System and method for optimizing enterprise private network

Abstract

A system and method for optimizing an enterprise private network are provided. The method includes the following steps: a traffic analysis module uses traffic data associated with the enterprise private network to analyze whether abnormal traffic will occur; when the traffic analysis module analyzes that abnormal traffic will occur, an abnormal network segment location module uses configuration data associated with the enterprise private network to locate the abnormal network segment; and a network access control decision module uses the traffic data and the abnormal network segment to determine the network access control decision for the enterprise private network.

Description

[Technical Field]

[0001] This invention relates to a system and method for optimizing enterprise private networks. [Previous Technology]

[0002] The low latency of enterprise private networks, especially the ultra-low latency of 5G enterprise private networks, has enabled several previously unattainable application scenarios, including real-time robot operation in smart factories, remote diagnosis and surgery in telemedicine, and the realization of self-driving cars. However, if service delays occur due to insufficient bandwidth allocation, the above applications will be unable to be controlled in real time, resulting in inaccurate operation and difficulty in completing tasks. [Summary of the Invention]

[0003] The system for optimizing an enterprise private network of the present invention includes a traffic analysis module, an abnormal network segment location module, and a troubleshooting decision module. The traffic analysis module uses traffic data associated with the enterprise private network to analyze whether abnormal traffic will occur; when the traffic analysis module analyzes that abnormal traffic will occur, the abnormal network segment location module uses configuration data associated with the enterprise private network to locate the abnormal network segment; the troubleshooting decision module uses the traffic data and the abnormal network segment to determine the troubleshooting decision for the enterprise private network.

[0004] The method for optimizing an enterprise private network according to the present invention includes the following steps: a traffic analysis module uses traffic data associated with the enterprise private network to analyze whether abnormal traffic will occur; when the traffic analysis module analyzes that abnormal traffic will occur, an abnormal network segment location module uses configuration data associated with the enterprise private network to locate the abnormal network segment; and a traffic management decision module uses traffic data and the abnormal network segment to determine the traffic management decision for the enterprise private network.

Implementation Method

[0006] FIG1 is a schematic diagram illustrating a system 1000 for optimizing an enterprise private network according to an embodiment of the present invention. The system 1000 may include a traffic analysis module 1300, an abnormal network segment location module 1400, and a traffic management decision module 1500. In one embodiment, the system 1000 may include a data collection module 1100. In one embodiment, the system 1000 may include a traffic prediction module 1200. In one embodiment, the system 1000 may include a simulation verification module 1600. In one embodiment, the system 1000 may include an optimization control module 1700. In one embodiment, the system 1000 may include a storage module 1800. In one embodiment, each module shown in FIG1 may be software and / or firmware code executed by a processor. In one embodiment, the enterprise private network device 2000 may be associated with mobile terminal resources 2100, base station resources 2200, MEC (Multi-access Edge Computing or Mobile Edge Computing) resources 2300, core network resources 2400, server resources 2500, and network equipment resources 2600.

[0007] Figure 2 is a flowchart illustrating a method for optimizing an enterprise private network according to an embodiment of the present invention, wherein the method can be implemented by the system 1000 shown in Figure 1. Please refer to Figures 1 and 2 simultaneously.

[0008] In step S23, the traffic analysis module 1300 can use traffic data associated with the enterprise private network to analyze whether abnormal traffic will occur.

[0009] In step S25, when the traffic analysis module 1300 analyzes that abnormal traffic will occur, the abnormal network segment location module 1400 can use the configuration data associated with the enterprise private network to locate the abnormal network segment.

[0010] In step S27, the unblocking decision-making module 1500 can use traffic data and abnormal network segments to determine the unblocking decision of the enterprise private network.

[0011] Figure 3 is an example implementation of the system 1000 shown in Figure 1. Figures 4A, 4B, and 5 are further illustrations of Figure 3. Please refer to Figures 1, 2, 3, 4A, 4B, and 5 simultaneously.

[0012] In one embodiment, the traffic data may include the current traffic volume. In step S21, the data collection module 1100 may receive the current traffic volume and configuration data from the enterprise private network device 2000. For example, the data collection module 1100 may receive or poll the current traffic volume and configuration data of the enterprise private network device 2000, for example, through NetFlow, sFlow, or SNMP, but the present invention is not limited thereto. Then, the data collection module 1100 may store the current traffic volume and configuration data in the storage module 1800.

[0013] In one embodiment, the traffic data may include historical traffic volume, current traffic volume, and future traffic volume. In step S22, the traffic prediction module 1200 can use historical traffic volume and current traffic volume to predict future traffic volume. Specifically, the traffic prediction module 1200 can be an AI module. The traffic prediction module 1200 can perform time-series regression using machine learning regression algorithms (such as random forest regression algorithms). Alternatively, the traffic prediction module 1200 can predict future traffic volume using deep learning algorithms for solving time series problems, such as RNN recurrent neural networks, LSTM long short-range memory, GRU gated recurrent units, etc. Here, it is assumed that the traffic prediction module 1200 uses random forest regression for time series prediction, with 5 minutes of traffic data as a unit. The traffic prediction module 1200 can predict the next 7 data points using the first 10 data points and store the predicted future traffic volume in the storage module 1800.

[0014] In step S23, the traffic analysis module 1300 can use traffic data associated with the enterprise private network to analyze whether abnormal traffic will occur. In one embodiment, the traffic analysis module 1300 can use traffic data associated with the enterprise private network, overload threshold, and number of overload exceedances to analyze whether abnormal traffic will occur. Specifically, the traffic analysis module 1300 can use parameters such as traffic data within a time interval (including historical, current, and future), overload threshold, and number of overload exceedances to analyze whether abnormal traffic will occur. Here, it is assumed that the traffic analysis module 1300 analyzes the traffic data of base station BS1, and the time interval is 7 future points (35 minutes) and 4 current points (20 minutes). Further, it is assumed that the traffic analysis module 1300 analyzes that base station BS1 exceeds 90% of the overload threshold for 4 consecutive data points in the future. Therefore, the traffic analysis module 1300 can analyze that a traffic overload anomaly will occur in the future.

[0015] Next, in step S25, when the traffic analysis module 1300 analyzes that abnormal traffic will occur, the abnormal network segment location module 1400 can use the configuration data associated with the enterprise private network to locate the abnormal network segment. Assume that the abnormal network segment (bottleneck) located by the abnormal network segment location module 1400 is base station BS1. For ease of explanation, the following embodiment assumes that base station BS2 still has traffic space to receive "directed traffic".

[0016] In step S27, the unblocking decision-making module 1500 can use traffic data and abnormal network segments to determine the unblocking decision of the enterprise private network. Specifically, the unblocking decision-making module 1500 and the simulation verification module 1600 can work together to locate which of the following categories the abnormal network segment belongs to based on configuration data:

[0017] (1) Terminal -> Base Station -> MEC Edge Cloud;

[0018] (2) MEC edge cloud -> core network;

[0019] (3) Nuclear network -> Application services.

[0020] It should be noted that (1) to (3) above can each correspond to different decision-making processes and traffic diversion simulation operations. The following will continue to explain.

[0021] [(1) Terminal -> Base Station -> MEC Edge Cloud]

[0022] In one embodiment, the abnormal network segment may include a terminal, a base station, and an MEC edge cloud (i.e., step T00 shown in FIG3). The traffic diversion decision may include a diversion decision and a spread spectrum decision. The base station may include a traffic exceeding the overload threshold base station and a traffic not reaching the overload threshold base station. The simulation verification module 1600 may perform a diversion simulation operation to divert the terminal from the traffic exceeding the overload threshold base station to the traffic not reaching the overload threshold base station. When the diversion simulation operation is successful, the traffic diversion decision module 1500 may determine the diversion decision. On the other hand, when the diversion simulation operation fails, the simulation verification module 1600 may perform a spread spectrum simulation operation. When the spread spectrum simulation operation is successful, the traffic diversion decision module 1500 may determine the spread spectrum decision. In detail, when the abnormal network segment location module 1400 locates the abnormal network segment as belonging to "(1) terminal -> base station -> MEC edge cloud" in step S25, in step T01, the traffic diversion decision module 1500 can analyze the traffic of each terminal on the base station that has reached the heavy load threshold, and sort them according to the traffic size, starting with the terminal with the most traffic, and perform traffic diversion simulation operation in sequence. Here it is assumed that for the base station BS1 that has reached the heavy load threshold, the traffic diversion decision module 1500 analyzes and sorts the traffic size of each terminal on the base station BS1 as terminal UE01, terminal UE02, terminal UE03... Then, the simulation verification module 1600 can start to perform traffic diversion simulation operation.

[0023] In step T02, the simulation verification module 1600 can identify the base station with a low traffic load threshold that covers the location of the terminal based on the configuration data and the terminal's historical connection records. If there are multiple base stations with low traffic load thresholds, the base station with the lightest load (traffic volume) is selected. Assume that the simulation verification module 1600 finds that terminal UE01 is within the signal coverage area of ​​base station BS2, and that base station BS2 is the base station with the lowest load and low traffic load threshold. Furthermore, assume that the simulation verification module 1600 determines that the remaining traffic space of base station BS2 (500MBps) is greater than the 100Mbps required by terminal UE01.

[0024] In step T03, if the remaining traffic space of the base station is greater than the traffic demand of the terminal, the simulation verification module 1600 can perform a traffic diversion simulation operation. As mentioned above, assume that the simulation verification module 1600 performs a traffic diversion simulation operation from base station BS1 to base station BS2 for terminal UE01. Furthermore, assume that the simulation verification module 1600 finds that the heavy load on base station BS1 can be successfully relieved (i.e., the traffic on base station BS1 will be lower than the heavy load threshold), and the traffic on base station BS2 is still lower than the heavy load threshold.

[0025] In step T04, if the result of the diversion simulation operation has resolved the overload of the base station whose traffic exceeds the overload threshold, the diversion decision module 1500 can send out a diversion decision. Continuing with the above embodiment, the diversion decision sent by the diversion decision module 1500 is a diversion decision to "dive terminal UE01 from base station BS1 to base station BS2".

[0026] In step S29, the optimization control module 1700 can transmit a diversion decision to the enterprise private network device 2000. Continuing with the above embodiment, the optimization control module 1700 can transmit a diversion decision to the enterprise private network device 2000. In other words, the optimization control module 1700 can send instructions to the corresponding controller to make diversion decisions or spread spectrum decisions, etc. Figure 4B further illustrates the diversion decision "diverting terminal UE01 from base station BS1 to base station BS2".

[0027] After completing step S29, step S21 can be executed again. In other words, the data collection module 1100, traffic prediction module 1200, traffic analysis module 1300, abnormal network segment location module 1400, smoothing decision-making module 1500, simulation verification module 1600, and optimization control module 1700 shown in Figure 1 can form a closed-loop optimization of the enterprise private network to achieve a low-latency autonomous enterprise private network.

[0028] Please refer back to Figure 4A. On the other hand, if the result of the diversion simulation operation cannot resolve the overload of the base station whose traffic exceeds the overload threshold, and the simulation verification module 1600 has completed the diversion simulation operation for all terminals, then in step T05, the simulation verification module 1600 can perform a spread spectrum simulation operation. If the result of the spread spectrum simulation operation has resolved the overload of the base station whose traffic exceeds the overload threshold, then the relief decision module 1500 can send out a relief decision (i.e., send out a spread spectrum decision). If the result of the spread spectrum simulation operation cannot resolve the overload of the base station whose traffic exceeds the overload threshold, then in step T06, the relief decision module 1500 can notify the administrator that the overload cannot be automatically relieved.

[0029] [(2)MEC Edge Cloud -> Core Network]

[0030] Abnormal network segments may include MEC edge cloud and core network (i.e., step U00 shown in Figure 3). The smoothing decision may include spread spectrum decision. When traffic data indicates that the used bandwidth of the core network is greater than the heavy load threshold, the simulation verification module 1600 may perform a spread spectrum simulation operation. When the spread spectrum simulation operation is successful, the smoothing decision decision module 1500 may determine the spread spectrum decision.

[0031] [(3) Nuclear Network -> Application Services]

[0032] In one embodiment, the abnormal network segment may include the core network and application services (i.e., step V00 shown in FIG3). The smoothing decision may include traffic diversion decision and spread spectrum decision. The simulation verification module 1600 may perform traffic diversion simulation operation. When the traffic diversion simulation operation is successful, the smoothing decision decision module 1500 may determine the traffic diversion decision. On the other hand, when the traffic diversion simulation operation fails, the simulation verification module 1600 may perform spread spectrum simulation operation. When the spread spectrum simulation operation is successful, the smoothing decision decision module 1500 may determine the spread spectrum decision. In detail, when the abnormal network segment location module 1400 locates the abnormal network segment as belonging to "(3) core network -> application services" in step S25, in step V01, the smoothing decision decision module 1500 may analyze the traffic of each user on the circuit that has reached the heavy load threshold, and sort them according to the traffic size, starting with the user with the most traffic, and sequentially perform traffic diversion simulation operation.

[0033] In step V02, the simulation verification module 1600 can find possible routing circuit information based on the configuration data.

[0034] In step V03, if the remaining traffic space of the circuit is greater than the user's traffic demand, the simulation verification module 1600 can perform a traffic diversion simulation operation.

[0035] In step V04, if the result of the flow diversion simulation operation has resolved the heavy load of the circuit, the dredging decision module 1500 can send out the dredging decision.

[0036] On the other hand, if the result of the current diversion simulation operation cannot resolve the circuit's overload, and the simulation verification module 1600 has completed the current diversion simulation operation for all terminals, then in step V05, the simulation verification module 1600 may perform a spread spectrum simulation operation. If the result of the spread spectrum simulation operation resolves the circuit's overload, then the unblocking decision module 1500 may send out an unblocking decision (i.e., send out a spread spectrum decision). If the result of the spread spectrum simulation operation cannot resolve the circuit's overload, then in step V06, the unblocking decision module 1500 may notify the administrator that the overload cannot be automatically unblocked.

[0037] In summary, the system and method for optimizing enterprise private networks of the present invention can utilize the traffic data of the enterprise private network to analyze whether abnormal traffic will occur, and can utilize the configuration data of the enterprise private network to locate abnormal network segments. Furthermore, the present invention can use traffic data and abnormal network segments to determine the network outage response. Based on this, the present invention can form a closed-loop optimization system for enterprise private networks to achieve a low-latency autonomous enterprise private network. [Simplified Explanation of the Diagram]

[0005] Figure 1 is a schematic diagram of a system for optimizing an enterprise private network according to an embodiment of the present invention. Figure 2 is a flowchart of a method for optimizing an enterprise private network according to an embodiment of the present invention. Figure 3 is an example of an implementation of the system shown in Figure 1. Figures 4A, 4B and 5 are further illustrations of Figure 3.

Claims

1. A system for optimizing enterprise private networks, comprising: Traffic analysis module; Abnormal network segment location module; The system also includes a traffic management decision-making module, wherein the traffic analysis module uses traffic data associated with the enterprise private network to analyze whether abnormal traffic will occur; when the traffic analysis module analyzes that the abnormal traffic will occur, the abnormal network segment location module uses configuration data associated with the enterprise private network to locate the abnormal network segment; the traffic management decision-making module uses the traffic data and the abnormal network segment to determine the traffic management decision for the enterprise private network, wherein the abnormal network segment includes a first abnormal network segment, a second abnormal network segment, and a third abnormal network segment, wherein the first abnormal network segment includes terminals, base stations, and MEC edge cloud, wherein the second abnormal network segment includes the MEC edge cloud and core network, wherein the third abnormal network segment includes the core network and application services, wherein the system further includes a simulation verification module, wherein when the abnormal network segment is the second abnormal network segment, the traffic management decision includes a spread spectrum decision, and the system performs the following: when the traffic data indicates that the used bandwidth of the core network is greater than the overload threshold, the simulation verification module performs a spread spectrum simulation operation; When the spread spectrum simulation operation is successful, the dredging decision module determines the spread spectrum decision.

2. The system as described in claim 1 further includes a data collection module, wherein the traffic data includes current traffic volume, and wherein the data collection module receives the current traffic volume and the configuration data from the enterprise private network equipment.

3. The system as described in claim 1 further includes a traffic prediction module, wherein the traffic data includes historical traffic volume, current traffic volume, and future traffic volume, wherein the traffic prediction module uses the historical traffic volume and the current traffic volume to predict the future traffic volume.

4. The system as described in claim 1, wherein the traffic analysis module uses traffic data associated with the enterprise private network, overload thresholds, and overload exceedance counts to analyze whether the abnormal traffic will occur.

5. The system as described in claim 1, wherein the base station includes a traffic overload threshold base station and a traffic underload threshold base station, wherein when the abnormal network segment is the first abnormal network segment, the traffic diversion decision includes a diversion decision and the spread spectrum decision, and the system performs the following: the simulation verification module performs a diversion simulation operation to divert the terminal from the traffic overload threshold base station to the traffic underload threshold base station; when the diversion simulation operation is successful, the traffic diversion decision module determines the diversion decision; when the diversion simulation operation fails, the simulation verification module performs the spread spectrum simulation operation; when the spread spectrum simulation operation is successful, the traffic diversion decision module determines the spread spectrum decision.

6. The system as described in claim 1, wherein when the abnormal network segment is the third abnormal network segment, the unblocking decision includes a diversion decision and the spread spectrum decision, and the system performs the following: the simulation verification module performs a diversion simulation operation; when the diversion simulation operation is successful, the unblocking decision decision module determines the diversion decision; when the diversion simulation operation fails, the simulation verification module performs the spread spectrum simulation operation; when the spread spectrum simulation operation is successful, the unblocking decision decision module determines the spread spectrum decision.

7. The system as described in claim 1, wherein the enterprise private network equipment is associated with mobile terminal resources, base station resources, MEC resources, core network resources, server resources, and network equipment resources.

8. The system as described in claim 1 further includes an optimization control module, wherein the optimization control module transmits the dredging decision to the enterprise private network equipment.

9. A method for optimizing an enterprise private network, suitable for a system including a traffic analysis module, an abnormal network segment location module, and a troubleshooting decision module, wherein the method includes the following steps: the traffic analysis module analyzes whether abnormal traffic will occur using traffic data associated with the enterprise private network; when the traffic analysis module analyzes that the abnormal traffic will occur, the abnormal network segment location module locates the abnormal network segment using configuration data associated with the enterprise private network; and the troubleshooting decision module determines a troubleshooting decision for the enterprise private network using the traffic data and the abnormal network segment, wherein the abnormal network segment includes a first abnormal network segment, a second abnormal network segment, and a third abnormal network segment, wherein the first abnormal network segment includes a terminal, a base station, and an MEC edge cloud, wherein the second abnormal network segment includes the MEC edge cloud and a core network, and wherein the third abnormal network segment includes the core network and application services, wherein the system further includes a simulation verification module, wherein when the abnormal network segment is the second abnormal network segment, the troubleshooting decision includes a spread spectrum decision, and the system performs the following: When the traffic data indicates that the used bandwidth of the nuclear network is greater than the heavy load threshold, the simulation verification module performs a spread spectrum simulation operation; when the spread spectrum simulation operation is successful, the unblocking decision-making module determines the spread spectrum decision.

Images