Tuning method and apparatus based on dynamic tuning threshold, and computing device
By introducing a fuzzy inference system and tuning algorithm into the network, dynamically adjusting the network shunt ratio and path, the problems of poor adaptability and high manual tuning cost of static threshold tuning methods are solved, and more efficient network tuning and Improved business experience.
Patent Information
- Application Number
- PCT/CN2024/090639
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-04
- Filing Date
- 2024-04-29
- Publication Date
- 2024-11-07
AI Technical Summary
The existing static threshold trigger tuning methods lack network adaptability, resulting in invalid tuning, and the manual trigger tuning methods have high operational expenditure and poor replicability.
A dynamic tuning threshold method based on fuzzy inference system and tuning algorithm is adopted to generate dynamic tuning thresholds based on real-time network traffic information, and the network shunt ratio and path are automatically adjusted to reduce the number of invalid tuning times and reduce network congestion.
It realizes network adaptability for network tuning, reduces the number of invalid tuning times and operation and maintenance costs, improves business experience, and is suitable for different network states.
Smart Images

Figure CN2024090639_07112024_PF_FP_ABST
Abstract
Description
A tuning method, device and computing equipment based on dynamic tuning threshold
[0001] This application claims priority to the Chinese patent application filed with the China Patent Office on May 4, 2023, with application number 202310490448.5 and application name “A tuning method, device and computing device based on dynamic tuning threshold”, the entire contents of which are incorporated by reference into this application. Technical Field
[0002] The present application relates to the field of communication technology, and in particular to a tuning method, apparatus, and computing device based on a dynamic tuning threshold. Background Art
[0003] To achieve load balancing, commonly used network tuning methods include static threshold-triggered tuning, where the user sets a link congestion threshold based on experience. When link utilization (currently carried link traffic / link capacity) exceeds the threshold, the tuning model in the control device is triggered. Other commonly used network tuning methods include manually triggered tuning, where network operations personnel manually adjust services based on network conditions (such as alarms) and operational experience.
[0004] Among them, the static threshold-triggered tuning method does not take into account the actual network traffic conditions, and there are large differences between different layers and different dedicated lines. Therefore, the fixed static threshold makes this method less adaptable to the entire network, leading to problems such as ineffective tuning. The manually triggered tuning method has high operating expenses, relies on manual experience, and has poor replicability.
[0005] Summary of the Invention
[0006] The present application provides a tuning method, apparatus, and computing device based on a dynamic tuning threshold. By coupling a fuzzy inference system with a tuning algorithm, a corresponding dynamic tuning threshold is generated based on real-time information such as network traffic. The control device determines whether to perform network tuning based on the dynamic tuning threshold and the acquired data. The network diversion ratio and path are tuned based on the network traffic information and the dynamic tuning threshold, which can reduce the number of invalid tuning times and reduce congestion.
[0007] In a first aspect, the present application provides a tuning method based on a dynamic tuning threshold, which is applied to a control device. The method includes: the control device obtains one or more real-time data such as link traffic, diversion ratio, path, segment list, packet loss information, and service experience data sent by the network device. Thereafter, the control device calls a fuzzy inference system to generate a first tuning threshold based on the obtained real-time data. When the real-time data triggers the first tuning threshold, the control device uses a tuning algorithm to perform network tuning, adjust the diversion ratio of the path, or split the path as needed.
[0008] The above-mentioned tuning method based on dynamic tuning threshold generates a dynamic tuning threshold suitable for the current network situation by controlling the device based on the real-time link traffic and other data obtained, and performs network tuning according to the dynamic tuning threshold. It has the adaptability of the entire network, can reduce the number of invalid tuning and the overall tuning number, and reduce network congestion and packet loss.
[0009] In one possible implementation, the control device invokes a fuzzy inference system to generate a first tuning threshold. The fuzzification interface in the fuzzy inference system uses a membership function to convert real-time data into fuzzy inputs. The inference engine, fuzzy rules, and fuzzy inputs generate a control variable. Finally, the defuzzification interface converts the control variable into a clear control variable, which is the first tuning threshold output in the form of data. Because data such as link traffic is highly uncertain, and users have higher requirements for network management, scheduling, and load balancing, the control device leverages the fuzzy inference system's ability to adaptively adjust parameters based on uncertain networks to generate dynamic tuning thresholds. This allows for more appropriate tuning thresholds, facilitates network load balancing, and achieves a better service experience.
[0010] In one possible implementation, before performing network optimization based on a first optimization threshold generated in real time, the control device first updates the optimization threshold corresponding to the optimization algorithm based on the first optimization threshold, so that the optimization algorithm updates the diversion ratio and path when data such as real-time link traffic reaches the first optimization threshold. In this process, the control device uses the generated dynamic optimization threshold to update the optimization threshold corresponding to the optimization algorithm, making the optimization algorithm adaptable to different real-time network states and generating optimization solutions such as diversion ratios and paths that are more consistent with the real-time network state. This can reduce the number of invalid optimizations and achieve better optimization results.
[0011] In one possible implementation, after generating a first split ratio and a first path through network tuning, the control device inputs the first split ratio and the first path into a fuzzy inference system, uses the fuzzy inference system to perform function fitting, and uses a gradient descent algorithm to update the parameters of the fuzzy inference system, including updating fuzzy rules and membership functions. This process couples the fuzzy inference system and the tuning algorithm, allowing for interactive and iterative updates, enabling the fuzzy inference system to generate more accurate dynamic tuning thresholds.
[0012] In a second aspect, the present application provides a tuning method based on a dynamic tuning threshold, which is applied to network devices and control devices. The method includes: the network device calls a fuzzy inference system based on one or more real-time data such as link traffic, diversion ratio, path, segment list, packet loss information, and service experience data to generate a first tuning threshold, and sends the first tuning threshold to the control device. When the real-time data triggers the first tuning threshold, the control device uses a tuning algorithm to perform network tuning, adjust the diversion ratio of the path, or split the path as needed.
[0013] If network devices have strong computing capabilities, they can directly generate dynamic tuning thresholds based on acquired real-time data and report these thresholds to the control device, allowing the control device to perform network optimization based on these thresholds. This process fully utilizes the computing power of network devices, reduces the overhead of sending real-time data to the control device, and conserves computing and storage resources on the control device.
[0014] In one possible implementation, the network device calls the fuzzy inference system to generate a first tuning threshold. The fuzzification interface in the fuzzy inference system uses a membership function to convert real-time data into fuzzy input, and uses the inference engine, fuzzy rules and fuzzy input to generate a control quantity. Finally, the control quantity is converted into a clear control quantity through the defuzzification interface, which is the first tuning threshold output in the form of data.
[0015] In one possible implementation, when the control device generates a first diversion ratio and a first path through network tuning, the network device inputs the first diversion ratio and the first path into a fuzzy inference system, performs function fitting on them using the fuzzy inference system, and updates the parameters of the fuzzy inference system using a gradient descent algorithm, including updating fuzzy rules and membership functions.
[0016] In a third aspect, the present application provides a tuning device based on a dynamic tuning threshold, which includes an acquisition module, a threshold generation module, a tuning module, and a sending module. The acquisition module is used to obtain first data sent by a network device, wherein the first data includes one or more of real-time link traffic, diversion ratio, path, segment list, packet loss information, and service experience data; the threshold generation module is used to generate a first tuning threshold that conforms to the real-time network status based on the first data; the tuning module is used to perform network tuning when the first data touches the first tuning threshold, and generate a first diversion ratio and a first path; the sending module is used to send the first diversion ratio and the first path to the network device.
[0017] In one possible implementation, the modules included in the tuning device based on dynamic tuning thresholds can be deployed in the same computing device, or when the network device has strong computing capabilities, the tuning module and the sending module can be deployed in the control device, and the acquisition module and the threshold generation module can be deployed in the network device.
[0018] In a fourth aspect, the present application provides a computing device comprising at least one processor and a memory, wherein the memory stores instructions, and when the instructions are executed by at least one processor, the at least one processor executes the method provided in the first aspect.
[0019] In a fifth aspect, the present application provides a computer-readable storage medium, which includes computer program instructions. When the computer program instructions are executed by a computing device cluster, the computing device cluster executes the method provided in the first aspect.
[0020] Based on the implementation methods provided in the above aspects, this application can also be further combined to provide more implementation methods. BRIEF DESCRIPTION OF THE DRAWINGS
[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following briefly introduces the drawings required for describing the embodiments.
[0022] FIG1 is a schematic diagram of a static threshold tuning method provided in an embodiment of the present application;
[0023] FIG2 is a schematic diagram of the structure of a tuning system based on a dynamic tuning threshold according to an embodiment of the present application;
[0024] FIG3 is a schematic diagram of a process of a tuning method based on a dynamic tuning threshold provided in an embodiment of the present application;
[0025] FIG4 is a schematic diagram of the structure of a fuzzy inference system provided in an embodiment of the present application;
[0026] FIG5 is a schematic structural diagram of a joint tuning framework for a fuzzy inference system adaptively adjusted with a network according to an embodiment of the present application;
[0027] FIG6 is a schematic structural diagram of a tuning device based on a dynamic tuning threshold according to an embodiment of the present application;
[0028] FIG7 is a schematic structural diagram of a control device provided in an embodiment of the present application;
[0029] FIG8 is a schematic structural diagram of a network device provided in an embodiment of the present application. DETAILED DESCRIPTION
[0030] Network congestion is a persistently overloaded network state in which the network equipment's demand for network resources (such as link bandwidth, storage space, and processor processing power) exceeds its inherent processing power and capacity. Network congestion can cause network transmission performance to degrade, leading to data loss, increased latency, and other problems.
[0031] At present, in order to avoid network congestion, commonly used network tuning methods include static threshold trigger tuning methods or manual trigger tuning methods. As shown in Figure 1, Figure 1 is a schematic diagram of a static threshold tuning method provided by an embodiment of the present application. In the static threshold trigger tuning method, the static threshold is determined based on an empirical value and can be the same for the entire network. For example, the static threshold for the entire network in (a) is 0.8. It can also be set based on different links. For example, the static threshold in (b) can be 0.8 and 0.7. The static threshold remains unchanged after determination. When the link utilization (currently carried link traffic / link capacity) exceeds the link static threshold, the tuning algorithm in the control device is triggered to perform (a) diversion ratio adjustment, or (b) path splitting on demand. Alternatively, in the method of manually triggering tuning, the network operation and maintenance personnel manually adjust the diversion ratio and path based on network conditions (such as alarm information, etc.) and operation and maintenance experience, thereby reducing network congestion and avoiding packet loss caused by network congestion, business damage, etc. The above static threshold-triggered tuning method does not take into account the real-time network traffic conditions, and there are large differences between different layers and different dedicated lines. The fixed static threshold leads to poor adaptability of the method across the entire network, resulting in multiple invalid tuning issues. The above manual triggering tuning method has high operating expenses and relies on manual experience, making it less replicable.
[0032] Therefore, the present application provides a tuning method based on a dynamic tuning threshold. By coupling a fuzzy inference system with a tuning algorithm, the fuzzy inference system in the control device generates a dynamic tuning threshold based on information such as the current link traffic, diversion ratio, and path reported by the network device. The tuning algorithm calculates based on the dynamic tuning threshold and the current link traffic, and updates the diversion ratio and path of the link. Compared with existing tuning methods, the above method triggers the tuning algorithm according to different tuning thresholds when the information reported by the network device is different. It has network-wide adaptability, can reduce the number of invalid tuning and overall tuning, and can adjust the diversion ratio and path according to real-time traffic without manual work, which can reduce operation and maintenance costs. The above tuning method based on a dynamic tuning threshold provided by the present application is applied to the system shown in Figure 2.
[0033] As shown in Figure 2, Figure 2 is a structural diagram of a tuning system based on a dynamic tuning threshold provided in an embodiment of the present application. The system includes a user device 210, a control device 220 and a network device 230, wherein the user device, the control device and the network device are connected through a network.
[0034] In a specific implementation, the user device can be an electronic device with data transmission and reception capabilities and display functions, such as a mobile phone, a tablet computer, or a desktop computer, and this application does not specifically limit this. The user device is used to visualize the tuning method based on the dynamic tuning threshold and display the data or image sent by the control device via the network.
[0035] In a specific implementation, the control device is a device that has both computing and storage capabilities. It can be a physical server, such as an X86 server, an ARM server, etc., or a virtual machine (VM) based on a general physical server combined with network function virtualization (NFV) technology. This application does not make specific restrictions on this. The control device is used to execute the Simple Network Management Protocol (SNMP) application and provide network management functions. In this application, the control device couples the fuzzy inference system with the tuning algorithm, which can be used to generate dynamic tuning thresholds and tuning solutions including information such as diversion ratios and paths.
[0036] In a specific implementation, the network device may be a host, router, or other device, which is used to store and forward data based on information such as the diversion ratio and path sent by the control device. It may also be used to send one or more of the information such as real-time link traffic, diversion ratio, path, segment list, packet loss information, and service experience data to the control device. The network device may also send more data, which is not specifically limited in this application.
[0037] In one possible implementation, when the network device has strong computing capabilities, the network device can execute some of the steps executed by the control device, and the generated dynamic tuning threshold can reduce the overhead of sending data on the network device and save storage space and computing resources of the control device.
[0038] As shown in FIG3 , FIG3 is a flowchart of a tuning method based on a dynamic tuning threshold provided in an embodiment of the present application. The method is applied to the system shown in FIG2 and includes the following steps.
[0039] Step S310: The control device obtains first data sent by the network device, wherein the first data is changing real-time data.
[0040] The control device sends collection-related parameter configuration to the network device, so that the network device sends first data to the control device, wherein the first data includes one or more of real-time link traffic, diversion ratio, path, segment list, packet loss information and service experience data.
[0041] The segment identity document-list (SID-list) in the first data is an ordered list composed of one or more instruction identifiers (SID). Segment is the instruction to be executed by the node for the received data packet. All paths in the network can be determined according to the SID-list. The service experience data in the first data includes indicators such as the number of video freezes and the cumulative time. Combined with the packet loss information, it is applied to the tuning method of the dynamic tuning threshold, which can be used for business network linkage and more accurately determine the dynamic tuning threshold. The first data can also include more types of data, which are not specifically limited in this application.
[0042] When the number of network devices is small, the control device can collect the first data from the network devices via the Simple Network Management Protocol (SNMP). When there are a large number of network devices, the control device can collect the first data from the network devices using network telemetry. The network devices can automatically transmit the first data to the control device, resulting in a shorter collection interval and more real-time and accurate first data, which is conducive to generating a more accurate dynamic threshold. The control device can also use more methods to obtain the first data, which are not specifically limited in this application.
[0043] In one possible implementation, the control device receives a startup instruction from the user device and an initialization parameter setting instruction for calculating the dynamic tuning threshold before sending the acquisition-related parameter configuration to the network device. If the user device does not send the initialization parameters, the system default initialization parameters are called.
[0044] In one possible implementation, after obtaining the first data, the control device also processes the first data and analyzes the statistical characteristics of the first data, for example, performing traffic regularization on the link traffic. In the case of a fuzzy inference system used by the control device in subsequent processes, the calculation effect of the regularized data is better, and a more accurate dynamic tuning threshold can be generated.
[0045] Step S320: The control device generates a first tuning threshold according to the first data.
[0046] The control device calls the fuzzy inference system according to the acquired first data to calculate and generate the first tuning threshold.
[0047] As shown in FIG4 , FIG4 is a schematic diagram of the structure of a fuzzy reasoning system provided in an embodiment of the present application. The fuzzy reasoning system 400 includes a fuzzification interface 410, a rule base 420, an inference engine 430, and a defuzzification interface 440. The fuzzification interface is used to convert real input into fuzzy input that can be used by the system by constructing a membership function. In the present application, the first data is converted into the corresponding fuzzy input. The rule base includes a variety of rules established based on expert experience. The rules can be given manually or obtained through learning. In the fuzzy reasoning process, the rules are provided for the inference engine. For example, in the present application, rules for generating dynamic tuning thresholds are provided. The inference engine is used to complete fuzzy reasoning and generate control quantities based on the fuzzy input and the rules provided by the rule base. The defuzzification interface is used to convert the control quantities generated by the inference engine into clear control quantities that can be understood and applied, that is, to generate the first tuning threshold represented in numerical form.
[0048] Because existing technologies for network control and scheduling are based on deterministic network status and traffic flow obtained through monitoring a single time slice, and the technical theories employed are deterministic and precise solutions based on graph theory and operations research, the static thresholds used in existing technologies lack adaptability. Currently, users have higher requirements for network control and scheduling, seeking a faster and better experience. Furthermore, data such as network traffic and network status are subject to significant uncertainty. Control devices utilize fuzzy inference systems to achieve adaptive parameter adjustment based on uncertain networks, generating dynamic tuning thresholds based on the acquired data. This method can generate tuning thresholds that are suitable for real-time network conditions, thereby achieving network-wide adaptability in the tuning method and achieving better network control and scheduling results.
[0049] In one possible implementation, when the network device has strong computing power, the network device can execute the above steps S310 and S320. The network device processes the first data and generates a first tuning threshold using a fuzzy inference system. This can fully utilize the computing power of the device, reduce the overhead of sending data on the device, and save storage space and computing resources of the control device.
[0050] In one possible implementation, the control device sends the generated first tuning threshold and the reasoning process for generating the first tuning threshold (for example, the membership function image used in the reasoning process, the control quantity generated by the inference engine, etc.) to the user device to realize the visualization of the reasoning process.
[0051] Step S330: The control device determines whether the first data reaches a first tuning threshold. If the first data does not reach the first tuning threshold, the control device executes step S340; if the first data reaches the first tuning threshold, the control device executes step S350.
[0052] Step S340: The control device keeps the current diversion ratio and path unchanged.
[0053] When the first data does not reach the first tuning threshold, the tuning algorithm in the control device does not need to generate a new split ratio and path, and maintains the current split ratio and path.
[0054] Step S350: The control device performs network optimization and sends the generated first split ratio and first path to the network device.
[0055] Before performing network optimization, the control device updates the optimization threshold corresponding to the optimization algorithm to a first optimization threshold and determines whether to perform network optimization based on the first data and the first optimization threshold. Compared to static threshold optimization methods, the control device continuously updates the optimization threshold of the optimization algorithm, making the optimization algorithm adaptable to different network conditions, reducing the number of invalid optimizations, and alleviating congestion.
[0056] When the first data reaches a first tuning threshold, the control device performs network tuning using a tuning algorithm, generates a first split ratio and a first path, and sends them to the network device.
[0057] In one possible embodiment, when the tuning algorithm is coupled with a fuzzy inference system that generates a first tuning threshold, the control device updates the fuzzy rules and membership functions in the fuzzy inference system according to the first diversion ratio and the first path after generating the first diversion ratio and the first path. As shown in Figure 5, Figure 5 is a structural diagram of a joint tuning framework for a fuzzy inference system with network adaptive adjustment provided by an embodiment of the present application. First, the fuzzy inference system is initialized and set by the user, or a dynamic tuning threshold is generated using default parameters, and the dynamic tuning threshold is used to update the tuning threshold corresponding to the tuning algorithm. The tuning algorithm performs network tuning according to the dynamic tuning threshold, solves the diversion ratio and path adjustment problem, and generates the optimal solution, the first diversion ratio and the first path. The control device inputs the first diversion ratio and the first path into the fuzzy inference system, causes the fuzzy inference system to perform function fitting on the first diversion ratio and the first path, updates the fuzzy rules and membership functions using the gradient descent method, generates a new dynamic tuning threshold, repeats the above steps, and performs multiple iterations until the maximum number of cycles or the accuracy requirement of the optimal solution is reached.
[0058] The above method couples the tuning algorithm with the fuzzy inference system, which can solve the subjectivity and uncertainty problems caused by the fuzzy inference system relying on personal experience and repeated trial and error in membership functions and related parameters. It can enable the fuzzy inference system to generate more accurate dynamic tuning thresholds and achieve better tuning effects.
[0059] In one possible implementation, after generating the first split ratio and the first path, the control device predicts the traffic flow of different paths at the next moment based on the first split ratio and the first path. The control device may use the predicted traffic flow information as the first data obtained and repeat the above steps to perform optimization. This process allows the control device to perform optimization in advance, thereby reducing congestion.
[0060] The network device transmits data according to the received first split ratio and first path, and continues to report real-time data such as link traffic to the control device at the next moment. The control device repeats the above steps to continue network optimization.
[0061] In one possible implementation, the above-mentioned dynamic tuning threshold-based tuning method provided by the present application is applicable to network tuning scenarios based on segment routing over IPv6 (SRv6) policies for Internet Protocol version 6 (IPv6), as well as mobile networks and any other scenarios where threshold-triggered tuning is used. For example, in a mobile network, traffic-based channel adjustment can calculate a dynamic tuning threshold based on traffic flow, channel information, and the like, and trigger channel adjustment when traffic reaches the dynamic tuning threshold.
[0062] In one possible implementation, the fuzzy inference system in the control device can be coupled with the tuning algorithm to continuously update the fuzzy rules and membership functions to calculate the dynamic tuning threshold. It can also be independent of the tuning algorithm and calculate the dynamic tuning threshold based on expert experience or manually set fuzzy rules and membership functions. This application does not make specific limitations on this.
[0063] At present, the link tuning threshold in the static threshold triggered tuning method is determined based on experience and is fixed. When the link load increases and reaches the link tuning threshold, the tuning is triggered, and the tuning algorithm in the control device updates the diversion ratio or path. In the case of low network adjustability, the above tuning process may be invalid tuning, and as the load increases, there will be increased congestion, packet loss, etc. Therefore, it is necessary to generate a tuning method suitable for different network conditions. Since the fuzzy inference system converts expert experience into fuzzy language rules, the generalization ability is less affected by data and has strong robustness and adaptability. Therefore, the present application calls the fuzzy inference system in the control device or network device, which can achieve the effect of controlling the tuning threshold according to complex data such as uncertain link traffic, thereby achieving better network management and control effects.
[0064] In summary, the present application provides a tuning method based on a dynamic tuning threshold. By coupling the fuzzy reasoning system and the tuning algorithm, the fuzzy reasoning system continuously updates the fuzzy rules and membership functions, and generates a more accurate dynamic tuning threshold based on real-time link traffic and other information, so that the tuning algorithm performs network tuning according to the dynamic tuning threshold, and generates a tuning plan for iteration of the fuzzy reasoning system, so that the tuning method is theoretical and applicable to different network states, reducing the number of invalid tuning times, reducing congestion, and reducing traffic packet loss caused by business disorder.
[0065] In one specific implementation, a fuzzy inference system dynamically generates thresholds for optimization based on real-time traffic data. A control device receives a startup command from a user device over the network and begins calculating the dynamic optimization thresholds. The control device sends collection-related parameter configurations to network devices. Using a network telemetry subscription method, the receiving network devices automatically report each tunnel's SID list, real-time link traffic information, tunnel SID list paths, and traffic split ratios to the control device at regular intervals.
[0066] Based on the information received from the network devices, the control device invokes the fuzzy inference system to generate a first dynamic tuning threshold. The control device then sends a visualization of the membership function, fuzzy rules, and the first dynamic tuning threshold used in the fuzzy inference process to the user device, visualizing the fuzzy inference process. Furthermore, the user device confirms the reported first dynamic tuning threshold, prompting the control device to send the first dynamic tuning threshold to the network device. If the link utilization (link real-time traffic / link capacity) exceeds the first dynamic tuning threshold, the control device invokes the optimization algorithm to perform network optimization, updates the split ratio and path, sends the updated split ratio and path to the network device, and updates the membership function and fuzzy rules based on the updated split ratio and path. After the control device sends the updated split ratio and path to the network device, the network device continues to report information such as the link real-time traffic to the control device, repeating the above steps.
[0067] After multiple network optimizations, the control device counts the number of optimizations and congestion conditions corresponding to the dynamic optimization threshold optimization method within the current period or a certain period, generates visual data or charts, and sends them to the user device. The control device can also send the number of optimizations and congestion conditions corresponding to the static threshold optimization method to the user device for comparison with the optimization method provided in this application.
[0068] In other specific implementations, the control device may also generate a dynamic tuning threshold based on tunnel packet loss information, or based on tunnel packet loss information and service experience data, which is not specifically limited in this application.
[0069] As shown in FIG6 , FIG6 is a tuning device based on a dynamic tuning threshold provided by the present application. The tuning device 600 can be applied to the system shown in FIG2 . The device includes an acquisition module 610, a threshold generation module 620, a tuning module 630, and a sending module 640. The acquisition module is used to obtain first data sent by a network device, wherein the first data includes one or more of real-time link traffic, diversion ratio, path, segment list, packet loss information, and service experience data; the threshold generation module is used to generate a first tuning threshold that conforms to the real-time network status based on the first data; the tuning module is used to perform network tuning when the first data touches the first tuning threshold and generate a first diversion ratio and a first path; and the sending module is used to send the first diversion ratio and the first path to the network device.
[0070] The acquisition module is used to execute step S310 of the method shown in Figure 3, the threshold generation module is used to execute step S320 of the method shown in Figure 3, the tuning module is used to execute steps S330 to S350 of the method shown in Figure 3, and the sending module is used to execute step S350 of the method shown in Figure 3. The above modules can be deployed in the control device in Figure 2, or can be deployed separately in the control device and the network device, which is not specifically limited in this application.
[0071] As shown in FIG7 , FIG7 is a schematic diagram of the structure of a control device provided by the present application, wherein the control device 220 includes a processor 710, a memory 720, a communication interface 730, and a bus 740. The processor, memory, and communication interface may be interconnected via an internal bus or may communicate via wireless transmission or other means.
[0072] The processor 710 can be composed of at least one general-purpose processor, such as a central processing unit (CPU), or a combination of a CPU and a hardware chip. The hardware chip can be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The PLD can be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof. The processor 710 is used to execute various types of digital storage instructions. The processor can implement any function of the module in the device shown in Figure 6 by executing corresponding instructions.
[0073] Memory 720 can be a volatile memory, such as random access memory (RAM), dynamic RAM (DRAM), static RAM (SRAM), synchronous dynamic RAM (SDRAM), double data rate RAM (DDR), cache, etc. Memory can also include a combination of the above. Memory 720 can include all modules in the tuning device 600 based on dynamic tuning thresholds, or only the tuning module and the sending module. Processor 710 can execute the steps in FIG. 3 and implement the functions in FIG. 6 by executing the program code corresponding to each module. The memory can also store data such as the generated dynamic tuning threshold, split ratio, and path.
[0074] The communication interface 730 can be used to receive instructions issued by user equipment, or data reported by network equipment, etc., and can also send data such as dynamic tuning thresholds to the network equipment. This application does not make specific limitations on this.
[0075] It should be noted that FIG7 is only a possible implementation of the embodiment of the present application. In actual applications, the control device may also include more or fewer components, which is not limited here.
[0076] As shown in Figure 8, Figure 8 is a schematic diagram of the structure of a network device provided by the present application, wherein the network device 230 includes a processor 810, a memory 820, a communication interface 830, and a bus 840. Among them, the processor, memory, and communication interface can be interconnected via an internal bus or can communicate via means such as wireless transmission. The implementation methods of the processor 810, memory 820, and communication interface 830 are similar to those of the processor 710, memory 720, and communication interface 730 shown in Figure 7, and are not described in detail here.
[0077] If the network device has strong computing capabilities, the memory 820 may include a collection module and a threshold generation module. The network device does not need to report real-time link traffic and other data. The processor 810 can generate dynamic tuning thresholds through a fuzzy inference system by executing the program code corresponding to each module in the memory 820, and report them to the control device via the communication interface 830. This can reduce the network device's data transmission overhead and save storage space and computing resources of the control device. The memory 820 may also include a built-in transceiver module and other modules for receiving information such as dynamic tuning thresholds and tuning plans sent by the control device, as well as reporting real-time data such as link traffic, diversion ratio, and path to the control device. The memory may also include more or fewer modules, which are not specifically limited in this application.
[0078] It should be noted that FIG8 is only a possible implementation of an embodiment of the present application. In actual applications, the network device may also include more or fewer components, which is not limited here.
[0079] An embodiment of the present application further provides a computer-readable storage medium, in which instructions are stored. When the computer-readable storage medium is executed on a processor, the method flow shown in FIG3 is implemented.
[0080] An embodiment of the present application further provides a computer program product. When the computer program product runs on a processor, the method flow shown in FIG3 is implemented.
[0081] When the computer program instructions are loaded or executed on a computer, the processes or functions according to the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.
[0082] Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. A computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage node such as a server or data center that contains at least one set of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tape), optical media (e.g., high-density digital video discs (DVDs), or semiconductor media).
[0083] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and such modifications or substitutions are intended to be within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be subject to the scope of protection of the claims.
Claims
1. A tuning method based on a dynamic tuning threshold, applied to control equipment, It is characterized in that The method comprises: The control device acquires first data sent by the network device, wherein the first data includes one or more of real-time link traffic, split ratio, path, segment list, packet loss information, and service experience data; The control device generates a first tuning threshold according to the first data; The control device performs network optimization when the first data reaches the first optimization threshold, and sends the generated first split ratio and first path to the network device.
2. The method according to claim 1, It is characterized in that Generating a first tuning threshold according to the first data includes: The control device calls a fuzzy inference system to generate the first tuning threshold according to the first data, fuzzy rules related to the tuning threshold in the fuzzy inference system, and a membership function.
3. The method according to claim 1 or 2, It is characterized in that Before the control device performs network optimization when the first data reaches the first optimization threshold, the method further includes: The control device updates the tuning threshold corresponding to the tuning algorithm to the first tuning threshold, wherein the tuning algorithm is used to perform network tuning.
4. The method according to claim 3, It is characterized in that After sending the first split ratio and the first path to the network device, the method further includes: The control device updates the fuzzy rule and the membership function according to the first diversion ratio and the first path.
5. A tuning method based on dynamic tuning threshold, applied to network equipment and control equipment, the method include: The network device generates a first tuning threshold according to first data, wherein the first data includes one or more of real-time link traffic, split ratio, path, segment list, packet loss information, and service experience data; The network device sends the first tuning threshold to the control device; The control device performs network optimization when the first data reaches the first optimization threshold, and sends the generated first split ratio and first path to the network device.
6. The method according to claim 5, It is characterized in that The network device generating a first tuning threshold according to the first data includes: The network device calls a fuzzy inference system to generate the first tuning threshold according to the first data, fuzzy rules related to the tuning threshold in the fuzzy inference system, and a membership function.
7. The method according to claim 5 or 6, It is characterized in that After the network device receives the first split ratio and the first path sent by the control device, the method further includes: The network device updates the fuzzy rule and the membership function according to the first split ratio and the first path.
8. A tuning device based on a dynamic tuning threshold, It is characterized in that The device includes a collection module, a threshold generation module, a tuning module and a sending module: The acquisition module is used to obtain first data sent by the network device, wherein the first data includes one or more of real-time link traffic, split ratio, path, segment list, packet loss information and service experience data; The threshold generating module is used to generate a first tuning threshold according to the first data; The tuning module is configured to perform network tuning when the first data reaches the first tuning threshold, and generate a first split ratio and a first path; The sending module is used to send the first split ratio and the first path to a network device.
9. A computing device, It is characterized in that The computing device comprises a processor and a memory, wherein the memory stores a computer program, and the processor executes the computer program to enable the computing device to perform the method according to any one of claims 1 to 4.
10. A computer-readable storage medium, It is characterized in that The computer-readable storage medium stores a program, and when the program is executed on a computing device, the computing device is caused to execute the method according to any one of claims 1 to 4.
Citation Information
Patent Citations
On-chip optical network adaptive route planning method based on fuzzy logic
CN114598946A
Parameter tuning method and device and related equipment
CN114692878A
Method, device and system for determining network optimization strategy
CN115776443A
Fuzzy inference network for classification of high-dimensional data
US20030018592A1
