Network resource allocation method, apparatus, device, storage medium, and program product

Abstract

The application discloses a network resource allocation method, device, equipment, storage medium and program product, the method comprises the following steps: according to the index weight of multiple network performance indexes under the scene type to which the target scene belongs of the scene weight configuration indication, multiple key network performance indexes under the target scene are selected;The multiple key network performance indexes of each first network element under the target scene are monitored;Then, according to multiple key network performance indexes, the expansion and shrinkage of the first network element are obtained multiple second network elements;Then, according to the multiple network performance indexes of each second network element and the scene weight configuration, the weight of each second network element is adjusted, and each second network element is re-allocated network resources according to the adjusted weight;The application can improve the transmission efficiency, reliability of the whole SDN network, improve the network service quality, and improve the utilization rate of the SDN network resources.

Description

Technical Field

[0001] This invention relates to the field of network resource allocation technology, and in particular to a network resource allocation method, apparatus, device, storage medium, and program product. Background Technology

[0002] With the rapid development of network technology, SDN (Software Defined Networking) has been widely used due to its advantages such as centralized control and flexible programmability. SDN technology can dynamically distribute network traffic to devices such as bare metal servers, enabling more flexible network security for these devices. However, in an SDN network architecture, different network elements exhibit differences in performance in terms of bandwidth, transmission latency, and packet loss rate. Even the same network element can change dynamically at different times. Currently, traditional SDN network resource allocation methods mainly fall into two categories: one is to allocate traffic based on simple, fixed rules; the other is to perform load balancing based on static or semi-static configuration methods.

[0003] However, traditional SDN network resource allocation methods are often limited to single business scenarios or fixed network environments, adopting relatively fixed strategies and lacking the ability to collect multi-dimensional performance indicators in real time. This makes it difficult to adaptively adjust traffic allocation according to real-time performance changes of network elements, resulting in problems such as excessive transmission latency and packet loss that affect network service quality. Summary of the Invention

[0004] To address the problems existing in the prior art, embodiments of the present invention provide a network resource allocation method, apparatus, device, storage medium, and program product, which can realize dynamic adjustment of resource allocation, improve the transmission efficiency and reliability of SDN networks, enhance network service quality, and improve the utilization rate of SDN network resources.

[0005] In a first aspect, embodiments of the present invention provide a network resource allocation method, including:

[0006] Based on the preset scenario-based weight configuration, the indicator weights of multiple network performance indicators under the scenario type to which the target scenario belongs are indicated, and multiple key network performance indicators under the target scenario are selected.

[0007] Monitor multiple key network performance indicators of each first network element in the target scenario;

[0008] Based on multiple key network performance indicators, multiple second network elements are obtained by expanding or shrinking multiple first network elements;

[0009] Based on the multiple network performance indicators and the scenario-based weight configuration of each second network element, the weights of each second network element are adjusted, and network resources are redistributed to each second network element according to the adjusted weights; wherein, the weights of the second network elements are used to indicate the resource redistribution weights of the second network elements.

[0010] As an improvement to the above scheme, the step of obtaining multiple second network elements by expanding or shrinking multiple first network elements based on multiple key network performance indicators includes:

[0011] Detect whether multiple key network performance indicators meet preset network element expansion or network element reduction conditions;

[0012] Under the condition of network element expansion, network element addition processing is performed according to the expansion / shrinkage step size corresponding to the target scenario to obtain multiple second network elements; wherein, the second network element includes the first network element and the newly added network element;

[0013] If the network element scaling conditions are met, network element deletion is performed according to the scaling step size corresponding to the target scenario to obtain multiple second network elements; wherein, the second network elements include the remaining first network elements after the deletion of network elements.

[0014] As an improvement to the above scheme, the network element expansion conditions include: multiple key network performance indicators all exceed their respective upper thresholds; the network element reduction conditions include: multiple key network performance indicators all fall below their respective lower thresholds.

[0015] As an improvement to the above scheme, the step of adjusting the weights of each second network element based on multiple network performance indicators and the scenario-based weight configuration includes:

[0016] Calculate the sum of the first weights based on the original weights of each of the second network elements;

[0017] Based on the multiple network performance indicators and the scenario-based weight configuration of each second network element, calculate the multi-dimensional comprehensive weight of each second network element;

[0018] Based on the sum of the first weights and the multi-dimensional comprehensive weights of each second network element, the target weight of the corresponding second network element is calculated, and the original weights of each second network element are adjusted to the target weights.

[0019] As an improvement to the above scheme, the step of calculating the multi-dimensional comprehensive weight of each second network element based on multiple network performance indicators and the scenario-based weight configuration includes:

[0020] For each of the second network elements, multiple network performance indicators of the second network element are detected, and the values ​​of the detected multiple network performance indicators of the second network element are normalized to obtain normalized index values ​​of the multiple network performance indicators.

[0021] The weights of multiple network performance metrics under the scenario type to which the target scenario belongs are determined from the scenario-based weight configuration.

[0022] Based on the normalized index values ​​of multiple network performance indicators of each second network element and the index weights of the corresponding network performance indicators, the multi-dimensional comprehensive weight of each second network element is calculated.

[0023] As an improvement to the above scheme, the step of calculating the multi-dimensional comprehensive weight of each second network element based on the normalized index values ​​of multiple network performance indicators and the corresponding index weights of each network performance indicator includes:

[0024] Based on the weights of the multiple network performance indicators, the normalized values ​​of the multiple network performance indicators are weighted and summed to obtain the multi-dimensional comprehensive score of the second network element.

[0025] The multi-dimensional comprehensive scores of all second network elements are summed to obtain the total score.

[0026] The multi-dimensional comprehensive weight of each second network element is obtained based on its proportion in the total score.

[0027] As an improvement to the above scheme, the step of calculating the target weight of the corresponding second network element based on the sum of the first weights and the multi-dimensional comprehensive weights of each second network element includes:

[0028] The multi-dimensional comprehensive weights of all the second network elements are summed to obtain the total second weights.

[0029] The target weight of the second network element is calculated based on the proportion of the multi-dimensional comprehensive weight of each second network element in the total weight of the second weight and the total weight of the first weight.

[0030] As an improvement to the above scheme, the step of reallocating network resources to each of the second network elements according to the adjusted weights includes:

[0031] The original weight and target weight of each second network element are compared to determine the weight difference; wherein, the original weight is the weight of the corresponding second network element before adjustment, and the target weight is the weight of the corresponding second network element after adjustment.

[0032] If the weight difference of any second network element is not less than a preset difference threshold, the weight of each second network element is adjusted from its original weight to its target weight.

[0033] As an improvement to the above solution, in the target scenario where the SDN network does not integrate bare metal servers, the network performance indicators include: the SDN network's bps utilization, pps utilization, cps utilization, average latency, jitter rate, and packet loss rate.

[0034] In the target scenario of SDN network converged with bare metal server, the network performance indicators include: the SDN network's bps utilization, pps utilization, cps utilization, average latency, jitter rate, and packet loss rate, and the bare metal server's CPU utilization, memory bandwidth utilization, and network interface card queue utilization.

[0035] As an improvement to the above solution, in the target scenario of SDN network convergence with bare metal servers, the method further includes:

[0036] Monitor the CPU utilization and network interface card queue usage of each bare metal server;

[0037] If any of the bare metal servers meets the preset first traffic diversion condition, the network traffic of the second network element associated with the corresponding bare metal server is diverted to the second network element associated with other bare metal servers that also meet the preset second traffic diversion condition.

[0038] The first traffic splitting condition includes: the CPU utilization rate of the bare metal server exceeds a preset first utilization rate threshold or the network card queue utilization rate exceeds a preset first utilization rate threshold; the second traffic splitting condition includes: the CPU utilization rate of the bare metal server is less than a preset second utilization rate threshold or the network card queue utilization rate is less than a preset second utilization rate threshold.

[0039] As an improvement to the above solution, in the target scenario of SDN network convergence with bare metal servers, the method further includes:

[0040] Real-time monitoring of the rate of change of multiple network performance indicators for each of the second network elements;

[0041] Suspicious traffic is identified based on the rate of change of each of the aforementioned network performance indicators; wherein, the suspicious traffic is network traffic corresponding to at least one of the aforementioned network performance indicators having a rate of change exceeding a preset rate of change threshold within a first set time period;

[0042] The suspicious traffic is subjected to DoS attack traffic detection; wherein, the DoS attack traffic is network traffic that includes preset attack characteristics.

[0043] As an improvement to the above solution, the DoS attack traffic detection of the suspicious traffic includes:

[0044] Feature extraction is performed on the suspicious traffic to obtain the traffic characteristics of the suspicious traffic;

[0045] The traffic characteristics of the suspicious traffic are matched with a preset attack feature database; wherein, the attack feature database includes attack features of multiple DoS attack traffic templates.

[0046] If the traffic characteristics of the suspicious traffic match the attack characteristics of any of the DoS attack traffic templates, the suspicious traffic is determined to be DoS attack traffic.

[0047] As an improvement to the above solution, the method further includes:

[0048] Upon detecting DoS attack traffic, the weights of all second network elements and their associated bare metal servers in the target scenario are reset to zero, and the weights of multiple network performance indicators in the target scenario are increased.

[0049] If no DoS attack traffic is detected during the second time period, the weights of all the second network elements and their associated bare metal servers in the target scenario, as well as the weights of multiple network performance indicators, are reset.

[0050] In a second aspect, embodiments of the present invention provide a network resource allocation device, comprising:

[0051] The key indicator determination module is used to select multiple key network performance indicators under the target scenario based on the indicator weights of multiple network performance indicators under the scenario type of the target scenario according to the preset scenario-based weight configuration.

[0052] A network performance monitoring module is used to monitor multiple key network performance indicators of each first network element in the target scenario;

[0053] The network element expansion / shrinkage module is used to obtain multiple second network elements by expanding or shrinking multiple first network elements according to multiple key network performance indicators.

[0054] The dynamic weight adjustment module is used to adjust the weight of each second network element according to multiple network performance indicators and the scenario-based weight configuration, and to redistribute network resources of each second network element according to the adjusted weight; wherein, the weight of the second network element is used to indicate the resource redistribution weight of the second network element.

[0055] Thirdly, embodiments of the present invention provide a network resource allocation device, including: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor executes the computer program to implement the network resource allocation method as described in any one of the first aspects.

[0056] Fourthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the network resource allocation method as described in any one of the first aspects.

[0057] Fifthly, embodiments of the present invention provide a computer program product, including a computer program / instruction that, when executed by a processor, implements the network resource allocation method as described in any one of the first aspects.

[0058] Compared to existing technologies, this invention provides a network resource allocation method, apparatus, device, storage medium, and program product. This involves monitoring multiple key network performance indicators for each first network element in a target scenario; then, based on these indicators, expanding or shrinking the capacity of the first network elements to obtain multiple second network elements; subsequently, adjusting the weights of each second network element according to its multiple network performance indicators and scenario-based weight configuration, and redistributing network resources according to the adjusted weights. This invention, by real-time monitoring of multiple network performance indicators in an SDN network, first... SDN network elements are expanded or reduced in size to meet current business needs. Then, the weights of network elements in the SDN network are dynamically adjusted across multiple network performance indicators. By adjusting the number of network elements, the number of elements participating in resource allocation is aligned with the overall business requirements of the SDN network, improving overall transmission efficiency. Furthermore, dynamically adjusting the weights of network elements allows for the prioritization of resource allocation, enabling rational traffic routing, improving overall reliability, enhancing network service quality, and increasing resource utilization. Attached Figure Description

[0059] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0060] Figure 1This is a flowchart of a network resource allocation method provided in an embodiment of the present invention;

[0061] Figure 2 This is an overall schematic diagram of the adaptive dynamic adjustment of weights provided in an embodiment of the present invention;

[0062] Figure 3 This is a schematic diagram of the multi-dimensional dynamic weight adjustment process provided in an embodiment of the present invention;

[0063] Figure 4 This is a structural block diagram of a network resource allocation device provided in an embodiment of the present invention;

[0064] Figure 5 This is a structural block diagram of a network resource allocation device provided in an embodiment of the present invention. Detailed Implementation

[0065] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0066] It is understood that the various numerical designations used in the embodiments of this invention are merely for descriptive convenience and are not intended to limit the scope of this application. The order of the process numbers does not imply the order of execution; the execution order of each process should be determined by its function and internal logic.

[0067] In embodiments of the invention, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, without necessarily requiring or implying any such actual relationship or order between these entities or operations. The terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes said element. The term "a plurality or several" refers to two or more.

[0068] Please see Figure 1 , Figure 1 This is a flowchart illustrating a network resource allocation method provided in an embodiment of the present invention. The network resource allocation method can be executed by an SDN controller, and the method specifically includes:

[0069] S11: Select multiple key network performance indicators under the target scenario based on the indicator weights of multiple network performance indicators under the scenario type indicated by the preset scenario-based weight configuration.

[0070] S12: Monitor multiple key network performance indicators of each first network element in the target scenario;

[0071] The scenario-based weight configuration includes the weights of network performance indicators under multiple scenario types. Each network element (e.g., the first network element, the second network element) includes multiple network performance indicators. For each network element, several key network performance indicators under different scenario types can be selected from its multiple network performance indicators based on the weights of the network performance indicators under different scenario types. The network performance indicators include, but are not limited to, network indicators such as bps (Bits Per Second) utilization, pps (Packets Per Second) utilization, cps (Connections Per Second) utilization, average latency, jitter rate, and packet loss rate of the SDN network. Furthermore, the network performance indicators may also include hardware indicators. For example, in a scenario considering the integration of SDN networks with bare metal servers, the network performance indicators may also include hardware indicators such as CPU utilization, memory bandwidth utilization, and network card queue utilization. This embodiment of the invention does not impose specific limitations.

[0072] In this embodiment of the invention, the SDN network is pre-configured with scenario-based weights by the scenario-based weight configuration module of the SDN controller to adapt to the weight configuration of various network service scenarios, thereby guiding the collection of multi-dimensional indicators in different scenarios.

[0073] For example, suppose the total number of network elements in the current SDN network environment is n, that is, there are n network elements (such as the first network element) in the current SDN network environment, labeled as N1~N. n The SDN controller's collectors gather network performance metrics across multiple dimensions from n network elements, including six metrics such as bps utilization, pps utilization, cps utilization, average latency, jitter rate, and packet loss rate. A predefined normalized metric weight vector P = [p1, p2, p3, p4, p5, p6] is defined for different scenario types, where p1, p2, p3, p4, p5, and p6 correspond to the normalized metric weights for bps utilization, pps utilization, cps utilization, latency, jitter rate, and packet loss rate, respectively.

[0074] The scenario types include, but are not limited to, three typical application service scenario types: eMBB (Enhanced Mobile Broadband), mMTC (Massive Machine-Type Communications), and uRLLC (Ultra-Reliable Low-Latency Communications). These three scenario types basically cover all typical service scenarios in SDN networks. For the eMBB scenario, high bandwidth (bps) and high throughput (pps) are the core requirements to meet the high traffic demands of high-definition video, VR / AR, etc., with basic but not extreme requirements for latency and packet loss. For the mMTC scenario, supporting massive connections (cps) is the core requirement, focusing on connection capacity, with lower bandwidth (bps) requirements and less sensitivity to latency and jitter. For the mMTC scenario, ultra-low latency (average latency) and ultra-high reliability (packet loss rate, jitter rate) are the core requirements to meet the critical scenarios of industrial control, autonomous driving, etc., with lower requirements for bandwidth and connection count. Indicator weights are allocated based on the characteristics of different scenario types. The table below shows an example of the weight allocation of network performance metrics (i.e., scenario-based weight configuration) under different scenario types.

[0075]

[0076] S13: Based on multiple key network performance indicators, multiple second network elements are obtained by expanding or shrinking the capacity of multiple first network elements;

[0077] The second network element refers to the network element obtained after scaling up or down the network element in the SDN network environment. For example, the original first network element and the newly added network element in the SDN network environment; or the first network element remaining after deleting some of the first network elements in the SDN network environment.

[0078] The SDN controller's service-driven scaling module, based on preset scenario-based weight configurations, can select one or more key network performance indicators (KPIs) from the scenario type to which the target scenario belongs. For example, multiple network performance indicators with high weights can be monitored in real time as key KPIs, serving as trigger conditions for scaling mechanisms. In an eMBB scenario, for instance, bps utilization, pps utilization, and average latency can be selected as three key KPIs. The monitoring values ​​of these key KPIs determine whether network element scaling is necessary. This service-driven approach, using multiple key KPIs, more accurately matches service requirements and avoids resource waste or performance bottlenecks caused by fluctuations in a single KPI.

[0079] S14: Based on the multiple network performance indicators and the scenario-based weight configuration of each second network element, adjust the weight of each second network element, and reallocate network resources for each second network element according to the adjusted weight; wherein, the weight of the second network element is used to indicate the resource reallocation weight of the second network element.

[0080] It is understandable that the weight of a network element (such as the second network element) is used to measure the resource allocation weight of that network element. The higher the resource allocation weight, the more network resources that network element is allocated.

[0081] In this embodiment of the invention, the multi-dimensional dynamic weight adjustment module of the SDN controller monitors multiple network performance indicators of the second network element in real time, and combines the indicator weights of multiple network performance indicators under the scenario type of the target scenario in the scenario-based weight configuration to dynamically adjust the weight of the second network element itself. This achieves dynamic adjustment of resource allocation weights, and network resources are redistributed according to the dynamically adjusted weights of each second network element. This embodiment of the invention monitors multiple network performance indicators in the SDN network in real time, first expanding or shrinking the network elements in the SDN network to make the SDN network meet current business needs, and then dynamically adjusting the weights of the network elements in the SDN network based on multiple network performance indicators. By adjusting the number of network elements in the SDN network, the number of network elements participating in resource allocation meets the overall business needs of the SDN network, improving the overall transmission efficiency of the SDN network. Simultaneously, by dynamically adjusting the weights of network elements in the SDN network, the proportion and priority of network elements in resource allocation can be adjusted, thereby achieving reasonable guidance of network traffic, improving the overall reliability of the SDN network, enhancing network service quality, and increasing the utilization rate of SDN network resources.

[0082] As an optional embodiment, S11: Based on the indicator weights of multiple network performance indicators under the scenario type of the target scenario indicated by the preset scenario-based weight configuration, select multiple key network performance indicators under the target scenario, including:

[0083] Based on the preset scenario-based weight configuration, the weights of multiple network performance indicators under the scenario type to which the target scenario belongs are determined.

[0084] The network performance indicators are sorted from largest to smallest according to their weights, and the network performance indicators at the top of the sort are selected as the key network performance indicators in the target scenario.

[0085] In this embodiment of the invention, based on the indicator weights for each scenario type in the scenario-based weight configuration, a set number (e.g., 3) of network performance indicators can be selected as key network performance indicators for the corresponding scenario type. Subsequently, for different target scenarios, only the corresponding scenario type needs to be matched to find the key network performance indicators for that target scenario, which can then be used as conditions to trigger the scaling mechanism. For example, taking the scenario-based weight configuration shown in the table above as an example, if the target scenario type is eMBB, then bps utilization, pps utilization, and average latency are selected as key network performance indicators; if the target scenario type is mMTC, then bps utilization, pps utilization, and cps utilization are selected as key network performance indicators; if the target scenario type is uRLLC, then average latency, jitter rate, and packet loss rate are selected as key network performance indicators.

[0086] This invention selects multiple key network performance indicators for different scenario types as the basis for determining network element expansion and contraction, which can more accurately match the needs of business scenarios and avoid resource waste or performance bottlenecks caused by fluctuations in a single indicator.

[0087] As an optional embodiment, S12: Based on multiple key network performance indicators, multiple second network elements are obtained by expanding or shrinking the capacity of multiple first network elements, including:

[0088] Detect whether multiple key network performance indicators meet preset network element expansion or network element reduction conditions;

[0089] Under the condition of network element expansion, network element addition processing is performed according to the expansion / shrinkage step size corresponding to the target scenario to obtain multiple second network elements; wherein, the second network element includes the first network element and the newly added network element;

[0090] The network element expansion conditions include: multiple key network performance indicators all exceed their respective threshold limits;

[0091] When the network element scaling down condition is met, network element deletion is performed according to the scaling up / down step size corresponding to the target scenario to obtain multiple second network elements; wherein, the second network elements include the remaining first network elements after the deletion of network elements;

[0092] The network element scaling-down conditions include: multiple key network performance indicators are all less than their respective lower threshold limits.

[0093] In this embodiment of the invention, the scaling step size (i.e., the single adjustment amount, indicating the number of network elements added or deleted in a single operation) for different scenario types is preset. The specific value of the scaling step size is not specifically limited in this embodiment of the invention. For example, the scaling step sizes for eMBB, mMTC, and uRLLC scenarios are 2, 5, and 3, respectively, as shown in the table below.

[0094]

[0095] Based on the scenario type of the target scenario, key network performance indicators (KPIs) of each first network element in the corresponding SDN network are monitored. For each key network performance indicator, it is determined whether the value of the key network performance indicator exceeds or falls below its corresponding threshold upper limit. If the values ​​of all key network performance indicators exceed their respective threshold upper limits, it is determined that network element expansion is required, increasing the number of network elements. If the values ​​of all key network performance indicators are below their respective threshold lower limits, it is determined that network element reduction is required, deleting network elements. If the values ​​of key network performance indicators are between their corresponding threshold upper limits and the threshold lower limit, the system directly enters the multi-dimensional dynamic weight adjustment module to dynamically adjust the network element weights. Figure 2 As shown. Specifically, network elements are added or deleted in a step-by-step manner according to the expansion and contraction step size, until the value of the key network performance index of the second network element in the expanded and contracted SDN network does not exceed the upper limit of its respective threshold and is not lower than the lower limit of its respective threshold.

[0096] As an optional implementation, for each key network performance indicator, the average value of the corresponding key network performance indicator for all first network elements can be calculated. Then, it can be determined whether the average value of the key network performance indicator exceeds or falls below its corresponding threshold upper limit. If the average value of all key network performance indicators exceeds their respective corresponding threshold upper limits, it is determined that network element expansion is required, increasing the number of network elements. If the average value of all key network performance indicators is below their respective corresponding threshold lower limits, it is determined that network element reduction is required, decreasing the number of network elements.

[0097] In this invention, the lower and upper thresholds for various key network performance indicators may be the same or different under different scenario types. This embodiment of the invention does not limit the values ​​of the lower and upper thresholds for various key network performance indicators under different scenario types. It is understood that in this embodiment of the invention, the first network element refers to a network element in the SDN network environment before scaling up or down, and the second network element refers to a network element in the SDN network environment after scaling up or down.

[0098] For example, in an eMBB target scenario, if the bps utilization, pps utilization, and average latency of all first-level network elements all exceed their respective upper or lower thresholds, then scaling up or down two network elements is triggered. Similarly, in a mMTC target scenario, if the bps utilization, pps utilization, and cps utilization of all first-level network elements all exceed their respective upper or lower thresholds, then scaling up or down five network elements is triggered. And in a uRLLC target scenario, if the average latency, jitter rate, and packet loss rate of all first-level network elements all exceed their respective upper or lower thresholds, then scaling up or down three network elements is triggered. This business-driven approach to judging multiple key network performance indicators allows for more accurate matching of business needs, avoiding resource waste or performance bottlenecks caused by fluctuations in a single indicator.

[0099] As an optional embodiment, adjusting the weights of each second network element based on multiple network performance indicators and the scenario-based weight configuration includes:

[0100] Calculate the sum of the first weights based on the original weights of each of the second network elements;

[0101] For example, the sum of the original weights of all second network elements is calculated to obtain the current total weight, i.e., the first total weight, which is used for the weight allocation of subsequent network elements. The specific calculation is as follows:

[0102] (1);

[0103] in, Indicates the second network element The original weights; n represents the total number of network elements in the current SDN network environment, and here it represents the total number of the second network element; This represents the sum of the first weights.

[0104] Based on the multiple network performance indicators and the scenario-based weight configuration of each second network element, calculate the multi-dimensional comprehensive weight of each second network element;

[0105] Specifically, based on the multiple network performance indicators and the scenario-based weight configuration of each second network element, a multi-dimensional comprehensive weight is calculated for each second network element, including:

[0106] For each of the second network elements, multiple network performance indicators of the second network element are detected, and the values ​​of the detected multiple network performance indicators of the second network element are normalized to obtain normalized index values ​​of the multiple network performance indicators.

[0107] For example, the values ​​of various network performance indicators of the second network element are normalized using the following index normalization function to obtain the normalized index values ​​of the corresponding network performance indicators.

[0108] (2);

[0109] in, This represents the normalized index value of network performance index j for the second network element i. This represents the value of the network performance index j of the second network element i.

[0110] The weights of multiple network performance metrics under the scenario type to which the target scenario belongs are determined from the scenario-based weight configuration.

[0111] For example, the weights of each network performance indicator under the scenario type to which the target scenario belongs are obtained from the preset scenario-based weight configuration. For example, if the target scenario is a real-time conferencing scenario, which belongs to the eMBB scenario type, the weights of each network performance indicator p1, p2, p3, p4, p5, and p6 are 0.4, 0.25, 0.08, 0.12, 0.12, and 0.07, respectively.

[0112] Based on the normalized index values ​​of multiple network performance indicators of each second network element and the index weights of the corresponding network performance indicators, the multi-dimensional comprehensive weight of each second network element is calculated.

[0113] Specifically, based on the normalized index values ​​of multiple network performance indicators for each second network element and the corresponding index weights of the network performance indicators, a multi-dimensional comprehensive weight for each second network element is calculated, including:

[0114] Based on the weights of the multiple network performance indicators, the normalized values ​​of the multiple network performance indicators are weighted and summed to obtain the multi-dimensional comprehensive score of the second network element.

[0115] For example, the normalized index values ​​of the network performance indicators of each second network element calculated above. Combined with the network performance index weight p j Calculate the multi-dimensional comprehensive score S for each second network element. i The specific calculations are as follows:

[0116] (3);

[0117] in, This represents the weight of network performance metric j, where j m, where m represents the number of network performance metrics, for example, m=6. S iThis represents the multi-dimensional comprehensive score of the second network element i.

[0118] The multi-dimensional comprehensive scores of all second network elements are summed to obtain the total score.

[0119] The multi-dimensional comprehensive weight of each second network element is obtained based on its proportion in the total score.

[0120] For example, the specific calculation of the multi-dimensional comprehensive weight is as follows:

[0121] (4);

[0122] in, q represents the multi-dimensional comprehensive weight of the second network element i, and q represents the normalized comprehensive score index that integrates multiple dimensions of network performance indicators and has a scenario-specific tendency.

[0123] Based on the sum of the first weights and the multi-dimensional comprehensive weights of each second network element, the target weight of the corresponding second network element is calculated, and the original weights of each second network element are adjusted to the target weights.

[0124] Specifically, the multi-dimensional comprehensive weights of all the second network elements are summed to obtain the total second weights;

[0125] The target weight of the second network element is calculated based on the proportion of the multi-dimensional comprehensive weight of each second network element in the total weight of the second weight and the total weight of the first weight.

[0126] For example, the target weight of each second network element is obtained by calculating the sum of the second weights of the multidimensional comprehensive weights of each second network element obtained above, and by multiplying the ratio of the multidimensional comprehensive weight of each second network element to the sum of the second weights and the sum of the first weights.

[0127] (5);

[0128] in, This represents the final calculated second network element. The target weights that need to be adjusted This represents the sum of the second weights.

[0129] Finally, the weights of each second network element are changed from the original weights. Adjust to target weight The process of dynamically adjusting the network element weights in the multi-dimensional dynamic weight adjustment module is as follows: Figure 3As shown. This embodiment of the invention introduces a multi-dimensional performance indicator fusion mechanism and a scenario-based weight configuration system to predefine differentiated indicator weights for different business scenarios. The indicator weights are strongly bound to business requirements, normalizing indicators of different dimensions and solving the problem of inconsistent indicator dimensions. At the same time, it breaks through the limitations of traditional single bandwidth indicators. By generating a comprehensive network element score through weighted summation and dynamically redistributing the weights, it ensures that high-weight network elements in the SDN network receive more traffic, realizing the rational guidance of network traffic and improving the overall transmission efficiency and reliability of the SDN network.

[0130] It is understandable that the higher the weight of a network element in an SDN network, the greater the allocated network resources / distributed network traffic.

[0131] As an optional embodiment, the reallocation of network resources to each of the second network elements according to the adjusted weights includes:

[0132] The original weight and target weight of each second network element are compared to determine the weight difference; wherein, the original weight is the weight of the corresponding second network element before adjustment, and the target weight is the weight of the corresponding second network element after adjustment.

[0133] For example, after calculating the target weights for each second network element, it is necessary to determine the difference between the original weights and the target weights of the second network elements to avoid the potential impact of overly frequent and unnecessary weight updates on the stability of the SDN network. The formula for calculating the weight difference of each second network element is as follows: (6);

[0134] Among them, the absolute value symbol This is used to ensure that the resulting weight difference is positive. This represents the weight difference of the second network element i.

[0135] If the weight difference of any second network element is not less than a preset difference threshold, the weight of each second network element is adjusted from its original weight to its target weight.

[0136] For example, based on the target weights of each second network element calculated above, it is determined whether an SDN network element resource allocation strategy needs to be executed. When If all differences are less than a preset threshold, no network resource reallocation occurs. Otherwise, based on the updated target weights of each second network element, the SDN controller reallocates network resources according to the distribution of target weights of the second network elements in the SDN network. For example, when making traffic forwarding decisions, the data traffic is guided to select a suitable network element (such as the second network element with the largest weight ratio) for forwarding according to the ratio of the current weight of each second network element to the sum of the current weights of all second network elements (i.e., the weight ratio). This achieves dynamic optimization of SDN network resources and ensures that network performance is at its best.

[0137] This invention introduces a multi-dimensional performance indicator fusion mechanism and a scenario-based weight configuration system to predefine differentiated indicator weights for different business scenarios. The indicator weights are strongly bound to business requirements, normalizing indicators of different dimensions and solving the problem of inconsistent indicator dimensions. At the same time, it breaks through the limitations of traditional single bandwidth indicators. By generating a comprehensive network element score through weighted summation, and through dynamic redistribution of weights, it ensures that high-weight network elements in the SDN network receive more traffic. In addition, the introduction of multi-dimensional judgment of weight differences can control the adjustment frequency of network element weights and reduce network oscillations, thereby realizing a closed-loop mapping of "business requirements - indicator weights - resource allocation".

[0138] As an optional embodiment, in the target scenario where the SDN network does not integrate bare metal servers, the network performance indicators include: the SDN network's bps utilization, pps utilization, cps utilization, average latency, jitter rate, and packet loss rate.

[0139] As another optional embodiment, in the target scenario of SDN network converged with bare metal server, the network performance indicators include: the SDN network's bps utilization, pps utilization, cps utilization, average latency, jitter rate, and packet loss rate, and the bare metal server's CPU utilization, memory bandwidth utilization, and network interface card queue utilization.

[0140] In this embodiment of the invention, for bare metal servers with SDN network convergence, hardware indicators such as CPU utilization, memory bandwidth utilization, and network card queue utilization of the bare metal server can also be collected through the server management interface and SDN controller in conjunction with the existing network indicators to form a two-level hardware and software indicator system.

[0141] Based on the principle of SDN scenario-based weight configuration, a new weight for server hardware metrics is added to address different business scenarios. This constructs a scenario-based weight configuration with nine metrics across two levels: software and hardware. An example of a scenario-based weight configuration that integrates software and hardware metrics is shown in the table below.

[0142]

[0143] It should be noted that the dynamic adjustment method for network element weights after hardware metrics of the converged server is the same as that after hardware metrics of the non-converged server, the only difference being that the number of network performance metrics and the number of metric weights are expanded from 6 to 9. The comprehensive score and weight of the interaction between the network element and the bare metal server can be calculated using the above formula (1-5), i.e., the comprehensive score and weight of network elements in the SDN network environment.

[0144] It is understood that bare metal servers are associated with network elements of the SDN network. For example, a bare metal server can be associated with a network element. This embodiment of the invention does not impose specific limitations.

[0145] Furthermore, in the target scenario of SDN network converged bare metal server, the method further includes:

[0146] Monitor the CPU utilization and network interface card queue usage of each bare metal server;

[0147] If any of the bare metal servers meets the preset first traffic diversion condition, the network traffic of the second network element associated with the corresponding bare metal server is diverted to the second network element associated with other bare metal servers that also meet the preset second traffic diversion condition.

[0148] The first traffic splitting condition includes: the CPU utilization rate of the bare metal server exceeds a preset first utilization rate threshold or the network card queue utilization rate exceeds a preset first utilization rate threshold; the second traffic splitting condition includes: the CPU utilization rate of the bare metal server is less than a preset second utilization rate threshold or the network card queue utilization rate is less than a preset second utilization rate threshold.

[0149] For the scenario of SDN network converged with bare metal servers, a bare metal carrying capacity threshold triggering mechanism has been further introduced. Scenario-specific thresholds are set for server hardware indicators. When any hardware indicator of a bare metal server exceeds its corresponding threshold limit, the SDN network traffic is adjusted accordingly.

[0150] In this invention, the lower and upper thresholds for various hardware metrics can be the same or different under different scenario types. This embodiment of the invention does not limit the values ​​of the lower and upper thresholds for various hardware metrics under different scenario types.

[0151] For example, in an eMBB scenario, if a bare metal server's CPU utilization exceeds 80% (i.e., the first utilization threshold, which is the upper limit of the threshold) or its network interface card (NIC) queue utilization exceeds 90% (i.e., the first utilization threshold, which is the upper limit of the threshold), then the traffic allocation weight of the network element (e.g., the second network element) associated with that bare metal server will be reduced by 20%, and network traffic will be redirected to other bare metal servers' associated network elements (i.e., the lower limit of the threshold) with CPU utilization below 50% (i.e., the second utilization threshold, which is the lower limit of the threshold) and NIC queue utilization below 60% (i.e., the second utilization threshold, which is the upper limit of the threshold), to avoid overloading a single server.

[0152] In the target scenario of SDN network convergence with bare metal servers, the method further includes:

[0153] Real-time monitoring of the rate of change of multiple network performance indicators for each of the second network elements;

[0154] Suspicious traffic is identified based on the rate of change of each of the aforementioned network performance indicators; wherein, the suspicious traffic is network traffic corresponding to at least one of the aforementioned network performance indicators having a rate of change exceeding a preset rate of change threshold within a first set time period;

[0155] The suspicious traffic is subjected to DoS attack traffic detection; wherein, the DoS attack traffic is network traffic that includes preset attack characteristics.

[0156] The process of detecting DoS attack traffic from the suspicious traffic includes:

[0157] Feature extraction is performed on the suspicious traffic to obtain the traffic characteristics of the suspicious traffic;

[0158] The traffic characteristics of the suspicious traffic are matched with a preset attack feature database; wherein, the attack feature database includes attack features of multiple DoS attack traffic templates.

[0159] If the traffic characteristics of the suspicious traffic match the attack characteristics of any of the DoS attack traffic templates, the suspicious traffic is determined to be DoS attack traffic.

[0160] In this embodiment of the invention, a multi-dimensional indicator anomaly pattern is used to identify DoS attack traffic. The specific identification process is as follows:

[0161] Anomaly fluctuation detection: Real-time monitoring of the change rate of nine network performance indicators. If the change rate of a certain indicator exceeds the preset change rate threshold (e.g., 50%) within a first set time period (e.g., 10 seconds), for example, the bps utilization of a certain second network element suddenly increases from 20% to 80%, or the number of cps connections suddenly increases from 1,000 to 100,000, the network traffic carried by the second network element is marked as suspicious traffic.

[0162] Attack Feature Matching: Suspicious traffic is characterized by feature extraction, matching the attack features of DoS attack traffic templates. Attack features include: percentage of spoofed source IPs, percentage of packets with empty payloads, number of connection requests sent by the same source IP within a unit of time, and percentage of duplicate packets. For example, spoofed source IPs exceeding 90%, packets with empty payloads exceeding 80% (eMBB scenario spoofing attack), the same source IP sending more than 1000 connection requests within 1 second (mMTC scenario spoofing attack), and duplicate packets exceeding 70% in low-latency service traffic (uRLLC scenario spoofing attack) are examples of spoofed eMBB scenarios. This embodiment of the invention does not impose specific limitations on these scenarios.

[0163] This invention collects various DoS attack traffic in advance as DoS attack traffic templates and extracts attack features to build an attack feature library. By matching the traffic features of suspicious traffic with the attack features in the attack feature library (such as similarity matching), if the traffic features of suspicious traffic match any attack feature, the suspicious traffic can be determined to be DoS attack traffic.

[0164] This invention, through a dual-condition approach of detecting abnormal fluctuations in joint indicators and matching attack features, identifies DoS attack traffic disguised as services, effectively improving the accuracy of DoS attack traffic identification.

[0165] Furthermore, the method also includes:

[0166] Upon detecting DoS attack traffic, the weights of all second network elements and their associated bare metal servers in the target scenario are reset to zero, and the weights of multiple network performance indicators in the target scenario are increased.

[0167] If no DoS attack traffic is detected during the second time period, the weights of all the second network elements and their associated bare metal servers in the target scenario, as well as the weights of multiple network performance indicators, are reset.

[0168] For example, the above-mentioned DoS attack traffic identification method can detect whether the network traffic carried by each second network element is DoS attack traffic; when DoS attack traffic is detected, the security privilege reset zero mechanism is triggered, and the specific process is as follows:

[0169] If DoS attack traffic is detected, the linkage adaptation weight of all second network elements and their associated bare metal servers in the target scenario is set to 0. In this embodiment of the invention, the weight of the second network element is regarded as the linkage adaptation weight of the second network element and its associated bare metal server being set to 0. At the same time, the linkage adaptation weight of the target scenario is increased. In this embodiment of the invention, the weight of multiple network performance indicators in the target scenario is increased, for example, by 30%.

[0170] If no DoS attack traffic is detected during the second time period (e.g., 30 consecutive seconds) after the attack ends, the linkage adaptation weight of the second network element and its associated bare metal server, as well as the weights of multiple network performance indicators, are reset. For example, the second network element is reset to the target weight, and the weights of multiple network performance indicators are reset to the values ​​in the corresponding scenario-based weight configuration.

[0171] For example, if the DoS attack traffic is a disguised attack for a certain business scenario (such as a disguised attack for a uRLLC scenario, which is matched with a disguised attack for a uRLLC scenario), then the weights of the network element (such as the second network element) and the bare metal server corresponding to that scenario are all set to 0, thus blocking the DoS attack traffic from entering the core link of the SDN network and the backend server.

[0172] Meanwhile, the weighting of linkage adaptation in normal business scenarios will be increased by 30%. For example, the weighting of metrics in normal uRLLC scenarios will be adjusted from [bps utilization: 0.03, pps utilization: 0.02, cps utilization: 0.02, average latency: 0.4, jitter rate: 0.2, packet loss rate: 0.25, CPU utilization: 0.05, memory bandwidth utilization: 0.02, network card queue utilization: 0.01] to [bps utilization: 0.039, pps utilization: 0.026, cps utilization: 0.026, average latency: 0.52, jitter rate: 0.26, packet loss rate: 0.325, CPU utilization: 0.065, memory bandwidth utilization: 0.026, network card queue utilization: 0.013] to ensure resource supply for normal business.

[0173] And after the attack ends (no uRLLC scene spoofing attack is detected for 30 consecutive seconds), the original scene-based weight configuration is automatically restored to avoid long-term impact on normal business operations.

[0174] Bare metal servers are the "central brain" of CNC machine tools. In smart factories, the efficient and precise operation of CNC machine tools relies on the support of bare metal servers, and the two form a real-time data interaction closed loop through an SDN network. Taking "smart factory CNC machine tool collaborative processing" in the uRLLC scenario as an example, this scenario requires the network to meet extreme indicators such as latency <10ms, jitter <1ms, and packet loss rate <0.001%. At the same time, it is necessary to ensure the stable hardware support of the bare metal server (configuration: 8-core CPU, 16GB memory, 25Gbps network card) to ensure that multiple CNC machine tools can process precision parts (such as aircraft engine blades) synchronously in real time.

[0175] The existing static weight allocation method has the following problems: (1) Insufficient reliability: It ignores key indicators such as jitter rate and packet loss rate, and does not associate with server CPU utilization. For example, if a machine tool meets the latency standard (8ms) but jitter suddenly increases to 3ms and server CPU utilization rises to 85%, it will cause the tool feed rate to deviate and cause parts to be scrapped; (2) Waste of resources: In order to ensure "extreme reliability", 50% of redundant resources are reserved for a long time. Even if the machine tool is in a low load state, it will not be released, resulting in insufficient resources for other equipment (such as AGV robots); (3) Lagging dynamic response: When 3 CNC machine tools are suddenly added in the workshop (load surge), the static weight cannot quickly adjust the resource allocation, which can easily cause link congestion and the latency will instantly exceed 20ms; (4) Lack of security protection: It cannot identify DoS attacks. Attack traffic disguises normal business, causing server overload and business interruption.

[0176] After the above-mentioned weight adjustment mechanism is used to adjust the weight of the network element (such as the second network element) associated with the bare metal server in the embodiment of the present invention, the following can be achieved: (1) Multi-dimensional dynamic weight adjustment: The predefined indicator weights of uRLLC scenario (such as average latency 0.4, jitter rate 0.2, packet loss rate 0.25) are used to strongly bind the core indicators with the industrial control requirements. The 6 indicators of each machine tool are collected by the SDN controller and a comprehensive score is generated after normalization. The weight is dynamically adjusted by the score. (2) Linkage and adaptation with bare metal, integrating CPU utilization rate, memory bandwidth utilization rate, and network card queue utilization rate indicators. For example, when the CPU utilization rate of a certain server-associated network element rises to 85% due to the connection of 3 high-load machine tools, the SDN controller reduces the weight of the network element from 30% to 10%, and the traffic is guided to other server-associated network elements with a CPU utilization rate of 40%. Finally, the CPU utilization rate of the 3 servers is stabilized at 50%-60%, and the data processing latency is reduced to less than 8ms. (3) DoS attack: A DoS attack disguised itself as a uRLLC scenario, sending empty data packets at 1ms intervals, causing the server's network card queue utilization to surge to 95%. The security weight control module detected "the number of CPS connections suddenly increased from 100 to 50,000 within 10 seconds (abnormal fluctuation) + the proportion of empty data packets was 90% (attack characteristics)," and determined it to be an attack. The weight of the disguised scenario was reset to 0, while the weight of the normal scenario was increased by 30%. During the attack, the normal CNC machine tool data transmission was not interrupted, and the latency was stable at 9ms. The weight configuration was automatically restored 30 seconds after the attack ended. Ultimately, "zero machining error, no idle resources, second-level self-healing of faults, and effective interception of attacks" were achieved, meeting the core requirements of industrial control for "ultra-high reliability and ultra-low latency."

[0177] Compared with the prior art, the beneficial effects of the embodiments of the present invention are as follows:

[0178] (1) By introducing a multi-dimensional network performance index fusion mechanism, multiple network performance indicators are collected, and indicators of different dimensions are normalized to solve the problem of inconsistent indicator dimensions. At the same time, it can cover multiple characteristics such as traffic load, latency sensitivity, and transmission stability, break through the limitations of traditional single bandwidth indicators, avoid misjudgment caused by fluctuations in a single indicator (such as bandwidth), and improve the accuracy of network element status perception.

[0179] (2) Compared with the existing technology that uses fixed weights or general weight models, the embodiments of the present invention pre-allocate the priority of indicators (i.e., indicator weights) according to business characteristics, and achieve accurate matching of business needs through scenario-based weight configuration. Differentiated weight vectors are predefined for different business scenarios. The weight vectors are strongly bound to business needs, realizing direct mapping between business needs and indicator weights, avoiding the inadequacy of general models for specific scenarios. Based on indicator weights, a comprehensive score of network elements is generated by weighted summation. Through dynamic redistribution of weights, it is ensured that network elements with high weight indicators obtain more traffic. Multi-dimensional judgment of weight differences is introduced to control the adjustment frequency, reduce network oscillations, and realize a closed-loop mapping of "business needs - indicator weights - resource allocation".

[0180] (3) By introducing a business-driven scaling mechanism, key network performance indicators are selected as trigger conditions based on scenario-based weight configuration. This avoids frequent adjustments due to fluctuations in non-key indicators. Multiple key network performance indicators can be used to accurately match business needs and reduce resource waste.

[0181] (4) The embodiments of the present invention fully consider the dynamic changes of key network performance indicators and indicator priorities (i.e. indicator weights) under different business scenarios, dynamically adjust resource allocation, quickly alleviate network congestion, and avoid resource waste.

[0182] See Figure 4 , Figure 4 This is a structural block diagram of a network resource allocation device provided in an embodiment of the present invention. The network resource allocation device includes:

[0183] The key indicator determination module 11 is used to select multiple key network performance indicators under the target scenario based on the indicator weights of multiple network performance indicators under the scenario type of the target scenario according to the preset scenario-based weight configuration.

[0184] The network performance monitoring module 12 is used to monitor multiple key network performance indicators of each first network element in the target scenario;

[0185] The network element expansion / shrinkage module 13 is used to obtain multiple second network elements by expanding or shrinking multiple first network elements according to multiple key network performance indicators.

[0186] The dynamic weight adjustment module 14 is used to adjust the weight of each second network element according to multiple network performance indicators and the scenario-based weight configuration of each second network element, and to redistribute network resources of each second network element according to the adjusted weight; wherein, the weight of the second network element is used to indicate the resource redistribution weight of the second network element.

[0187] In an optional embodiment, the network element expansion / shrinkage module 13 includes:

[0188] The judgment unit is used to detect whether multiple key network performance indicators meet preset network element expansion conditions or network element reduction conditions.

[0189] The expansion unit is used to perform network element addition processing according to the expansion / shrinkage step size corresponding to the target scenario when the network element expansion conditions are met, to obtain multiple second network elements; wherein, the second network element includes the first network element and the newly added network element;

[0190] The shrinking unit is used to perform network element deletion processing according to the expansion / shrinking step size corresponding to the target scenario when the network element shrinking conditions are met, to obtain multiple second network elements; wherein, the second network elements include the remaining first network elements after the network elements are deleted.

[0191] In one optional embodiment, the network element expansion condition includes: multiple key network performance indicators all exceed their respective upper threshold limits; the network element reduction condition includes: multiple key network performance indicators all fall below their respective lower threshold limits.

[0192] In an optional embodiment, the dynamic weight adjustment module 14 includes:

[0193] The first calculation unit is used to calculate the sum of the first weights based on the original weights of each of the second network elements;

[0194] The second calculation unit is used to calculate the multi-dimensional comprehensive weight of each second network element based on the multiple network performance indicators and the scenario-based weight configuration of each second network element.

[0195] The third calculation unit is used to calculate the target weight of the corresponding second network element based on the first weight sum and the multi-dimensional comprehensive weight of each second network element, and to adjust the original weight of each second network element to the target weight.

[0196] In one optional embodiment, the second computing unit includes:

[0197] The normalization subunit is used to detect multiple network performance indicators of each second network element, and normalize the detected values ​​of multiple network performance indicators of the second network element to obtain normalized index values ​​of multiple network performance indicators.

[0198] The weight determination subunit is used to determine the weights of multiple network performance indicators under the scenario type to which the target scenario belongs from the scenario-based weight configuration.

[0199] The comprehensive weight calculation subunit is used to calculate the multi-dimensional comprehensive weight of each second network element based on the normalized index values ​​of multiple network performance indicators and the index weights of the corresponding network performance indicators.

[0200] In one optional embodiment, the comprehensive weight calculation subunit includes:

[0201] The first weighted summation subunit is used to perform weighted summation of the normalized index values ​​of the multiple network performance indicators according to the index weights of the multiple network performance indicators, so as to obtain the multi-dimensional comprehensive score of the second network element.

[0202] The first summation subunit is used to sum the multi-dimensional comprehensive scores of all the second network elements to obtain the total score.

[0203] The comprehensive weight determination sub-unit is used to obtain the multi-dimensional comprehensive weight of the corresponding second network element based on the proportion of the multi-dimensional comprehensive score of each second network element in the total score.

[0204] In one optional embodiment, the third computing unit includes:

[0205] The second summation subunit is used to sum the multi-dimensional comprehensive weights of all the second network elements to obtain the second weight sum.

[0206] The target weight calculation subunit is used to calculate the target weight of the second network element based on the proportion of the multi-dimensional comprehensive weight of each second network element in the total of the second weights and the total of the first weights.

[0207] In an optional embodiment, the dynamic weight adjustment module 14 further includes:

[0208] The weight difference judgment unit is used to judge the weight difference between the original weight and the target weight of each second network element; wherein, the original weight is the weight of the corresponding second network element before adjustment, and the target weight is the weight of the corresponding second network element after adjustment;

[0209] The weight adjustment unit is used to adjust the weight of each second network element from its original weight to its target weight, provided that the weight difference of any second network element is not less than a preset difference threshold.

[0210] In an optional embodiment, in a target scenario where the SDN network does not integrate bare metal servers, the network performance metrics include: the SDN network's bps utilization, pps utilization, cps utilization, average latency, jitter rate, and packet loss rate.

[0211] In the target scenario of SDN network converged with bare metal server, the network performance indicators include: the SDN network's bps utilization, pps utilization, cps utilization, average latency, jitter rate, and packet loss rate, and the bare metal server's CPU utilization, memory bandwidth utilization, and network interface card queue utilization.

[0212] In an optional embodiment, the device further includes:

[0213] The metrics monitoring module is used to monitor the CPU utilization and network interface card queue utilization of each bare metal server in the target scenario of SDN network converged bare metal servers;

[0214] The traffic splitting module is used to split the network traffic of the second network element associated with the corresponding bare metal server to other second network elements associated with bare metal servers that meet the preset second splitting conditions, when any of the bare metal servers meets the preset first splitting conditions.

[0215] The first traffic splitting condition includes: the CPU utilization rate of the bare metal server exceeds a preset first utilization rate threshold or the network card queue utilization rate exceeds a preset first utilization rate threshold; the second traffic splitting condition includes: the CPU utilization rate of the bare metal server is less than a preset second utilization rate threshold or the network card queue utilization rate is less than a preset second utilization rate threshold.

[0216] In an optional embodiment, the device further includes:

[0217] The rate of change monitoring module is used to monitor the rate of change of multiple network performance indicators of each of the second network elements in real time in the target scenario of SDN network converged bare metal server;

[0218] The suspicious traffic identification module is used to identify suspicious traffic based on the rate of change of each of the network performance indicators; wherein, the suspicious traffic is network traffic corresponding to at least one of the network performance indicators having a rate of change exceeding a preset rate of change threshold within a first set time period;

[0219] An attack traffic identification module is used to detect DoS attack traffic on the suspicious traffic; wherein, the DoS attack traffic is network traffic that includes preset attack characteristics.

[0220] In one optional embodiment, the attack traffic identification module includes:

[0221] The feature extraction unit is used to extract features from the suspicious traffic to obtain the traffic features of the suspicious traffic;

[0222] The feature matching unit is used to match the traffic features of the suspicious traffic with a preset attack feature library; wherein, the attack feature library includes attack features of multiple DoS attack traffic templates.

[0223] The traffic determination unit is used to determine that the suspicious traffic is DoS attack traffic if the traffic characteristics of the suspicious traffic match the attack characteristics of any of the DoS attack traffic templates.

[0224] In an optional embodiment, the device further includes:

[0225] The weight adjustment module is used to reset the weights of all the second network elements and their associated bare metal servers in the target scenario to zero when a DoS attack is detected, and to increase the weights of multiple network performance indicators in the target scenario.

[0226] The weight reset module is used to reset the weights of all the second network elements and their associated bare metal servers, as well as the weights of multiple network performance indicators, in the target scenario when no DoS attack traffic is detected during the second time period.

[0227] It should be noted that the working process of each module in the network resource allocation device described in the embodiments of the present invention can refer to the working process of the network resource allocation method described in the above embodiments, and the technical effect achieved is the same as that of the network resource allocation method described in the above embodiments, so it will not be repeated here.

[0228] See Figure 5 , Figure 5 This is a structural block diagram of a network resource allocation device provided in an embodiment of the present invention. The network resource allocation device includes a processor 21, a memory 22, and a computer program stored in the memory 22 and executable on the processor 21. When the processor 21 executes the computer program, it implements the steps in the above-described network resource allocation method embodiments, such as steps S11 to S14.

[0229] For example, the computer program may be divided into one or more modules / units, which are stored in the memory 22 and executed by the processor 21 to complete the present invention. The one or more modules / units may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the network resource allocation device.

[0230] The network resource allocation device may include, but is not limited to, a processor 21 and a memory 22. Those skilled in the art will understand that the schematic diagram is merely an example of a network resource allocation device and does not constitute a limitation on the device. It may include more or fewer components than illustrated, or combine certain components, or use different components. For example, the network resource allocation device may also include input / output devices, network access devices, buses, etc.

[0231] The processor 21 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor 21 is the control center of the network resource allocation device, connecting various parts of the entire network resource allocation device through various interfaces and lines.

[0232] The memory 22 can be used to store the computer programs and / or modules. The processor 21 implements various functions of the network resource allocation device by running or executing the computer programs and / or modules stored in the memory 22 and calling the data stored in the memory 22. The memory 22 may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phonebook, etc.). In addition, the memory 22 may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0233] If the modules / units integrated into the network resource allocation device are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by the processor 21, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc.

[0234] It should be noted that the device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network elements. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can be specifically implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0235] The above description represents the preferred embodiments of the present invention. It should be noted that, for those skilled in the art, various improvements and modifications can be made without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for allocating network resources, characterized in that, include: Based on the preset scenario-based weight configuration, the indicator weights of multiple network performance indicators under the scenario type to which the target scenario belongs are indicated, and multiple key network performance indicators under the target scenario are selected. Monitor multiple key network performance indicators of each first network element in the target scenario; Based on multiple key network performance indicators, multiple second network elements are obtained by expanding or shrinking multiple first network elements; Based on the multiple network performance indicators and the scenario-based weight configuration of each second network element, the weights of each second network element are adjusted, and network resources are reallocated to each second network element according to the adjusted weights; wherein, the weights of the second network elements are used to indicate the resource reallocation weights of the second network elements. The step of adjusting the weights of each second network element based on multiple network performance indicators and the scenario-based weight configuration includes: Calculate the sum of the first weights based on the original weights of each of the second network elements; Based on the multiple network performance indicators and the scenario-based weight configuration of each second network element, calculate the multi-dimensional comprehensive weight of each second network element; Based on the sum of the first weights and the multi-dimensional comprehensive weights of each second network element, the target weight of the corresponding second network element is calculated, and the original weights of each second network element are adjusted to the target weights.

2. The network resource allocation method as described in claim 1, characterized in that, The process of obtaining multiple second network elements by expanding or shrinking multiple first network elements based on multiple key network performance indicators includes: Detect whether multiple key network performance indicators meet preset network element expansion or network element reduction conditions; Under the condition of network element expansion, network element addition processing is performed according to the expansion / shrinkage step size corresponding to the target scenario to obtain multiple second network elements; wherein, the second network element includes the first network element and the newly added network element; If the network element scaling conditions are met, network element deletion is performed according to the scaling step size corresponding to the target scenario to obtain multiple second network elements; wherein, the second network elements include the remaining first network elements after the deletion of network elements.

3. The network resource allocation method as described in claim 2, characterized in that, The network element expansion conditions include: multiple key network performance indicators all exceed their respective upper threshold limits; the network element reduction conditions include: multiple key network performance indicators all fall below their respective lower threshold limits.

4. The network resource allocation method as described in claim 1, characterized in that, The step of calculating the multi-dimensional comprehensive weight of each second network element based on multiple network performance indicators and the scenario-based weight configuration includes: For each of the second network elements, multiple network performance indicators of the second network element are detected, and the values ​​of the detected multiple network performance indicators of the second network element are normalized to obtain normalized index values ​​of the multiple network performance indicators. The weights of multiple network performance metrics under the scenario type to which the target scenario belongs are determined from the scenario-based weight configuration. Based on the normalized index values ​​of multiple network performance indicators of each second network element and the index weights of the corresponding network performance indicators, the multi-dimensional comprehensive weight of each second network element is calculated.

5. The network resource allocation method as described in claim 4, characterized in that, The step of calculating the multi-dimensional comprehensive weight of each second network element based on the normalized index values ​​of multiple network performance indicators and the corresponding index weights of each network performance indicator includes: Based on the weights of the multiple network performance indicators, the normalized values ​​of the multiple network performance indicators are weighted and summed to obtain the multi-dimensional comprehensive score of the second network element. The multi-dimensional comprehensive scores of all second network elements are summed to obtain the total score. The multi-dimensional comprehensive weight of each second network element is obtained based on its proportion in the total score.

6. The network resource allocation method as described in claim 1, characterized in that, The step of calculating the target weight of the corresponding second network element based on the sum of the first weights and the multi-dimensional comprehensive weights of each second network element includes: The multi-dimensional comprehensive weights of all the second network elements are summed to obtain the total second weights. The target weight of the second network element is calculated based on the proportion of the multi-dimensional comprehensive weight of each second network element in the total weight of the second weight and the total weight of the first weight.

7. The network resource allocation method as described in claim 1, characterized in that, The process of reallocating network resources to each of the second network elements according to the adjusted weights includes: The original weight and target weight of each second network element are compared to determine the weight difference; wherein, the original weight is the weight of the corresponding second network element before adjustment, and the target weight is the weight of the corresponding second network element after adjustment. If the weight difference of any second network element is not less than a preset difference threshold, the weight of each second network element is adjusted from its original weight to its target weight.

8. The network resource allocation method as described in claim 1, characterized in that, In the target scenario where the SDN network does not integrate bare metal servers, the network performance indicators include: the SDN network's bps utilization, pps utilization, cps utilization, average latency, jitter rate, and packet loss rate. In the target scenario of SDN network converged with bare metal server, the network performance indicators include: the SDN network's bps utilization, pps utilization, cps utilization, average latency, jitter rate, and packet loss rate, and the bare metal server's CPU utilization, memory bandwidth utilization, and network interface card queue utilization.

9. The network resource allocation method as described in claim 8, characterized in that, In the target scenario of SDN network convergence with bare metal servers, the method further includes: Monitor the CPU utilization and network interface card queue usage of each bare metal server; If any of the bare metal servers meets the preset first traffic diversion condition, the network traffic of the second network element associated with the corresponding bare metal server is diverted to the second network element associated with other bare metal servers that also meet the preset second traffic diversion condition. The first traffic splitting condition includes: the CPU utilization rate of the bare metal server exceeds a preset first utilization rate threshold or the network card queue utilization rate exceeds a preset first utilization rate threshold; the second traffic splitting condition includes: the CPU utilization rate of the bare metal server is less than a preset second utilization rate threshold or the network card queue utilization rate is less than a preset second utilization rate threshold.

10. The network resource allocation method as described in claim 8, characterized in that, In the target scenario of SDN network convergence with bare metal servers, the method further includes: Real-time monitoring of the rate of change of multiple network performance indicators for each of the second network elements; Suspicious traffic is identified based on the rate of change of each of the aforementioned network performance indicators; wherein, the suspicious traffic is network traffic corresponding to at least one of the aforementioned network performance indicators having a rate of change exceeding a preset rate of change threshold within a first set time period; The suspicious traffic is subjected to DoS attack traffic detection; wherein, the DoS attack traffic is network traffic that includes preset attack characteristics.

11. The network resource allocation method as described in claim 10, characterized in that, The process of detecting DoS attack traffic in the suspicious traffic includes: Feature extraction is performed on the suspicious traffic to obtain the traffic characteristics of the suspicious traffic; The traffic characteristics of the suspicious traffic are matched with a preset attack feature database; wherein, the attack feature database includes attack features of multiple DoS attack traffic templates. If the traffic characteristics of the suspicious traffic match the attack characteristics of any of the DoS attack traffic templates, the suspicious traffic is determined to be DoS attack traffic.

12. The network resource allocation method as described in claim 11, characterized in that, The method further includes: Upon detecting DoS attack traffic, the weights of all second network elements and their associated bare metal servers in the target scenario are reset to zero, and the weights of multiple network performance indicators in the target scenario are increased. If no DoS attack traffic is detected during the second time period, the weights of all the second network elements and their associated bare metal servers in the target scenario, as well as the weights of multiple network performance indicators, are reset.

13. A network resource allocation device, characterized in that, include: The key indicator determination module is used to select multiple key network performance indicators under the target scenario based on the indicator weights of multiple network performance indicators under the scenario type of the target scenario according to the preset scenario-based weight configuration. A network performance monitoring module is used to monitor multiple key network performance indicators of each first network element in the target scenario; The network element expansion / shrinkage module is used to obtain multiple second network elements by expanding or shrinking multiple first network elements according to multiple key network performance indicators. The dynamic weight adjustment module is used to adjust the weight of each second network element according to multiple network performance indicators and the scenario-based weight configuration, and to reallocate network resources of each second network element according to the adjusted weight; wherein, the weight of the second network element is used to indicate the resource reallocation weight of the second network element. The dynamic weight adjustment module includes: The first calculation unit is used to calculate the sum of the first weights based on the original weights of each of the second network elements; The second calculation unit is used to calculate the multi-dimensional comprehensive weight of each second network element based on the multiple network performance indicators and the scenario-based weight configuration of each second network element. The third calculation unit is used to calculate the target weight of the corresponding second network element based on the sum of the first weights and the multi-dimensional comprehensive weight of each second network element, and to adjust the original weight of each second network element to the target weight.

14. A network resource allocation device, characterized in that, include: A processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein the processor, when executing the computer program, implements the network resource allocation method as described in any one of claims 1 to 12.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to perform the network resource allocation method as described in any one of claims 1 to 12.

16. A computer program product comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by a processor, they implement the network resource allocation method according to any one of claims 1 to 12.

Images