Optimization method for enhancing network reliability based on fractional power diagram characteristics

Abstract

The invention discloses an optimization method for enhancing network reliability based on fractional power diagram characteristics, which comprises the following steps of: analyzing an original network topology, and calculating fractional power values of nodes and links so as to determine a connection relationship and a weight in a fractional power diagram, thereby constructing a corresponding fractional power diagram; defining a network reliability evaluation index suitable for the fractional power diagram, and accurately evaluating the reliability level of the network in different states through the index; according to the fractional power diagram characteristics and the reliability evaluation result, an optimization strategy is formulated, and backup and enhancement measures are taken for nodes or links which are low in fractional power value and have large influence on the network reliability; network state changes are monitored in real time, when the network breaks down or loads change, the fractional power diagram and the reliability index are recalculated, the optimization strategy is adjusted in time, and it is ensured that the network keeps high reliability all the time. According to the invention, network optimization is carried out based on fractional power diagram characteristics, and the network reliability is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to a network optimization method, in particular to an optimization method for enhancing network reliability. BACKGROUND

[0002] With the rapid development of the Internet, the network scale is expanding, and it is widely used in communication, finance and other fields. Network reliability directly affects business continuity, data transmission accuracy and user experience. In actual operation, network reliability faces many challenges. In terms of hardware, device aging, electrical failure, etc. often cause node failure, resulting in data transmission interruption. At the link level, natural disasters such as earthquakes, floods, etc. Misoperation of construction, and electromagnetic interference, may cause link interruption or signal attenuation, affecting the stability and integrity of data transmission. In the field of software and network architecture, the problem of load imbalance is prominent. A large number of user access requests in a short time will make the server and the link unable to bear the heavy burden, and the page loading will be slow or even paralyzed. In addition, the imperfections of network protocols are easily exploited by malicious software attacks, which seriously impact network reliability.

[0003] In the traditional method for enhancing network reliability, although redundancy design can improve reliability, it greatly increases the construction and maintenance cost, and the redundant resources are idle and wasted. The strategy relying on complex algorithms has high computational complexity, and in the face of large-scale networks, it consumes a lot of resources and time, and has poor real-time response ability, which is difficult to cope with the rapid changes of network state. At the same time, the traditional method lacks accurate means for evaluating the importance of nodes and links, and the optimization measures are difficult to accurately attack, so the overall effect is poor. SUMMARY

[0004] The purpose of the present application is to provide an Internet architecture optimization method based on the characteristics of fractional power graph for enhancing network reliability, which optimizes the network architecture by using the unique properties of fractional power graph, and improves the reliability, stability and data transmission efficiency of the network.

[0005] Technical scheme: An optimization method for enhancing network reliability based on the characteristics of fractional power graph, comprising:

[0006] Step 1: Analyze the original network topology, calculate the fractional power value of the nodes and links, and determine the connection relationship and weight in the fractional power graph to construct the corresponding fractional power graph; calculate the fractional power value of the nodes and links, and determine the connection relationship and weight in the fractional power graph;

[0007] Step 2: Evaluate the reliability level of the network in different states by using the network reliability comprehensive evaluation index suitable for the fractional power graph;

[0008] Step 3: Based on the fractional power graph characteristics and reliability evaluation results, develop optimization strategies: take backup measures for nodes with low fractional power values and large impact on network reliability; take link enhancement measures for links with low fractional power values;

[0009] Step 4: Establish a dynamic adjustment mechanism to monitor network state changes in real time. When network failures or load changes occur, recalculate the fractional power graph and reliability indicators, and adjust the optimization strategy in a timely manner based on the calculation results to ensure network reliability.

[0010] Further, in step 1, the calculation method of node fractional power value is: let the network node set be V = {v1, v2, …, v n}, the fractional power value of node v i is ; wherein N(v i ) is the neighbor node set of node v i , d(v j ) is the degree of neighbor node v j , and α is a fractional power index preset in (0, 1].

[0011] The calculation method of link fractional power value is: let the network link set be E = {e ij |v i ,v j ∈V}, the fractional power value of link e ij is

[0012] Construct the fractional power graph G f = (V f ,E f ), where the vertex set V f = V, and the link weight in the edge set E f is P(e ij ).

[0013] Further, the network reliability comprehensive evaluation index is an index that comprehensively considers node connectivity and data transmission delay; wherein the connectivity index of node v i is , wherein |N(v i )| represents the number of neighbor nodes of node v i , and n is the number of nodes in the network.

[0014] The data transmission delay index D is measured by calculating the average delay of data transmission between all nodes in the network. Let the data transmission delay from node v i to node v j be d ij , then the average delay is where |E| is the number of links in the link set E.

[0015] Network Reliability Comprehensive Evaluation Indicators C avg This is the average of the connectivity metrics for all nodes. ;D max ω1 and ω2 are the maximum allowable average data transmission delay in the network; ω1 and ω2 are weighting coefficients, and ω1 + ω2 = 1.

[0016] Furthermore, in step 3, the step of taking backup measures for nodes includes: if the fractional power value of a node is less than a pre-set node fractional power threshold θ, then the node is considered to have a low fractional power value; when a node with a low fractional power value is found to have a significant impact on network reliability, the backup measures are taken accordingly. k At that time, select with v in the network k Nodes v that are close to each other and have low load m As a backup node, establish a connection from v m to v k The backup link, the bandwidth of the backup link depends on node v k The business needs were determined.

[0017] Furthermore, in step 3, the step of taking link enhancement measures includes: if the link's power-law value is less than a preset link power-law threshold... If the fractional power of a link is low, then the link is considered to have a low fractional power value; for links with low fractional power values, e... pq If the original link bandwidth is B0, the bandwidth can be increased to B1 = B0 × (1 + β) by upgrading the hardware or adding parallel lines, where β is the bandwidth increase coefficient.

[0018] Furthermore, a node is deemed to have a significant impact on network reliability when any of the following verifiable conditions are met: 1) Historical faults cause production interruptions of ≥2 hours or office service blockages of ≥30 minutes; 2) When the performance of the simulated node is insufficient, office latency exceeds 400 milliseconds or production packet loss exceeds 10%; 3) The business bearer is the only / main bearer node for critical office / production services, and a fault will directly interrupt the services.

[0019] Furthermore, the selection of v in the network k Nodes v that are close to each other and have low load m As backup nodes, the method for determining node proximity is as follows: based on network topology, it is determined by the shortest path hop count or the physical deployment distance of the nodes, and a threshold is set; if the distance is below the threshold, it is considered close to v. k Close distance;

[0020] The method for identifying nodes with low load is as follows: nodes whose resource usage indicators are below a set threshold are considered nodes with low load; wherein the resource usage indicators include: reference node CPU utilization, memory utilization, and bandwidth utilization.

[0021] Further, for the part with high fractional power value, follow the business load matching principle, calculate resource demand according to business type, and maintain the resource utilization rate of the node / link with high fractional power value at 60%-80%.

[0022] Further, in step 4, the network state monitoring adopts a distributed monitoring mode, and a monitoring agent is deployed on the key nodes in the network. The monitoring agent collects network state information every τ time interval, including the load of the node, the bandwidth utilization of the link, and the data transmission delay.

[0023] Further, in step 4, when it is monitored that the node load exceeds the threshold L max , or the link bandwidth utilization exceeds the threshold U max , or the node connectivity changes, the fractional power graph and the network reliability comprehensive evaluation index are recalculated.

[0024] Beneficial effects: 1. Improve network reliability: Through the architecture optimization and redundant link setting based on the characteristics of fractional power graph, the fault tolerance of the network is greatly enhanced, and the network service interruption caused by node failure or link interruption is reduced.

[0025] 2. Improve data transmission efficiency: The optimized routing algorithm and resource allocation strategy can more reasonably guide data transmission, reduce network congestion, reduce data transmission delay, and improve overall data transmission efficiency.

[0026] 3. Enhance network scalability: The application of fractional power graph characteristics makes the network architecture have better scalability, which facilitates the flexible addition of nodes and links when the network scale is expanded, without causing great impact on the original network performance. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 is a flowchart of the method of the present application;

[0028] Figure 2 is a fractional power value distribution diagram in the embodiment;

[0029] Figure 3 is the network reliability optimization result presentation of the enterprise park in the embodiment. DETAILED DESCRIPTION

[0030] The present application will be further described in detail below in combination with the drawings and embodiments.

[0031] An optimization method for enhancing network reliability based on fractional power graph characteristics, as shown in the figure, includes the following steps: Figure 1

[0032] ​Step 1: Analyze the original network topology, calculate the fractional power values ​​of nodes and links based on their importance factors, and determine the connection relationships and weights in the fractional power graph to construct the corresponding fractional power graph.

[0033] First, obtain the original network topology data, assuming the set of network nodes is V = {v1, v2, ..., v...}. n The link set is E = {e} ij |v i ,v j Let v ∈ V}, where n is the number of nodes in the network. For each node v i Calculate its fractional power value P(v) i For link e ij Calculate its fractional power value P(e ij The following formula is used:

[0034]

[0035] Wherein, N(v) i ) is node v i The set of neighboring nodes, d(v j ) is a neighbor node v j The degree, α, is a pre-defined fractional power exponent, typically ranging from (0,1], and is adjusted according to the actual network characteristics.

[0036] By using fractional power values ​​to determine the connectivity and weights in the graph, the original network topology is transformed into a fractional power graph G, which facilitates the analysis of network reliability. f =(V f E f ), where the vertex set V f =V, edge set E f The link weight in the equation is P(e) ij ).

[0037] Step 2: Accurately assess the network reliability level under different states using a comprehensive network reliability evaluation index applicable to fractional power graphs.

[0038] For node v i Its connectivity index C(v) i The calculation method is as follows: Where, |N(v i | represents node v i The number of neighboring nodes of node v, this metric reflects the number of neighboring nodes of node v. i The degree of connection with other nodes in the network, with a value between 0 and 1, where 0 indicates that the node is completely isolated and 1 indicates that it is well connected to all reachable nodes.

[0039] The data transmission latency metric D is measured by calculating the average latency of data transmission between all nodes in the network. Let the slave node v... i to node v j The data transmission delay is d ij Then the average delay Here, |E| represents the number of links in the link set E.

[0040] Define a comprehensive network reliability evaluation index R, which comprehensively considers node connectivity and data transmission latency. The calculation method is as follows: in, That is, the average of the connectivity indices of all nodes; D max ω1 and ω2 are the maximum allowable average data transmission delay in the network; ω1 and ω2 are weighting coefficients, and ω1 + ω2 = 1. In general data transmission network scenarios, ω1 = 0.6 and ω2 = 0.4 can be set according to actual needs to balance the impact of node connectivity and data transmission delay on network reliability.

[0041] Step 3: Based on the characteristics of the fractional power graph and the reliability assessment results, formulate optimization strategies. For nodes with low fractional power values ​​and significant impact on network reliability, implement backups; for links with low fractional power values, implement link enhancement measures; for links with high fractional power values, allocate resources reasonably to improve the overall network reliability.

[0042] If the fractional power value of a node is less than the preset node fractional power threshold θ, such as θ = 0.3, then the node is considered to have a low fractional power value.

[0043] By combining data on service load, fault history (such as production interruption correlation), and the degree of network performance degradation under simulated faults, the scope and severity of the impact on office, production, and other business operations when a node fails or its performance is insufficient can be determined. Specifically, a node is considered to have a "significant impact" on network reliability when any of the following verifiable conditions are met:

[0044] 1) Historical faults causing production interruptions of ≥2 hours or office service disruptions of ≥30 minutes;

[0045] 2) When the performance of the simulation node is insufficient (computing power reduced by 50%, bandwidth limited to 80%), office latency exceeds 400 milliseconds and production packet loss exceeds 10%, threatening business continuity.

[0046] 3) The business bearer is the sole / primary bearer of critical office / production business (such as production instructions and order storage), and a failure will directly interrupt the business.

[0047] When a node v with a low fractional power value is found that has a significant impact on network reliability... k In such cases, backup measures should be taken. Specifically, this involves selecting a network with v k Nodes v that are close to each other and have low loadm As a backup node, establish a backup link from v m to v k , the bandwidth of the backup link is determined according to the business needs of node v k . Among them, the method for judging the distance between nodes is: according to the network topology, through the shortest path hop count (the number of links passed in the topology graph, the fewer the hops, the closer), or the physical deployment space distance of the node, set a threshold, such as the shortest path hop count < 3, or the physical deployment space distance of the node < 100 meters, below the threshold is "close to v k ". The method for judging the low-load node v m is: referring to the resource occupation indexes such as CPU usage, memory usage, and bandwidth occupation, combined with the normal load of the network, set a threshold, such as CPU usage < 30% for a long time, bandwidth occupation < 20% of the total bandwidth, the node below the threshold is "low-load node". In addition, for the low-score power node of the business key, the performance is improved by upgrading the hardware.

[0048] If the score power value of the link is less than the pre-set link score power threshold , such as , it is considered that the score power value of the link is low. For the link e pq with low score power value, link enhancement technology is adopted. Such as increasing the bandwidth of the link, if the original link bandwidth is B0, the bandwidth is improved to B1 = B0 x (1 + β) through upgrading the hardware or increasing the parallel line, β is the bandwidth improvement coefficient, which is valued according to the actual situation, generally between 0.2 to 0.5.

[0049] If the score power value of the node is greater than the pre-set node score power threshold θ', such as θ' = 0.8, it is considered that the score power value of the node is high. For the part with high score power value, reasonably allocate resources: follow the business load matching principle, calculate the resource demand according to the business type (office, production, etc.), so that the resource utilization rate of the node / link with high score power value is maintained at 60%-80% for a long time, which avoids idling and leaves redundancy.

[0050] Step 4: Establish a dynamic adjustment mechanism to monitor network state changes in real time. When the network fails or the load changes, recalculate the score power diagram and reliability index, and timely adjust and optimize the strategy to ensure that the network always maintains high reliability.

[0051] The network state monitoring adopts a distributed monitoring method, and a monitoring agent is deployed in the key nodes in the network. The monitoring agent collects network state information every τ time interval, including node load, link bandwidth utilization, data transmission delay, etc.

[0052] When it is monitored that the node load exceeds the threshold L max or the link bandwidth utilization exceeds the threshold Umax , or when node connectivity changes, triggering re-computation and adjustment.

[0053] Re-compute the fractional power graph, re-compute the fractional power values of nodes and links according to the new network state data, and update the fractional power graph according to the above fractional power graph construction steps.

[0054] Re-evaluate network reliability, re-compute network reliability indicators R according to the new fractional power graph data and the reliability evaluation indicator calculation process.

[0055] According to the new reliability evaluation results and the characteristics of the fractional power graph, adjust the optimization strategy, such as taking appropriate backup or enhancement measures for nodes or links with low fractional power values.

[0056] An apparatus for implementing the above-mentioned optimization method for enhancing network reliability based on the characteristics of the fractional power graph, comprising:

[0057] Data input module: receives and processes original network topology data and related parameter information. Supports input of multiple data formats such as text files, database records, etc. Preliminary cleaning and verification of input data to ensure data accuracy and integrity.

[0058] Model construction and loading module: constructs the fractional power graph model according to the data provided by the data input module. Calculate the fractional power values of nodes and links according to specific algorithms and rules, determine the connection relationship and weight in the graph.

[0059] Numerical calculation module: based on the fractional power graph model and related calculation logic, perform various numerical calculations. Calculate network reliability evaluation indicators, including node connectivity, data transmission delay and other factors. According to the calculation results, provide data support for the formulation of optimization strategy, such as determining the specific numerical standards of nodes and links that need to be backed up or enhanced.

[0060] Result analysis module: present the calculation results in an intuitive way, such as charts, reports, etc., to help users understand the reliability status of the network and the effect of the optimization strategy. At the same time, according to the analysis results, put forward further improvement suggestions and optimization direction, provide decision basis for the continuous optimization of the network.

[0061] The application will be further described in detail below with reference to the embodiments.

[0062] A certain enterprise park network contains 500 nodes and 1000 links, carrying office, production and other businesses. Some nodes are old, link bandwidth is narrow, packet loss is serious, and production has been interrupted for 2 hours due to server failure, network reliability is poor. Collect network topology data, calculate the fractional power values of nodes and links according to node business carrying capacity, link bandwidth and other parameters, and construct the fractional power graph, such asFigure 2 The fraction power value distribution diagram is shown, and the importance of nodes and links is clear. Figure 3 Present the enterprise park network reliability optimization result.

[0063] The above merely describes the preferred embodiments of the present application, but not for limiting the protection scope of the present application.

Claims

1. An optimization method for enhancing network reliability based on fractional power graph characteristics, characterized in that, include: Step 1: Analyze the original network topology, calculate the fractional power values ​​of nodes and links to determine the connection relationships and weights in the fractional power graph, and thus construct the corresponding fractional power graph; calculate the fractional power values ​​of nodes and links to determine the connection relationships and weights in the fractional power graph. Step 2: Evaluate the network reliability level under different states using a comprehensive network reliability evaluation index applicable to fractional power graphs; Step 3: Based on the characteristics of the fractional power graph and the reliability assessment results, formulate optimization strategies: take backup measures for nodes with low fractional power values ​​and significant impact on network reliability; take link enhancement measures for links with low fractional power values. Step 4: Establish a dynamic adjustment mechanism to monitor network status changes in real time. When network failures or load changes occur, recalculate the fractional power graph and reliability indicators, and adjust the optimization strategy in a timely manner based on the calculation results to ensure network reliability.

2. The method according to claim 1, characterized in that, In step 1, the method for calculating the power value of the node fraction is as follows: Let the set of network nodes be V = {v1, v2, ..., v...} n }, node v i fractional power Wherein, N(v) i ) is node v i The set of neighboring nodes, d(v j ) is a neighbor node v j The degree of α is a fractional power exponent preset in (0,1]. The method for calculating the power value of the link fraction is as follows: Let the set of network links be E = {e ij |v i ,v j ∈V}, link e ij fractional power Construct the fractional power graph G f =(V f E f ), where the vertex set V f =V, edge set E f The link weight in the equation is P(e) ij ).

3. The method according to claim 2, characterized in that, In step 2, the comprehensive network reliability evaluation index is an index that comprehensively considers node connectivity and data transmission latency; wherein, node v i connectivity index Where |N(v i | represents node v i The number of neighboring nodes, where n is the number of nodes in the network; Data transmission latency metric D is measured by calculating the average latency of data transmission between all nodes in the network. Let the latency be the latency between node v and node v. i to node v j The data transmission delay is d ij Then the average delay Where |E| is the number of links in the link set E; Network Reliability Comprehensive Evaluation Indicators Where C avg This is the average of the connectivity metrics for all nodes. D max ω1 and ω2 are the maximum allowable average data transmission delay in the network; ω1 and ω2 are weighting coefficients, and ω1 + ω2 = 1.

4. The method according to claim 1, characterized in that, In step 3, the step of taking backup measures for nodes includes: if the fractional power value of a node is less than a pre-set node fractional power threshold θ, then the node is considered to have a low fractional power value; when a node with a low fractional power value is found to have a significant impact on network reliability, then... k At that time, select with v in the network k Nodes v that are close to each other and have low load m As a backup node, establish a connection from v m to v k The backup link, the bandwidth of the backup link depends on node v k The business needs were determined.

5. The method according to claim 1, characterized in that, In step 3, the step of taking link enhancement measures includes: if the link's power-law value is less than a preset link power-law threshold... Then the link is considered to have a low fractional power value; for links with low fractional power values, e... pq If the original link bandwidth is B0, the bandwidth can be increased to B1 = B0 × (1 + β) by upgrading the hardware or adding parallel lines, where β is the bandwidth increase coefficient.

6. The method according to claim 4, characterized in that, A node is considered to have a significant impact on network reliability if any of the following verifiable conditions are met: 1) Historical faults cause production interruptions of ≥2 hours or office service blockages of ≥30 minutes; 2) When the performance of the simulated node is insufficient, office latency exceeds 400 milliseconds or production packet loss exceeds 10%; 3) The business bearer is the only / main bearer node for critical office / production business, and a fault will directly interrupt the business.

7. The method according to claim 4, characterized in that, The selection of v in the network k Nodes v that are close to each other and have low load m As backup nodes, the method for determining node proximity is as follows: based on network topology, it is determined by the shortest path hop count or the physical deployment distance of the nodes, and a threshold is set; if the distance is below the threshold, the node is considered close to v. k Close distance; The method for identifying nodes with low load is as follows: nodes whose resource usage index is lower than a set threshold are considered nodes with low load. The resource usage metrics mentioned include: reference node CPU utilization, memory utilization, and bandwidth utilization.

8. The method according to claim 1, characterized in that, For the portion with a high power value, the principle of matching business load is followed, and resource requirements are calculated based on the business type to maintain the resource utilization rate of nodes / links with high power values ​​at 60%-80%.

9. The method according to claim 1, characterized in that, In step 4, network status monitoring adopts a distributed monitoring method, deploying monitoring agents at key nodes in the network. The monitoring agents collect network status information every τ time interval, including node load, link bandwidth utilization, and data transmission latency.

10. The method according to claim 9, characterized in that, In step 4, when the node load is detected to exceed the threshold L max Or the link bandwidth utilization exceeds the threshold U max Or, when node connectivity changes, it triggers the recalculation of the fractional power graph and the comprehensive network reliability assessment index.

Images