior enforced by the metering logic of the P4 switch. Figure 4 illustrates that the sliced UDP flow consistently achieves a throughput exceeding the CIR but still below the PIR threshold. This phenomenon is attributed to the priority queuing mechanism of the P4 switch, where the CIR bandwidth is protected in the high priority queue, ensuring that the minimum service level is always met.
[0137] Figure 5A comparative analysis of the throughput of various flows over time under the P4-based network slicing strategy is shown. In this figure, there are two types of flows, flow 1 and flow 2. This experimental indirection highlights the subtle traffic management capabilities of the P4-based slicing system, especially the effectiveness of its strategy in dynamically adjusting to different load conditions. It further verifies the system's ability to maintain fidelity to the slice specification, guaranteeing service levels while ensuring fair bandwidth allocation between concurrent flows.
[0138] like Figure 6 As shown, the performance comparison between the network strategy generation model of the embodiment of the present application and the two models of PPO and DDPG is shown. The network strategy generation model of the embodiment of the present application always maintains a high reward trajectory throughout the iteration process, which reflects its excellent ability to converge to the optimal network configuration strategy. In contrast, PPO and DDPG show fluctuations in reward optimization, with DDPG showing more significant differences.
[0139] Based on the same inventive concept, a network resource management device based on a white box switch is provided, such as Figure 7 As shown, the device comprises:
[0140] The division unit 701 is used to divide the transmission network of the target switch into multiple network slices based on the initial network configuration policy.
[0141] The acquisition unit 702 is used to obtain the bandwidth requirement of the network data packet to be transmitted on the target switch, the bandwidth of the network slice in the unoccupied state and the future network status.
[0142] The judgment unit 703 is used to judge whether the initial network configuration strategy needs to be updated based on the bandwidth demand of the network data packet to be transmitted, the bandwidth of the network slice in the unoccupied state and the future network status.
[0143] The bandwidth of a network slice in an unoccupied state includes the bandwidth of a single network slice and the total bandwidth of all network slices in an unoccupied state.
[0144] The generation unit 704 is used to input the bandwidth requirement of the network data packet to be transmitted, the bandwidth of the network slice in the unoccupied state and the future network state into the trained network policy generation model to generate a network configuration update strategy when the initial network configuration strategy needs to be updated; the network configuration update strategy includes a network slice configuration strategy and a data transmission strategy.
[0145] In this unit, the training process of the network policy generation model includes:
[0146] Obtaining data to be trained; the data to be trained includes bandwidth requirements of historical transmission data, bandwidth of network slices in a current non-occupied state corresponding to the historical transmission data, and future network states corresponding to the historical transmission data;
[0147] The data to be trained, including the bandwidth demand of historical transmission data, the bandwidth of the network slice in the current non-occupied state corresponding to the historical transmission data, and the future network state corresponding to the historical transmission data, are divided into a training set and a test set;
[0148] Input the training set into the network strategy generation model to be trained for forward noise processing;
[0149] Perform reverse denoising iterative training based on the forward denoised data;
[0150] When the preset number of iterative training times is reached, each training result and test set is input into the quality assessment network for evaluation;
[0151] Solve the optimal solution of the quality assessment network according to the optimization theory model;
[0152] The network strategy generation model corresponding to the optimal solution is determined as the final network strategy generation model.
[0153] The redivision unit 705 is used to redivision the network slice based on the network slice configuration strategy and determine the target network slice.
[0154] In the unit, in the re-divided network slices, the bandwidth amount of each network slice is compared;
[0155] The network slice with the largest bandwidth is selected as the target network slice.
[0156] The sending unit 706 is used to send the data transmission policy to the target network slice, so that the target network slice transmits the network data packet to be transmitted according to the data transmission policy.
[0157] The process of transmitting a data packet through the data transmission strategy is to obtain the transmission rate of the network data packet to be transmitted;
[0158] Determine the transmission priority of the network data packets to be transmitted according to the transmission rate, committed information rate and peak information rate;
[0159] The transmission priority is sent to the token bucket corresponding to the target network slice; the token bucket transmits the network data packet according to the transmission priority.
[0160] In the embodiment of the present application, the acquisition unit is specifically used to acquire the traffic type and the number of network data packets to be transmitted, and the traffic type at least includes a combination of one or more of text, voice, video and picture;
[0161] Determine the bandwidth required for each traffic type data packet according to the traffic type and the preset correspondence between the traffic type and the bandwidth;
[0162] The bandwidth requirement of the network data packets to be transmitted is determined based on the bandwidth required for each data packet and the number of data packets.
[0163] It is also used to obtain the transmission rate of the token bucket of each non-occupied network slice and the transmission rate of the network data packet to be transmitted;
[0164] The future network state of the network slice in the non-occupied state is determined based on the rates of the two. If the transmission rate of the token bucket is greater than the transmission rate of the network data packets to be transmitted, the future network state is smooth; otherwise, the future network state is congested.
[0165] In an embodiment of the present application, the judging unit is used to judge whether the total bandwidth of all network slices in a non-occupied state is greater than or equal to the bandwidth requirement of the network data packet to be transmitted;
[0166] If the total bandwidth is greater than or equal to the bandwidth requirement of the network data packet to be transmitted, determine whether the maximum bandwidth in the network slice in the non-occupied state is less than the bandwidth requirement of the network data packet to be transmitted;
[0167] If the maximum bandwidth is less than the bandwidth requirement of the network data packet to be transmitted, the initial network configuration strategy needs to be updated;
[0168] If the maximum bandwidth is greater than the bandwidth requirement of the network data packet to be transmitted, it is determined whether the future network state in the non-occupied network slice corresponding to the maximum bandwidth is a congested state;
[0169] If so, the initial network configuration policy needs to be updated.
[0170] Based on the same technical concept, an embodiment of the present invention further provides an electronic device, such as Figure 8 As shown, it includes a processor 801 , a communication interface 802 , a memory 803 and a communication bus 804 , wherein the processor 801 , the communication interface 802 , and the memory 803 communicate with each other via the communication bus 804 .
[0171] Memory 803, used for storing computer programs;
[0172] The processor 801 is used to implement the steps of the network resource management method when executing the program stored in the memory 803.
[0173] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus.
[0174] The communication interface is used for communication between the above electronic device and other devices.
[0175] The memory may include a random access memory (RAM) or a non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.
[0176] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
[0177] The computer program product for performing the network resource management method provided in the embodiment of the present invention includes a computer-readable storage medium storing program code. The instructions included in the program code can be used to execute the method in the previous method embodiment. The specific implementation can be referred to the method embodiment, which will not be repeated here.
[0178] The network resource management device provided in the embodiment of the present invention can be specific hardware on the device or software or firmware installed on the device. The implementation principle and technical effects of the device provided in the embodiment of the present invention are the same as those of the aforementioned method embodiment. For the sake of brief description, the parts not mentioned in the device embodiment can refer to the corresponding contents in the aforementioned method embodiment. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices and units described above can all refer to the corresponding processes in the aforementioned method embodiment, and will not be repeated here.
[0179] In the embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.
[0180] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0181] In addition, each functional unit in the embodiment provided by the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
[0182] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.
[0183] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.
[0184] Finally, it should be noted that the above embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention. They should all be covered within the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.
Claims
1. A network resource management method, characterized in that: include: Divide the transmission network of the target switch into multiple network slices based on the initial network configuration policy; Obtain the bandwidth requirement of the network data packet to be transmitted on the target switch, the bandwidth of the network slice in the unoccupied state, and the future network status; Determine whether the initial network configuration strategy needs to be updated according to the bandwidth requirement of the network data packet to be transmitted, the bandwidth of the network slice in the non-occupied state, and the future network state; If necessary, the bandwidth requirement of the network data packet to be transmitted, the bandwidth of the network slice in the unoccupied state and the future network state are input into the trained network policy generation model to generate a network configuration update strategy; The network configuration update strategy includes a network slice configuration strategy and a data transmission strategy; The training process of the network strategy generation model includes: Acquire data to be trained; the data to be trained includes bandwidth demand of historical transmission data, bandwidth of network slices in a current non-occupied state corresponding to the historical transmission data, and future network state corresponding to the historical transmission data; The data to be trained includes the bandwidth demand of historical transmission data, the network slice bandwidth in the current non-occupied state corresponding to the historical transmission data, and the future network state corresponding to the historical transmission data, and is divided into a training set and a test set; Inputting the training set into the network strategy generation model to be trained for forward noise processing; Performing reverse denoising iterative training based on the forward denoised data; When the preset number of iterative training times is reached, each training result and test set is input into the quality assessment network for evaluation; Solving the optimal solution of the quality assessment network according to the optimization theory model; Determining the network strategy generation model corresponding to the optimal solution as the final network strategy generation model; Re-dividing the network slice based on the network slice configuration strategy and determining the target network slice; The data transmission strategy is sent to the target network slice so that the target network slice transmits the network data packet to be transmitted according to the data transmission strategy.
2. The method according to claim 1, characterized in that: The step of obtaining the bandwidth requirement of the network data packet to be transmitted on the target switch includes: Obtaining the traffic type and number of network data packets to be transmitted, wherein the traffic type includes at least one or a combination of text, voice, video and picture; Determine the bandwidth required for each traffic type of data packet according to the traffic type and the preset correspondence between the traffic type and the bandwidth; The bandwidth requirement of the network data packet to be transmitted is determined based on the bandwidth required by each data packet and the number of data packets.
3. The method according to claim 1, characterized in that The future network states of network slices that obtain non-occupied states include: Obtain the transmission rate of the token bucket of each non-occupied network slice and the transmission rate of the network data packet to be transmitted; The future network state of the network slice in the non-occupied state is determined based on the rates of the two, wherein if the transmission rate of the token bucket is greater than the transmission rate of the network data packet to be transmitted, the future network state is smooth, otherwise, the future network state is congested.
4. The method according to claim 1 or 2, characterized in that: The bandwidth of the network slice in the non-occupied state includes the bandwidth of a single network slice and the total bandwidth of all network slices in the non-occupied state, and judging whether the initial network configuration strategy needs to be updated according to the bandwidth requirement of the network data packet to be transmitted, the bandwidth of the network slice in the non-occupied state and the future network state includes: Determine whether the total bandwidth of all the network slices in the non-occupied state is greater than or equal to the bandwidth requirement of the network data packet to be transmitted; If the total bandwidth is greater than or equal to the bandwidth requirement of the network data packet to be transmitted, then determining whether the maximum bandwidth in the network slice in the non-occupied state is less than the bandwidth requirement of the network data packet to be transmitted; If the maximum bandwidth is less than the bandwidth requirement of the network data packet to be transmitted, it is necessary to update the initial network configuration strategy; If the maximum bandwidth is greater than the bandwidth requirement of the network data packet to be transmitted, determining whether the future network state in the non-occupied network slice corresponding to the maximum bandwidth is a congested state; If so, the initial network configuration strategy needs to be updated.
5. The method according to claim 1, characterized in that Determining the target network slice includes: In the re-divided network slices, compare the bandwidth of each network slice; The network slice with the largest bandwidth is selected as the target network slice.
6. The method according to claim 1, characterized in that The data transmission strategy includes a committed information rate and a peak information rate, and sending the data transmission strategy to the target network slice so that the target network slice transmits the network data packet to be transmitted according to the data transmission strategy includes: Get the transmission rate of the network data packet to be transmitted; Determine the transmission priority of the network data packet to be transmitted according to the transmission rate and the committed information rate and the peak information rate; The transmission priority is sent to the token bucket corresponding to the target network slice; the token bucket transmits the network data packet according to the transmission priority.
7. A network resource management device, characterized in that: The device comprises: A partitioning unit, configured to partition a transmission network of a target switch into a plurality of network slices based on an initial network configuration policy; An acquisition unit, used to acquire the bandwidth requirement of the network data packet to be transmitted on the target switch, the bandwidth of the network slice in the unoccupied state, and the future network state; A judging unit, configured to judge whether it is necessary to update the initial network configuration strategy according to the bandwidth requirement of the network data packet to be transmitted, the bandwidth of the network slice in the non-occupied state, and the future network state; A generating unit, configured to input the bandwidth requirement of the data packet to be transmitted, the bandwidth of the network slice in the non-occupied state and the future network state into the trained network policy generation model to generate a network configuration update strategy when the initial network configuration strategy needs to be updated; the network configuration update strategy includes a network slice configuration strategy and a data transmission strategy; The training process of the network strategy generation model includes: Acquire data to be trained; the data to be trained includes bandwidth demand of historical transmission data, bandwidth of network slices in a current non-occupied state corresponding to the historical transmission data, and future network state corresponding to the historical transmission data; The data to be trained includes the bandwidth demand of historical transmission data, the network slice bandwidth in the current non-occupied state corresponding to the historical transmission data, and the future network state corresponding to the historical transmission data, and is divided into a training set and a test set; Inputting the training set into the network strategy generation model to be trained for forward noise processing; Performing reverse denoising iterative training based on the forward denoised data; When the preset number of iterative training times is reached, each training result and test set is input into the quality assessment network for evaluation; Solving the optimal solution of the quality assessment network according to the optimization theory model; Determining the network strategy generation model corresponding to the optimal solution as the final network strategy generation model; A re-dividing unit, used to re-divide the network slice based on the network slice configuration policy and determine the target network slice; A sending unit is used to send the data transmission strategy to the target network slice, so that the target network slice transmits the network data packet to be transmitted according to the data transmission strategy.
8. An electronic device, characterized in that: It includes a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus; Memory, used to store computer programs; A processor, for implementing the method steps described in any one of claims 1 to 6 when executing a program stored in a memory.
9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method steps described in any one of claims 1 to 6 are implemented.
