Data center remote operation method and system based on acquisition gateway analysis

Abstract

The invention relates to the technical field of data center remote operation, and discloses a data center remote operation method and system based on collection gateway analysis, and the method comprises the steps: obtaining gateway resource data and data flow priority attributes; quantifying according to the resource data to obtain a cache length; when the cache length is greater than a length threshold value, processing by adopting a dynamic load balancing algorithm to obtain a migration data stream; analyzing according to the data stream to obtain a timestamp and a serial number; measuring according to the timestamp and the serial number to obtain transmission delay and packet loss rate; performing evaluation according to the transmission time delay and the packet loss rate, and performing link dynamic switching to obtain a backup transmission path when an evaluation result is a main link fault; according to the backup transmission path, adopting a distributed routing protocol to calculate and obtain a data transmission path; monitoring according to the data transmission path to obtain monitoring data; predicting according to the monitoring data to obtain an adjustment result; and performing optimization according to an adjustment result, and dynamically switching the main link and the standby link to perform data transmission. The method can solve the problem of low overall data transmission efficiency.

Description

Technical Field

[0001] The present invention relates to the technical field of remote operation of data centers, and in particular to a method and system for remote operation of data centers based on analysis of a collection gateway. Background Art

[0002] Currently, in the scenario of remote operation of data centers, the collection gateway, as a key node connecting the data center and remote terminals, its core task is to ensure the real-time transmission of data in different network environments. However, with the increasing complexity of network topologies and data transmission requirements, the gateway faces multiple technical challenges in actual operation. First, the gateway needs to process data streams from different terminals in real time while coping with uncertain factors such as network latency and bandwidth fluctuations, which requires the gateway to have high-efficiency real-time performance.

[0003] In an existing technology, the collection gateway is used as a key node connecting the data center and remote terminals. The gateway processes data streams from different terminals in real time, and dynamically adjusts the task allocation weights between gateways, and the topological relationship between gateways will change according to network failures or configuration changes. However, the real-time performance of the gateway is limited by its hardware resources and load conditions. When the load of a single gateway is too high, the real-time performance of data transmission will significantly decline. The dynamic adjustment of task allocation weights between gateways depends on the accurate perception of the real-time performance and load conditions of the gateways. Especially in the case of sudden traffic or network anomalies and rapid changes, it is difficult for the task allocation strategy to be adapted in real time.

[0004] In summary, there is a problem of low overall data transmission efficiency in the existing technology. Summary of the Invention

[0005] The present invention provides a method and system for remote operation of data centers based on analysis of a collection gateway to solve the problem of low overall data transmission efficiency in the existing technology.

[0006] In a first aspect, to solve the above technical problem, the present invention provides a method for remote operation of a data center based on analysis of a collection gateway, including:

[0007] Obtain gateway resource data and data stream priority attributes;

[0008] Perform a quantitative evaluation based on the gateway resource data and the data stream priority attributes to obtain the cache length of the gateway queue;

[0009] When the cache length is greater than a preset length threshold, use a dynamic load balancing algorithm for processing to obtain a migrated data stream;

[0010] Analyze the migrated data stream to obtain the timestamp and sequence number of the data stream;

[0011] Measure according to the timestamp and the serial number to obtain the transmission delay and packet loss rate of the primary and backup links;

[0012] Evaluate according to the transmission delay and the packet loss rate. When the evaluation result is that the primary link fails, perform link dynamic switching to obtain a backup transmission path;

[0013] Calculate according to the backup transmission path using a distributed routing protocol to obtain a first data transmission path;

[0014] Monitor according to the first data transmission path to obtain monitoring data;

[0015] Predict according to the monitoring data using a supervised learning algorithm to obtain an adjustment result;

[0016] Optimize according to the adjustment result to obtain an optimal data transmission path for data transmission.

[0017] In an alternative embodiment, the quantifying and evaluating according to the gateway resource data and the data flow priority attribute to obtain the cache length of the gateway queue includes:

[0018] Construct according to the gateway resource data to obtain a gateway resource utilization matrix;

[0019] Calculate according to the gateway resource utilization matrix using the analytic hierarchy process to obtain a matrix weight coefficient;

[0020] Calculate according to the matrix weight coefficient to obtain a task processing score value of the gateway;

[0021] Schedule according to the task processing score value using a weighted round-robin algorithm to obtain an optimal task allocation scheme;

[0022] Perform real-time monitoring according to the optimal task allocation scheme to obtain the cache length of the gateway queue.

[0023] In an alternative embodiment, when the cache length is greater than a preset length threshold, a dynamic load balancing algorithm is used for processing to obtain a migrated data flow, including:

[0024] Obtain the load value and CPU value of the current gateway;

[0025] When the cache length is greater than a preset length threshold, extract according to the cache length to obtain a priority attribute;

[0026] Calculate according to the priority attribute using a weighted round-robin strategy to obtain a round-robin value;

[0027] Calculate according to the load value and the CPU value to obtain a high-load gateway and a low-load gateway;

[0028] Migrate data from the high-load gateway to the low-load gateway according to the polling value to obtain a migration data stream.

[0029] In an alternative embodiment, the evaluating according to the transmission delay and the packet loss rate, and when the evaluation result is that the primary link fails, performing link dynamic switching to obtain a backup transmission path, includes:

[0030] Judge according to the transmission delay and the packet loss rate to obtain the current link state;

[0031] When the current link state is greater than a preset state threshold, determine that the primary link fails to obtain a link switching mechanism;

[0032] Perform link switching according to the link switching mechanism to obtain a primary transmission path;

[0033] Monitor according to the primary transmission path to obtain a data synchronization state;

[0034] Calculate according to the data synchronization state to obtain a synchronization delay;

[0035] When the synchronization delay is within a preset delay range, determine that the link switching is successful to obtain a backup transmission path.

[0036] In an alternative embodiment, the calculating according to the backup transmission path by using a distributed routing protocol to obtain a first data transmission path, includes:

[0037] Calculate according to the backup link by using a distributed routing protocol to obtain the connectivity status data and topological structure information of each network node in the link;

[0038] Calculate according to the connectivity status data to obtain a network link state;

[0039] Judge according to the network link state and the topological structure information to obtain the logical relationship between the primary link and the backup link;

[0040] Connect according to the logical relationship to obtain a primary transmission path;

[0041] Perform real-time monitoring according to the primary transmission path to obtain connectivity data;

[0042] When the connectivity data is less than a preset connectivity threshold, perform the logical relationship adjustment to obtain a first data transmission path.

[0043] In an alternative embodiment, the optimizing according to the adjustment result to obtain an optimal data transmission path for data transmission, includes:

[0044] Obtain redundant link status information;

[0045] Quantize according to the adjustment result to obtain an initial weight matrix;

[0046] Perform dynamic adjustment according to the initial weight matrix to obtain a weight matrix coefficient;

[0047] Calculate according to the weight matrix coefficient to obtain a queue cache weight;

[0048] Calculate according to the redundant link status information to obtain a data synchronization period parameter;

[0049] Iteratively optimize the initial weight matrix according to the queue cache weight and the data synchronization period parameter to obtain an optimal data transmission path for data transmission.

[0050] In an alternative embodiment, the iteratively optimizing the initial weight matrix according to the queue cache weight and the data synchronization period parameter to obtain an optimal data transmission path for data transmission includes:

[0051] Calculate according to the queue cache weight and the data synchronization period parameter to obtain a quality of service index;

[0052] Monitor according to the quality of service index to obtain a quality of service metric;

[0053] When the quality of service metric is less than the metric threshold, iteratively optimize the initial weight matrix to obtain an optimal data transmission path for data transmission.

[0054] In a second aspect, the present invention provides a data center remote operation system based on acquisition gateway analysis, including:

[0055] A data acquisition module, configured to acquire gateway resource data and data stream priority attributes;

[0056] A data evaluation module, configured to perform quantization evaluation according to the gateway resource data and the data stream priority attributes to obtain the cache length of the gateway queue;

[0057] A data processing module, configured to, when the cache length is greater than a preset length threshold, perform processing using a dynamic load balancing algorithm to obtain a migrated data stream;

[0058] A data analysis module, configured to analyze according to the migrated data stream to obtain the timestamp and sequence number of the data stream;

[0059] A data measurement module, configured to measure according to the timestamp and the sequence number to obtain the transmission delay and packet loss rate of the primary and standby links;

[0060] A link switching module, which is used to evaluate according to the transmission delay and the packet loss rate. When the evaluation result is that the primary link fails, it performs dynamic link switching to obtain a backup transmission path;

[0061] A data calculation module, which is used to calculate according to the backup transmission path by using a distributed routing protocol to obtain a first data transmission path;

[0062] A data monitoring module, which is used to monitor according to the first data transmission path to obtain monitoring data;

[0063] A data prediction module, which is used to predict according to the monitoring data by using a supervised learning algorithm to obtain an adjustment result;

[0064] A data optimization module, which is used to optimize according to the adjustment result to obtain an optimal data transmission path for data transmission.

[0065] In a third aspect, the present invention also provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the data center remote operation method based on the acquisition gateway analysis described in any one of the above.

[0066] In a fourth aspect, the present invention also provides a computer-readable storage medium, which includes a stored computer program. Wherein, when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the data center remote operation method based on the acquisition gateway analysis described in any one of the above.

[0067] Compared with the prior art, the present invention has the following beneficial effects:

[0068] (1) The present invention conducts a quantitative evaluation according to the gateway resource data and the data stream priority attribute to obtain the cache length of the data. By constructing a gateway resource utilization matrix, calculating the real-time task processing ability score value of the gateway, and using a weighted round-robin algorithm to schedule gateway tasks, it quantitatively evaluates the cache length and dynamically allocates tasks, optimizes the resource utilization rate of the gateway, reduces the pressure on high-load gateways, and improves the overall data transmission efficiency.

[0069] (2) The present invention evaluates according to the transmission delay and the packet loss rate. When the evaluation result is that the primary link fails, it performs dynamic link switching to obtain a backup transmission path; through a real-time monitoring and dynamic link switching mechanism, it ensures that when a network failure or link quality degradation occurs, it can quickly switch to the backup link, reducing data transmission delay and improving real-time performance.

[0070] (3) The present invention calculates a data transmission path according to the backup link by using a distributed routing protocol; and dynamically adjusts the data transmission path through the distributed routing protocol and real-time monitoring data, thereby avoiding resource waste and bottleneck problems and improving the utilization efficiency of network resources.

[0071] (4) The present invention optimizes according to the adjustment result to obtain an optimal data transmission path for data transmission; and through iterative optimization and a prediction mechanism, can cope with burst traffic and network anomalies in advance, reduce performance fluctuations, and enhance the stability and risk resistance of critical services. Description of the Drawings

[0072] Figure 1 is a schematic flowchart of a method for remote operation of a data center based on analysis of a collection gateway provided in the first embodiment of the present invention;

[0073] Figure 2 is a schematic structural diagram of remote operation of a data center provided by the present invention;

[0074] Figure 3 is a schematic structural diagram of a device for remote operation of a data center based on analysis of a collection gateway provided in the second embodiment of the present invention. Detailed Embodiments

[0075] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0076] Referring to Figure 1 , the first embodiment of the present invention provides a method for remote operation of a data center based on analysis of a collection gateway, including the following steps:

[0077] S11. Obtain gateway resource data and data stream priority attributes;

[0078] S12. Perform a quantitative evaluation according to the gateway resource data and the data stream priority attributes to obtain the cache length of the gateway queue;

[0079] S13. When the cache length is greater than a preset length threshold, perform processing by using a dynamic load balancing algorithm to obtain a migrated data stream;

[0080] S14. Analyze according to the migrated data stream to obtain the timestamp and sequence number of the data stream;

[0081] S15. Measure according to the timestamp and the sequence number to obtain the transmission delay and packet loss rate of the primary and backup links;

[0082] S16. Evaluate according to the transmission delay and the packet loss rate. When the evaluation result indicates that the primary link fails, perform link dynamic switching to obtain a backup transmission path.

[0083] S17. Calculate according to the backup transmission path using a distributed routing protocol to obtain a first data transmission path.

[0084] S18. Monitor according to the first data transmission path to obtain monitoring data.

[0085] S19. Predict according to the monitoring data using a supervised learning algorithm to obtain an adjustment result.

[0086] S110. Optimize according to the adjustment result to obtain an optimal data transmission path for data transmission.

[0087] In step S11, obtain gateway resource data and data stream priority attributes.

[0088] It should be noted that the gateway resource data mainly includes CPU utilization rate, memory occupancy rate, and I / O throughput. The CPU utilization rate and memory occupancy rate are obtained by deploying a resource monitoring agent (Agent) in the operation of each gateway node to collect the usage of hardware resources in real time, and then obtaining the CPU usage time (user state, kernel state, idle time, etc.) and the total memory, used memory, and free memory by calling (such as / proc / stat in Linux), and then calculating. Among them, the agent can be implemented using open-source tools (such as Telegraf, Collectd) or custom scripts. The disk and network I / O throughput are obtained through tools (such as iostat in Linux) or custom scripts, including disk read and write rates (MB / s) and network sending and receiving rates (Mbps). The collected resource data is transmitted to the central management node through a lightweight communication protocol (such as MQTT, gRPC), and the data is stored in a high-performance database (such as InfluxDB, Redis) to support real-time query and analysis.

[0089] It should be noted that the data flow priority attribute is an important basis for dynamic task allocation, mainly including the 5-tuple characteristics of the data flow and QoS requirements. A traffic analysis module is deployed at the network traffic entrance of the gateway to capture and analyze data packets. High-performance data packet processing frameworks (such as DPDK, PF_RING) or open-source tools (such as nDPI, Wireshark) can be used to obtain the data flow priority attribute. All data packets passing through the gateway are captured through the promiscuous mode of the network card. The data packets are parsed to extract the 5-tuple characteristics including the source IP address, destination IP address, source port number, destination port number, and protocol type (such as TCP, UDP). The QoS requirements are identified according to the DSCP field (Differentiated Services Code Point) in the data packet. If the DSCP field is unavailable, the QoS requirements are inferred according to the application layer protocol (such as HTTP, FTP, VoIP). For example, VoIP data flows usually require low latency and high priority, while file transfer data flows have higher bandwidth requirements. According to the QoS requirements and business needs, the data flows are divided into multiple priority categories (such as high, medium, low). For example, real-time video conferencing and VoIP can be set as high priority, online games and real-time data synchronization can be set as medium priority; file transfer and backup data can be set as low priority. The extracted data flow characteristics and priority attributes are stored in the central management node or distributed database. The gateway resource data and the data flow priority attribute are integrated to form complete context information.

[0090] In step S12, a quantization evaluation is performed according to the gateway resource data and the data flow priority attribute to obtain the cache length of the gateway queue.

[0091] In one implementation, the quantization evaluation according to the gateway resource data to obtain the cache length of the gateway queue includes:

[0092] Construct according to the gateway resource data to obtain a gateway resource utilization matrix;

[0093] Calculate using the analytic hierarchy process according to the gateway resource utilization matrix to obtain the matrix weight coefficient;

[0094] Calculate according to the matrix weight coefficient to obtain the task processing score value of the gateway;

[0095] Schedule using the weighted round-robin algorithm according to the task processing score value to obtain an optimal task allocation scheme;

[0096] Perform real-time monitoring according to the optimal task allocation scheme to obtain the cache length of the gateway queue.

[0097] Specifically, combining the data flow priority attributes of the data flow, these metrics are combined with the preset quality of service requirements to form a resource utilization matrix. Assume that the QoS requirements of the data flow include metrics such as latency and bandwidth, and the priority attributes are divided into three levels: high, medium, and low, with different weight values assigned respectively. The weight coefficients of CPU utilization, memory occupancy, and I / O throughput in the gateway resource utilization matrix are calculated through the Analytic Hierarchy Process (AHP). The Analytic Hierarchy Process is a commonly used multi-criteria decision analysis method. By constructing a judgment matrix, pairwise comparisons of each metric are made to determine their relative importance. For example, through comparative analysis, it is obtained that the weight of processor utilization is 0.5, the weight of memory occupancy is 0.3, and the weight of disk throughput is 0.2. This weight distribution reflects the dominant role of processor resources in task processing ability during actual operation. The collected gateway resource data is input into the scoring model. After normalizing each metric according to the matrix weight coefficients, the real-time task processing ability score value of the gateway is calculated. The construction of the scoring model is based on the gateway resource utilization matrix and weight coefficients. Its core is to combine the resource utilization metrics with the weight coefficients and calculate the task processing score value of the gateway through weighted summation. Specifically, the calculation process is that the score value is equal to the CPU weight multiplied by the CPU utilization value plus the memory occupancy multiplied by the memory weight value plus the I / O throughput multiplied by the I / O weight value. The training process of the scoring model is a key link to ensure its accuracy and adaptability. First, a large amount of historical operation data of gateways needs to be collected, including resource utilization metrics, data flow priority attributes, and corresponding performance manifestations (such as task processing latency, cache hit rate, etc.). Then, the collected data is preprocessed through feature engineering to extract key features (such as normalized resource utilization metrics). Next, machine learning algorithms (such as linear regression, random forest, etc.) are used to train the data to obtain the weight coefficients. For example, through a linear regression model, the influence degree of each resource metric on task processing ability can be learned, and the weight coefficients can be dynamically adjusted. Finally, the performance of the model is evaluated through methods such as cross-validation, and the model parameters are further optimized to ensure the accuracy and real-time nature of the scoring results. The training and optimization of the scoring model is an iterative process that enables the model to better adapt to the complex and changing network environment.

[0098] According to the gateway processing capacity score value and the data flow priority attribute, the weighted round-robin algorithm is used to schedule gateway tasks to obtain the optimal task allocation scheme. In the scenario of remote operation in the data center, the gateway, as a key node connecting the data center and remote terminals, is responsible for receiving data flows from various remote terminals (such as sensors, devices, user terminals, etc.). The generation of gateway tasks starts from the access of data flows. The weighted round-robin algorithm is a common load balancing algorithm. According to the priority of tasks and the gateway processing capacity score value, different weights are assigned to each gateway, and tasks are allocated according to the weight ratio. For example, for high-priority data flows, such as the environmental monitoring data flow with a priority of 3, the device status data flow with a priority of 2, and the log data flow with a priority of 1, the weighted round-robin algorithm will allocate more processing time to high-priority tasks according to the priority and gateway score. During the task scheduling process, the cache hit rate of the gateway is monitored in real time. The cache hit rate refers to the proportion of directly obtaining data from the cache, reflecting the effectiveness of the cache mechanism. If the cache hit rate is lower than the preset threshold, such as 60%, the cache optimization mechanism is triggered to adjust the cache policy. For example, by analyzing recent access patterns, frequently accessed environmental data is preferentially cached to reduce the disk access frequency. At the same time, the KNN algorithm is used to analyze the historical task processing data of the gateway to predict the resource usage trend of the gateway and perform resource allocation in advance. Finally, the random forest algorithm is used to jointly evaluate the task scheduling efficiency and cache hit rate of the gateway to generate a gateway performance optimization report to guide subsequent resource allocation and task scheduling strategy adjustment.

[0099] In step S13, when the cache length is greater than the preset length threshold, the dynamic load balancing algorithm is used for processing to obtain the migrated data flow.

[0100] In one implementation, the when the cache length is greater than the preset length threshold, the dynamic load balancing algorithm is used for processing to obtain the migrated data flow, includes:

[0101] When the cache length is greater than the preset length threshold, extraction is performed according to the cache length to obtain the priority attribute;

[0102] Calculation is performed according to the priority attribute using the weighted round-robin strategy to obtain the round-robin value;

[0103] Obtain the load value and CPU value of the current gateway;

[0104] Calculation is performed according to the load value and the CPU value to obtain the high-load gateway and the low-load gateway;

[0105] Migrate the data from the high-load gateway to the low-load gateway according to the round-robin value to obtain the migrated data flow.

[0106] Specifically, during the operation of the gateway in the data center, continuously monitor the cache length of each gateway queue, which is the trigger condition for the dynamic load balancing algorithm. The cache length refers to the number of data packets waiting to be processed in the gateway queue. When it exceeds the preset threshold, it indicates that the gateway faces an overload risk. For example, assume that the preset cache length threshold is 1000 data packets. The current cache length will be compared with this threshold in real time. If the cache length of a certain gateway reaches 1200 data packets, exceeding the preset value, it will be determined that the gateway is overloaded, thus triggering the dynamic load balancing algorithm to relieve its pressure and optimize the overall resource allocation. When the cache length exceeds the preset threshold, extract the priority attribute from the data stream to determine the importance and urgency of the data stream. The priority attribute is a key feature of the data stream and is usually divided into three levels: high (3), medium (2), and low (1). For example, the real-time monitoring data stream is marked as high priority (3) due to its high real-time requirements, the device status data stream is medium priority (2), and the log data stream is low priority (1). By parsing the header information of the data stream or according to the preset rules, accurately extract these priority attributes to provide a basis for subsequent data stream scheduling and migration, ensuring that high-priority data streams obtain priority during processing and migration.

[0107] Based on the extracted priority attributes, use the weighted round-robin strategy to calculate the round-robin value of each data stream. Weighted round-robin is a commonly used load balancing algorithm that determines the priority of data streams in migration by assigning weights to data streams with different priorities. For example, set the weight of high priority (3) to 4, the weight of medium priority (2) to 2, and the weight of low priority (1) to 1. Assume that there are three data streams in a certain gateway queue: Data stream A (priority 3, cache length 200), Data stream B (priority 2, cache length 300), and Data stream C (priority 1, cache length 700). Calculate the round-robin values according to the weights: the round-robin value of Data stream A is 200 multiplied by 4, which is 800; the round-robin value of Data stream B is 300 multiplied by 2, which is 600; and the round-robin value of Data stream C is 700 multiplied by 1, which is 700. The higher the round-robin value, the higher the priority of the data stream in migration, and the data stream with a higher round-robin value will be processed first. To accurately evaluate the load status of the gateway, obtain the current load value and CPU value of the gateway in real time. The load value reflects the current busy degree of the gateway, usually expressed as a percentage, while the CPU value represents the usage rate of the processor. For example, through the gateway monitoring module, real-time data collection shows that the load value of a certain gateway is 85% and the CPU value is 90%, indicating that the gateway is in a high-load state. These data are important bases for judging whether the gateway needs to perform load balancing, and the task allocation will be dynamically adjusted according to these real-time data to optimize the resource utilization efficiency.

[0108] Based on the obtained load values and CPU values, determine which gateways are in a high-load state and which are in a low-load state. For example, the preset load value threshold is 80%, and the CPU value threshold is 85%. If the load value of a certain gateway is 85% and the CPU value is 90%, it is marked as a high-load gateway; while for another gateway with a load value of 50% and a CPU value of 30%, it is marked as a low-load gateway. In this way, it is possible to quickly identify the gateways that need load balancing, providing clear targets and directions for subsequent data migration, thereby achieving reasonable resource allocation. According to the calculated polling value, migrate data from high-load gateways to low-load gateways to achieve load balancing. For example, preferentially migrate the low-priority data stream C (polling value 700) from the high-load gateway A (load value 85%, CPU value 90%) to the low-load gateway B (load value 50%, CPU value 30%). After the migration is completed, update the load values and CPU values of the two gateways to ensure that the impact of the migration process on the business is minimized. Through this dynamic migration mechanism, it is possible to effectively reduce the pressure on high-load gateways, while making full use of the idle resources of low-load gateways to improve the overall performance and stability. After the migration is completed, record the migrated data streams and generate detailed records of the migrated data streams. These records include information such as the source gateway, target gateway, priority attribute, and migration timestamp of the data stream. For example, the record shows that the data stream C is migrated from gateway A to gateway B, with a priority of 1 and a migration time of 10:00:00. These records not only help to monitor the effect of load balancing in real time but also provide data support for subsequent optimization strategies. By analyzing the records of the migrated data streams, it can be found that some data streams are frequently migrated within a specific time period, thereby optimizing the initial allocation strategy of the data streams and further improving the overall performance and service quality.

[0109] In step S14, analyze according to the migrated data stream to obtain the timestamp and sequence number of the data stream.

[0110] During the data migration process, it is necessary to collect detailed transmission information of the migrated data stream. The migrated data stream contains a large number of data packets, and each data packet carries key transmission information. By deploying a monitoring module at the network interface, capture these data packets in real time and extract their core information. For example, when the data stream is migrated from a high-load gateway to a low-load gateway, the monitoring module will record the sending time and receiving time of each data packet, and these time information will be used as the basic data for subsequent analysis.

[0111] The timestamp is a crucial attribute in the process of data packet transmission, which records the specific time point when the data packet is transmitted in the network. By analyzing the timestamp, the transmission delay of the data packet in the network can be accurately calculated. In the migration data stream, the timestamp of each data packet includes the sending timestamp and the receiving timestamp. The sending timestamp refers to the time when the data packet is sent from the source gateway, while the receiving timestamp refers to the time when the data packet arrives at the target gateway. These timestamps are extracted by parsing the header information of the data packet. For example, assuming that the sending timestamp of a data packet is 10:00:00.000 and the receiving timestamp is 10:00:00.020, then the transmission delay of this data packet can be calculated as 20 milliseconds. In this way, the transmission delay of the entire migration data stream can be quantitatively analyzed.

[0112] The sequence number is the unique identifier of the data packet in the data stream, which records the transmission order of the data packet in the network. By analyzing the sequence number, it can be detected whether the data packet is lost or transmitted out of order in the network. In the migration data stream, each data packet has a unique sequence number, and these sequence numbers are sequentially assigned according to the sending order of the data packets. The sequence number is extracted by parsing the header information of the data packet. For example, assuming that the consecutive sequence numbers in the data stream are 1, 2, 3, 5, then it can be found that the data packet with the sequence number 4 is lost. In this way, it can be detected whether packet loss occurs during the network transmission of the data stream, and the packet loss rate can be further calculated. For example, if 2 data packets are lost among 100 consecutive data packets, then the packet loss rate is 2%.

[0113] Based on the extracted timestamp and sequence number information, the link quality between gateways can be comprehensively evaluated. Transmission delay and packet loss rate are two key indicators for measuring link quality. By calculating the transmission delays of multiple data packets, the average delay and the maximum delay of the link can be obtained. At the same time, by detecting the continuity of the sequence numbers, the packet loss rate of the link can be calculated. For example, assuming that within a certain period of time, the average transmission delay of 1000 data packets is detected to be 30 milliseconds, the maximum delay is 50 milliseconds, and the packet loss rate is 1%, then it can be judged that the current link quality is good, but there are certain delay fluctuations. If the packet loss rate is too high or the delay is too large, the link optimization mechanism will be triggered, such as switching to the backup link or adjusting the data transmission path.

[0114] Based on the analysis results of timestamps and sequence numbers, the quality changes of the link can be perceived in real time, and the link can be dynamically adjusted to optimize data transmission. For example, when it is detected that the packet loss rate of the primary link exceeds a preset threshold (such as 2%) or the transmission delay exceeds a preset threshold (such as 100 milliseconds), the link switching mechanism will be triggered. Through the pre-configured backup link, the data traffic can be quickly switched to the standby link to ensure the continuity and reliability of data transmission. At the same time, the transmission delay and packet loss rate of the backup link will be continuously monitored to evaluate its available bandwidth. For example, if the transmission delay of the backup link is 20 milliseconds and the packet loss rate is 0.1%, it can be considered that the quality of the backup link is good and suitable for continuing to carry data traffic. Through this dynamic adjustment mechanism, the high efficiency and stability of data transmission can be maintained in a complex network environment.

[0115] In step S15, measurements are made according to the timestamp and the sequence number to obtain the transmission delay and packet loss rate of the primary and backup links.

[0116] In the network environment of the data center, first, the timestamp information is extracted from the data packet. The timestamp records the key time points of the data packet transmission in the network, including the sending time and the receiving time. By calculating the difference between the sending timestamp and the receiving timestamp, the transmission delay of the data packet can be accurately measured. For example, if the sending timestamp of a data packet is 10:00:00.000 and the receiving timestamp is 10:00:00.020, the transmission delay of this data packet is 20 milliseconds. The acquisition and analysis of timestamps are important bases for evaluating the link transmission efficiency and can help quickly identify link delay problems. In addition to timestamps, the sequence numbers of the data packets are also extracted. The sequence number is used to identify the transmission order of the data packets in the network. By analyzing the continuity of the sequence numbers, the loss of data packets can be detected. For example, if the received sequence numbers are 1, 2, 4, 5, it means that the data packet with sequence number 3 is lost, and the packet loss rate can be calculated accordingly. Assuming that 2 data packets are lost among 100 continuously sent data packets, the packet loss rate is 2%. The analysis of sequence numbers is crucial for evaluating the reliability of the link and can help detect link failures in a timely manner. Based on the analysis results of timestamps, the transmission delays of the primary link and the backup link are calculated respectively. The transmission delay refers to the time required for a data packet to travel from the sending end to the receiving end and is one of the key indicators for measuring the link performance. For example, by measuring the transmission delays of multiple data packets, the average delay of the primary link is calculated to be 30 milliseconds, and the average delay of the backup link is 50 milliseconds. By comparing the delays of the primary and backup links, the real-time performance of the link can be evaluated and a decision-making basis for subsequent link switching can be provided.

[0117] By analyzing the sequence numbers, calculate the packet loss rate of the primary and backup links. The packet loss rate refers to the proportion of lost data packets in the total transmitted data packets during network transmission, reflecting the reliability of the link. For example, by analyzing the packet sequence numbers over a period of time, it is found that the packet loss rate of the primary link is 1% and that of the backup link is 0.5%. A low packet loss rate indicates high stability and reliability of link transmission, while a high packet loss rate implies link failures or network congestion problems. Combine the transmission delay and packet loss rate of the primary and backup links to comprehensively evaluate the overall quality of the link. The evaluation of link quality is achieved by comparing the real-time measured delay and packet loss rate with preset thresholds. For example, the preset delay threshold for the primary link is 50 milliseconds, and the packet loss rate threshold is 1%. If the real-time measured delay of the primary link exceeds 50 milliseconds or the packet loss rate exceeds 1%, it is determined that the primary link is in an abnormal state. This evaluation mechanism can quickly identify link failures or performance degradation, providing a decision basis for link switching. When the transmission delay or packet loss rate of the primary link exceeds the preset threshold, it is determined that the primary link has failed, triggering the link switching mechanism. For example, if the delay of the primary link reaches 80 milliseconds, exceeding the preset threshold of 50 milliseconds, it will quickly switch to the backup link to ensure the continuity and stability of data transmission.

[0118] The implementation of the link switching mechanism needs to consider the smoothness and speed of switching to minimize the impact on services. For example, fast rerouting technology can be adopted to control the switching time within milliseconds, ensuring seamless switching of data transmission. After completing the link switching, it is necessary to confirm the status of the backup link to ensure that it can reliably carry data transmission. By continuously monitoring the transmission delay and packet loss rate of the backup link, evaluate the performance of the backup link. For example, if the delay of the backup link is 30 milliseconds and the packet loss rate is 0.5%, and these metrics are within the disaster recovery requirements (such as a maximum delay of 100 milliseconds and a packet loss rate below 2%), it is confirmed that the backup link switching is successful. Data transmission will continue to use the backup link, and its performance will be monitored in real time to ensure the efficiency and reliability of data transmission.

[0119] In step S16, evaluate based on the transmission delay and the packet loss rate. When the evaluation result indicates that the primary link has failed, perform link dynamic switching to obtain a backup transmission path.

[0120] In one implementation, the evaluating based on the transmission delay and the packet loss rate, and when the evaluation result indicates that the primary link has failed, performing link dynamic switching to obtain a backup transmission path includes:

[0121] Judge according to the transmission delay and the packet loss rate to obtain the current link state;

[0122] When the current link state is greater than the preset state threshold, it is determined that the primary link fails, and a link switching mechanism is obtained;

[0123] Perform link switching according to the link switching mechanism to obtain a primary transmission path;

[0124] Monitor according to the primary transmission path to obtain a data synchronization state;

[0125] Calculate according to the data synchronization state to obtain a synchronization delay;

[0126] When the synchronization delay is within the preset delay range, it is determined that the link switching is successful, and a backup transmission path is obtained.

[0127] Specifically, first evaluate the state of the current primary link according to the transmission delay and packet loss rate. The transmission delay and packet loss rate are key indicators for measuring the link performance. By monitoring these indicators in real time, it can be judged whether the link is working properly. For example, the preset delay threshold is 50 milliseconds, and the packet loss rate threshold is 1%. If the delay of the primary link is 60 milliseconds and the packet loss rate is 1.5% monitored in real time, the current link state exceeds the preset threshold, indicating that the link performance has deteriorated or there is a fault. When the current link state exceeds the preset state threshold, it is determined that the primary link has a fault. For example, if the delay of the primary link continuously exceeds 50 milliseconds and the packet loss rate exceeds 1%, the link switching mechanism will be triggered. According to the pre-configured backup link information, an available backup link is selected. The backup link is selected according to the preset strategy. The selection of the backup link is based on priority and load balancing to switch the data traffic from the primary link to the backup link. After the switching is completed, the system continuously monitors the performance indicators of the backup link to ensure that the switched link state is good. For example, the system will monitor the transmission delay, packet loss rate, and bandwidth utilization rate of the backup link. If the performance indicators of the backup link are within the preset range, the system confirms that the switching is successful. The purpose of the link switching mechanism is to switch the data traffic from the faulty primary link to the backup link to ensure the continuity of data transmission. For example, the link switching can be completed within 50 milliseconds after the fault is detected to minimize the impact of the switching process on the service.

[0128] According to the link switching mechanism, the data traffic is switched from the primary link to the backup link. After the switching is completed, the primary transmission path is determined, which is a temporary data transmission path selected based on the link status. For example, the connectivity status of the backup link is obtained through a distributed routing protocol (such as OSPF), and a path with the lowest latency and sufficient bandwidth is selected as the primary transmission path. At this time, data transmission starts through the backup link to ensure the continuity of data transmission. After switching to the backup link, the primary transmission path is monitored in real time to obtain the data synchronization status. The data synchronization status reflects the transmission consistency and latency of data between the primary and backup links. For example, by tracking the timestamps and sequence numbers of data packets, the transmission latency and packet loss rate of data packets on the backup link are monitored. If the transmission latency of a data packet is detected to be 30 milliseconds and the packet loss rate is 0.5%, it indicates that the data synchronization status is good. Based on the monitoring results of the data synchronization status, the synchronization delay is calculated. The synchronization delay refers to the delay of data transmission after the data is switched from the primary link to the backup link. For example, if the latency of the primary link is 60 milliseconds and the latency of the backup link is 30 milliseconds, the synchronization delay is 30 milliseconds. By calculating the synchronization delay, the impact of link switching on data transmission is evaluated, and it is ensured that the data transmission after switching meets the service requirements. When the synchronization delay is within the preset delay range, the link switching is determined to be successful. For example, the preset synchronization delay threshold is 100 milliseconds. If the calculated synchronization delay is 30 milliseconds, which is less than the preset threshold, the link switching is considered successful. At this time, the backup link is confirmed as the new transmission path and continues to carry the data traffic. The determination of the backup transmission path marks the completion of the link switching process, and the performance of the backup link will continue to be monitored to ensure the efficiency and stability of data transmission.

[0129] In step S17, according to the backup transmission path, a distributed routing protocol is used for calculation to obtain the first data transmission path.

[0130] In one implementation, the step of using a distributed routing protocol to calculate according to the backup transmission path to obtain the first data transmission path includes:

[0131] Using a distributed routing protocol to calculate according to the backup link to obtain the connectivity status data and topological structure information of each network gateway node in the link;

[0132] Calculating according to the connectivity status data to obtain the network link status;

[0133] Judging according to the network link status and the topological structure information to obtain the logical relationship between the primary link and the backup link;

[0134] Connecting according to the logical relationship to obtain the primary transmission path;

[0135] Perform real-time monitoring according to the primary transmission path to obtain connectivity data;

[0136] When the connectivity data is less than a preset connectivity threshold, perform the logical relationship adjustment to obtain a first data transmission path.

[0137] In a data center network, first analyze the backup links using distributed routing protocols (such as OSPF, BGP, etc.) to obtain the connectivity status data and topological structure information of each gateway node in the links. These data reflect the connection relationships between nodes in the network, the weights of the links (such as latency, bandwidth), and the existence of redundant paths. For example, through the OSPF protocol, each gateway node broadcasts its connectivity information with adjacent nodes, including the latency, bandwidth, and availability of the links. After other nodes receive this information, they update the local topological database to construct a complete network topology map. Based on the obtained connectivity status data, calculate the status of the network links. The link status includes the availability, latency, packet loss rate, and bandwidth utilization rate of the links.

[0138] Based on the network link status and topological structure information, determine the logical relationship between the primary link and the backup link. The logical relationship includes the priority of the links, redundant design, and the selection of data transmission paths. For example, it is determined that the primary link is unavailable due to a fault, and the backup link serves as an alternative path, and its topological structure and redundant design can meet the current data transmission requirements. At the same time, the performance of the backup link will also be evaluated to see if it is better than other redundant paths to ensure the efficiency and reliability of data transmission. According to the logical relationship between the primary link and the backup link, construct a primary transmission path. The primary transmission path is the optimal path selected based on the current network topology and link status. For example, select the backup link as the primary transmission path, and its path selection takes into account the latency, bandwidth, and redundant design of the link to ensure the efficiency and reliability of data transmission. The establishment of the primary transmission path is dynamic and will be adjusted according to the real-time network status.

[0139] Perform real-time monitoring on the primary transmission path to obtain connectivity data. The connectivity data includes indicators such as the real-time latency, packet loss rate, and bandwidth utilization rate of the link. For example, real-time monitor the connectivity of the backup link and find that its latency is 35 milliseconds, the packet loss rate is 0.3%, and the bandwidth utilization rate is 45%. These data reflect the actual operating status of the primary transmission path and provide a basis for subsequent path optimization. When the connectivity data is less than a preset connectivity threshold (such as the latency exceeds 50 milliseconds or the packet loss rate exceeds 1%), adjust the logical relationship between the primary link and the backup link to optimize the data transmission path. For example, if the latency of the backup link suddenly increases to 60 milliseconds, exceeding the preset threshold, re-evaluate the network topology, find other available redundant paths, and switch part of the traffic to the new path. Finally, obtain an optimized data transmission path to ensure the efficiency and stability of data transmission.

[0140] In step S18, real-time monitoring is performed according to the data transmission path to obtain monitoring data.

[0141] To ensure the stability and performance of the data transmission path, it is necessary to perform real-time monitoring on the established data transmission path. The real-time monitoring mechanism is implemented by deploying monitoring modules on key nodes of the network (such as gateways, switches). These monitoring modules can capture the transmission information of data packets in real time, including timestamps, sequence numbers, transmission delays, packet loss rates, etc. For example, the monitoring module deployed on each gateway node can record the sending and receiving times of data packets passing through the node in real time, as well as the integrity and order of the data packets. During the real-time monitoring process, key performance indicators of the data transmission path are focused on. The main performance indicators include: transmission delay; packet loss rate; bandwidth utilization; data packet integrity; The performance indicator data collected by the monitoring module will be transmitted to the analysis module in real time for preliminary analysis. The analysis module will classify and organize these data and extract key information. For example, it will calculate the average transmission delay over a period of time, the maximum packet loss rate, the peak value of bandwidth utilization, etc. These preliminary analysis results will provide a basis for subsequent in-depth analysis and decision-making. For example, if it is found that the average transmission delay of a certain link continues to increase, or the packet loss rate exceeds a preset threshold (such as 1%), then this link will be marked as a potential faulty link and the cause will be further analyzed. Based on the preliminary analysis results, a real-time monitoring report will be generated. These reports include the current state of the link, the change trend of performance indicators, and existing problems. For example, the monitoring report will show that the average transmission delay of a certain link has increased from 30 milliseconds to 50 milliseconds, and the packet loss rate has risen from 0.5% to 1.5% in the past 10 minutes. This information is crucial for network administrators and can help them quickly locate problems and take measures. At the same time, these reports will be stored in the historical database to provide a reference for subsequent trend analysis and optimization strategies.

[0142] In step S19, prediction is performed using a supervised learning algorithm according to the monitoring data to obtain an adjustment result.

[0143] First, collect the monitoring data from various gateway nodes in the network. These data include key performance indicators such as transmission delay, packet loss rate, bandwidth utilization, CPU utilization, and memory occupancy rate. These data reflect the current operating state of the network. Before making predictions, the data will be preprocessed, including data cleaning (removing outliers and noise), normalization (converting the data to a unified dimensional range), and feature extraction (extracting features useful for prediction). For example, features such as the average transmission delay in the past 10 minutes, the change trend of the packet loss rate, and the peak value of CPU utilization will be extracted to provide high-quality input data for the subsequent prediction model. Select a suitable supervised learning algorithm to build the prediction model. Supervised learning algorithms are a type of machine learning that makes predictions by learning the mapping relationship between input data and output labels. Common supervised learning algorithms include linear regression, decision tree, support vector machine (SVM), random forest, and neural network, etc. For example, for predicting the future performance indicators of a network link (such as transmission delay and packet loss rate), the random forest algorithm will be selected because it can handle complex non-linear relationships and has a certain robustness to noisy data; while for predicting the load change trend of the gateway, the long short-term memory network (LSTM) will be selected because it is good at processing time series data and can capture long-term dependencies in the data.

[0144] Use the preprocessed monitoring data to train the selected supervised learning algorithm. The training process includes dividing the data into a training set and a test set, using the training set to train the model, and evaluating the performance of the model through the test set. During the training process, the parameters of the model (such as the number of trees in the random forest, the learning rate in the neural network, etc.) will be adjusted to optimize the prediction accuracy of the model. For example, cross-validation will be used to select the optimal model parameters to ensure that the model can maintain good prediction performance on different data subsets. At the same time, various indicators during the model training process, such as mean squared error (MSE), mean absolute error (MAE), and prediction accuracy, will be recorded to evaluate the training effect of the model. The trained prediction model is used to predict the future state of the network. For example, the model can predict whether the transmission delay of a certain link will exceed the preset threshold within the next 5 minutes, or predict whether the CPU utilization of the gateway will exceed 80% within the next 10 minutes. According to the prediction results, corresponding adjustment results will be generated. If it is predicted that the transmission delay of the link will exceed the threshold, some traffic will be switched to other links in advance; if it is predicted that the CPU utilization of the gateway will be too high, the task allocation weights will be adjusted and some tasks will be migrated to other gateways. These adjustment results are aimed at optimizing network performance and avoiding potential performance bottlenecks or failure risks.

[0145] Implement corresponding optimization measures according to the adjustment results obtained from the prediction. For example, adjust the traffic allocation, update the task scheduling strategy, or reconfigure the network topology. After the implementation of the adjustment, continuously monitor the operating status of the network and evaluate the effectiveness of the adjustment measures. If the performance of the adjusted network meets the expectations, continue to operate; if the adjustment effect is not good, re-analyze the monitoring data, adjust the parameters of the prediction model or select other algorithms to further optimize the network performance. Through this continuous process of monitoring, prediction, and adjustment, it is possible to dynamically adapt to changes in the network state and ensure the efficient and stable operation of the data center network.

[0146] In step S110, optimize according to the adjustment result to obtain the optimal data transmission path for data transmission.

[0147] In one implementation, the optimizing according to the adjustment result to obtain the optimal data transmission path for data transmission includes:

[0148] Obtain the redundant link status information;

[0149] Quantify according to the adjustment result to obtain the initial weight matrix;

[0150] Perform dynamic adjustment according to the initial weight matrix to obtain the weight matrix coefficient;

[0151] Calculate according to the weight matrix coefficient to obtain the queue cache weight;

[0152] Calculate according to the redundant link status information to obtain the data synchronization period parameter;

[0153] Iteratively optimize the initial weight matrix according to the queue cache weight and the data synchronization period parameter to obtain the optimal data transmission path for data transmission.

[0154] In one implementation, the iteratively optimizing the initial weight matrix according to the queue cache weight and the data synchronization period parameter to obtain the optimal data transmission path for data transmission includes:

[0155] Calculate according to the queue cache weight and the data synchronization period parameter to obtain the quality of service index;

[0156] Monitor according to the quality of service index to obtain the quality of service metric;

[0157] When the quality of service metric is less than the metric threshold, iteratively optimize the initial weight matrix to obtain the optimal data transmission path for data transmission.

[0158] Specifically, first, obtain the status information of redundant links. This information includes key performance indicators such as link connectivity, latency, packet loss rate, and bandwidth utilization. These data reflect the current operating status of the redundant links and provide a basis for subsequent optimization calculations. For example, through a distributed routing protocol (such as OSPF), the status of redundant links is monitored in real time. It is found that the latency of a certain backup link is 30 milliseconds, the packet loss rate is 0.5%, and the bandwidth utilization is 40%. Quantify the adjustment results to generate an initial weight matrix. The initial weight matrix is used to represent the connection relationship between gateway nodes and the task assignment weights. For example, according to the adjustment results, an initial weight value is assigned to each gateway node, and these weight values reflect the importance and priority of the nodes in data transmission.

[0159] Perform dynamic adjustment based on the initial weight matrix and calculate the weight matrix coefficient. The weight matrix coefficient is used to dynamically adjust the task assignment between gateways to adapt to changes in the network state. For example, by monitoring the load status of gateways and link performance in real time, the weight values are dynamically adjusted. If the load of gateway A is too high, its weight value will be reduced, and some tasks will be assigned to gateway B or gateway C with lower loads. Calculate the queue cache weight according to the weight matrix coefficient. The queue cache weight is used to optimize the cache management strategy of gateway queues to ensure that high-priority data streams can be processed first. For example, according to the weight matrix coefficient, cache weights are assigned to each gateway queue, and high-priority data streams obtain higher cache weights. For example, for gateway A, its queue cache weight is 0.8, while for gateway C, its queue cache weight is 0.4.

[0160] Calculate the data synchronization cycle parameter according to the status information of the redundant links. The data synchronization cycle parameter determines the synchronization frequency of data between the primary link and the backup link. For example, if the latency of the backup link is low and the bandwidth is sufficient, the data synchronization cycle can be shortened to improve data consistency and real-time performance. For example, the calculated data synchronization cycle parameter is 100 milliseconds. Combine the queue cache weight and the data synchronization cycle parameter to calculate the Quality of Service Index (QoS Index). The Quality of Service Index is a comprehensive indicator used to evaluate the quality of service of the current network configuration. For example, the Quality of Service Index calculated according to the queue cache weight and the data synchronization cycle parameter is 0.85. Monitor the Quality of Service Index in real time to generate the Quality of Service Metric (QoS Metric). The Quality of Service Metric reflects the current quality of service status of the network, including latency, packet loss rate, bandwidth utilization, etc. For example, the monitored Quality of Service Metric is latency 30 milliseconds, packet loss rate 0.5%, and bandwidth utilization 40%.

[0161] If the quality of service metrics are lower than the preset thresholds (e.g., latency exceeds 50 milliseconds or packet loss rate exceeds 1%), the initial weight matrix will be iteratively optimized. For example, if the monitored latency is 60 milliseconds, which exceeds the preset threshold of 50 milliseconds, the weight matrix will be readjusted to optimize the data transmission path. Through multiple iterative optimizations, the optimal data transmission path is finally determined to ensure the efficiency and stability of data transmission. After the above iterative optimization process, the optimal data transmission path is finally determined and data transmission begins. The data stream is allocated to the corresponding gateways and links according to the optimized path to ensure that the data can be transmitted efficiently and stably. For example, a path with the lowest latency, the smallest packet loss rate, and sufficient bandwidth will be selected as the optimal transmission path. During the transmission process, the performance metrics of the link will be continuously monitored to ensure the continuity and reliability of data transmission.

[0162] To facilitate the understanding of the present invention, some preferred embodiments of the present invention will be further described below.

[0163] The working process of the present invention will be described below by taking a relatively common scenario as an example. Please also refer to Figure 2 , which is Figure 1 a schematic diagram of the working scenario of the method.

[0164] In a certain data center, in the face of gateway load imbalance and link failure risks, efficient and stable data transmission is achieved through a set of dynamic optimizations. First, by collecting resource data such as the CPU utilization rate, memory occupancy rate, and I / O throughput of the gateways, combined with the priority attributes of the data stream, the processing capabilities of the gateways are quantified and the task allocation is optimized to reduce the pressure on high-load gateways. Subsequently, based on the monitoring of the cache length and buffer length, the dynamic load balancing mechanism is triggered to migrate some data streams from high-load gateways to low-load gateways to further relieve the load pressure. At the same time, the transmission latency and packet loss rate of the link are monitored in real time. When the primary link status is abnormal, it is quickly switched to the backup link to ensure the continuity of data transmission. In addition, the network topology information is obtained through the distributed routing protocol to dynamically adjust the data transmission path and optimize the link usage strategy. On this basis, predictive analysis is carried out using real-time monitoring data to adjust the traffic allocation and gateway weights in advance to avoid potential performance bottlenecks. Finally, combined with the redundant link status and prediction results, the weight matrix and cache management strategy are iteratively optimized, and the data synchronization period is dynamically adjusted to ensure the efficient transmission of critical services. Through this series of dynamic adjustments and optimization measures, the data center has successfully solved the problems of load imbalance and link failure, significantly improved the network performance and stability, the latency of critical services has been reduced to 20 milliseconds, the packet loss rate has been maintained below 0.1%, and the network resource utilization rate has been greatly improved.

[0165] In summary, the present invention performs a quantitative evaluation based on the gateway resource data and the data stream priority attribute to obtain the cache length of the data. By constructing a gateway resource utilization matrix, the real-time task processing ability score value of the gateway is calculated, and the weighted round-robin algorithm is used to schedule gateway tasks, quantifying the evaluation of the cache length and dynamic task allocation, optimizing the resource utilization rate of the gateway, reducing the pressure on the high-load gateway, and improving the overall data transmission efficiency.

[0166] The present invention makes a judgment based on the time stamp and the serial number to obtain a link switching mechanism; through real-time monitoring and the dynamic link switching mechanism, it is ensured that when a network failure or a decline in link quality occurs, the backup link can be quickly switched to, reducing data transmission delay and improving real-time performance.

[0167] The present invention calculates the data transmission path according to the backup link by using a distributed routing protocol; through the distributed routing protocol and real-time monitoring data, the data transmission path is dynamically adjusted, avoiding resource waste and bottleneck problems, and improving the utilization efficiency of network resources.

[0168] The present invention performs iterative optimization according to the real-time adjustment result for data transmission; through iterative optimization and a prediction mechanism, it can cope with sudden traffic and network anomalies in advance, reduce performance fluctuations, and improve the stability and risk resistance of key services.

[0169] Refer to Figure 3 , the second embodiment of the present invention provides a data center remote operation system based on the analysis of a collection gateway, including:

[0170] A data acquisition module for acquiring gateway resource data and data stream priority attributes;

[0171] A data evaluation module for performing a quantitative evaluation based on the gateway resource data and the data stream priority attribute to obtain the cache length of the gateway queue;

[0172] A data processing module for, when the cache length is greater than a preset length threshold, performing processing by using a dynamic load balancing algorithm to obtain a migrated data stream;

[0173] A data analysis module for analyzing according to the migrated data stream to obtain the time stamp and serial number of the data stream;

[0174] A data measurement module for measuring according to the time stamp and the serial number to obtain the transmission delay and packet loss rate of the primary and backup links;

[0175] A link switching module for evaluating according to the transmission delay and the packet loss rate, and when the evaluation result is that the primary link fails, performing dynamic link switching to obtain a backup transmission path;

[0176] A data calculation module, configured to perform calculations using a distributed routing protocol according to the backup transmission path to obtain a first data transmission path;

[0177] A data monitoring module, configured to monitor according to the first data transmission path to obtain monitoring data;

[0178] A data prediction module, configured to perform prediction using a supervised learning algorithm according to the monitoring data to obtain an adjustment result;

[0179] A data optimization module, configured to perform optimization according to the adjustment result to obtain an optimal data transmission path for data transmission.

[0180] It should be noted that a data center remote operation system based on acquisition gateway analysis provided in an embodiment of the present invention is used to execute all process steps of a data center remote operation method based on acquisition gateway analysis in the above embodiment. The working principles and beneficial effects of the two correspond one by one, and thus will not be elaborated herein.

[0181] An embodiment of the present invention further provides an electronic device. The electronic device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as an algorithm program. When the processor executes the computer program, the steps in the above embodiments of various data center remote operation methods based on acquisition gateway analysis are implemented, such as Figure 1 the step S11 shown. Alternatively, when the processor executes the computer program, the functions of each module / unit in the above device embodiments are implemented, such as a data center remote operation module based on acquisition gateway analysis.

[0182] Exemplarily, the computer program can be divided into one or more modules / units. The one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program in the electronic device.

[0183] The electronic device can be a computing device such as a desktop computer, a notebook, a palm computer, and a smart tablet. The electronic device may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that the above components are only examples of the electronic device, and do not constitute a limitation on the electronic device. It may include more or fewer components than the above, or combine certain components, or different components. For example, the electronic device may further include input / output devices, network access devices, a bus, etc.

[0184] The so-called processor may be a Central Processing Unit (CPU), or may also be 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 may be a microprocessor or the processor may also be any conventional processor, etc. The processor is the control center of the electronic device and connects all parts of the entire electronic device through various interfaces and lines.

[0185] The memory can be used to store the computer program and / or modules. The processor realizes various functions of the electronic device by running or executing the computer program and / or modules stored in the memory, and by calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area. Among them, the program storage area can store applications required for operations and at least one function (such as a sound playback function, an image playback function, etc.); the data storage area can store data created according to the use of the mobile phone (such as audio data, phone book, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a FlashCard, at least one magnetic disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0186] Among them, if the modules / units integrated in the electronic device are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, to implement all or part of the processes in the above-described embodiment methods of the present invention, it can also be completed by a computer program instructing relevant hardware. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by a processor, the steps of the above various method embodiments can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in the form of source code, object code, executable file, or some intermediate form, etc. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disc, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the computer-readable medium does not include electrical carrier signals and telecommunication signals.

[0187] It should be noted that the device embodiments described above are only illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. In addition, in the attached drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which can be specifically implemented as one or more communication buses or signal lines. Those of ordinary skill in the art can understand and implement it without creative efforts.

[0188] The above-described specific embodiments have further elaborated on the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above is only the specific embodiments of the present invention and is not used to limit the protection scope of the present invention. It is particularly pointed out that for those skilled in the art, any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A remote operation method for a data center based on the analysis of a collection gateway, characterized in that Including: Obtain gateway resource data and data stream priority attributes; Perform quantitative evaluation based on the gateway resource data and the data stream priority attributes to obtain the cache length of the gateway queue; When the cache length is greater than a preset length threshold, use a dynamic load balancing algorithm for processing to obtain a migrated data stream; Analyze according to the migrated data stream to obtain the timestamp and sequence number of the data stream; Measure according to the timestamp and the sequence number to obtain the transmission delay and packet loss rate of the primary and backup links; Evaluate according to the transmission delay and the packet loss rate. When the evaluation result is that the primary link fails, perform link dynamic switching to obtain a backup transmission path; Calculate according to the backup transmission path using a distributed routing protocol to obtain a first data transmission path; Monitor according to the first data transmission path to obtain monitoring data; Use a supervised learning algorithm to predict according to the monitoring data to obtain an adjustment result; Optimize according to the adjustment result to obtain an optimal data transmission path for data transmission.

2. The method for remotely operating a data center based on the analysis of a collection gateway according to claim 1, wherein The performing quantitative evaluation based on the gateway resource data and the data stream priority attributes to obtain the cache length of the gateway queue includes: Construct according to the gateway resource data to obtain a gateway resource utilization matrix; Calculate using the analytic hierarchy process according to the gateway resource utilization matrix to obtain a matrix weight coefficient; Calculate according to the matrix weight coefficient to obtain a task processing score value of the gateway; Schedule using a weighted round-robin algorithm according to the task processing score value to obtain an optimal task allocation scheme; Perform real-time monitoring according to the optimal task allocation scheme to obtain the cache length of the gateway queue.

3. The method for remote operation of a data center based on analysis by a collection gateway according to claim 1, wherein, The when the cache length is greater than a preset length threshold, using a dynamic load balancing algorithm for processing to obtain a migrated data stream includes: Obtain the load value and CPU value of the current gateway; When the cache length is greater than a preset length threshold, extract according to the cache length to obtain a priority attribute; Calculate using a weighted round-robin strategy according to the priority attribute to obtain a round-robin value; Calculate according to the load value and the CPU value to obtain a high-load gateway and a low-load gateway; Migrate data from the high-load gateway to the low-load gateway according to the round-robin value to obtain a migrated data stream.

4. The method for remotely operating a data center based on analysis by a collection gateway according to claim 1, wherein The evaluating according to the transmission delay and the packet loss rate. When the evaluation result is that the primary link fails, performing link dynamic switching to obtain a backup transmission path includes: Judge according to the transmission delay and the packet loss rate to obtain the current link state; When the current link state is greater than a preset state threshold, determine that the primary link fails to obtain a link switching mechanism; Perform link switching according to the link switching mechanism to obtain a primary transmission path; Monitor according to the primary transmission path to obtain a data synchronization state; Calculate according to the data synchronization state to obtain a synchronization delay; When the synchronization delay is within a preset delay range, determine that the link switching is successful to obtain a backup transmission path.

5. The method for remotely operating a data center based on the analysis of a collection gateway according to claim 1, wherein The calculating according to the backup transmission path using a distributed routing protocol to obtain a first data transmission path includes: Calculate according to the backup link using a distributed routing protocol to obtain the connectivity status data and topological structure information of each gateway node in the link; Calculate according to the connectivity status data to obtain the network link status; Judge according to the network link status and the topological structure information to obtain the logical relationship between the primary link and the backup link; Connect according to the logical relationship to obtain a primary transmission path; Perform real-time monitoring according to the primary transmission path to obtain connectivity data; When the connectivity data is less than a preset connectivity threshold, perform the logical relationship adjustment to obtain a first data transmission path.

6. The method for remotely operating a data center based on analysis by a collection gateway according to claim 1, wherein, The optimizing according to the adjustment result to obtain an optimal data transmission path for data transmission includes: Obtain redundant link status information; Quantify according to the adjustment result to obtain an initial weight matrix; Perform dynamic adjustment according to the initial weight matrix to obtain a weight matrix coefficient; Calculate according to the weight matrix coefficient to obtain a queue buffer weight; Calculate according to the redundant link status information to obtain a data synchronization period parameter; Iteratively optimize the initial weight matrix according to the queue buffer weight and the data synchronization period parameter to obtain an optimal data transmission path for data transmission.

7. The method for remotely operating a data center based on the analysis of a collection gateway according to claim 6, wherein, The iteratively optimizing the initial weight matrix according to the queue buffer weight and the data synchronization period parameter to obtain an optimal data transmission path for data transmission includes: Calculate according to the queue buffer weight and the data synchronization period parameter to obtain a quality of service index; Monitor according to the quality of service index to obtain a quality of service metric; When the quality of service metric is less than a metric threshold, iteratively optimize the initial weight matrix to obtain an optimal data transmission path for data transmission.

8. A remote operation system for a data center based on the analysis of a collection gateway, characterized in that, Includes: A data acquisition module for acquiring gateway resource data and data stream priority attributes; A data evaluation module for performing quantitative evaluation according to the gateway resource data and the data stream priority attributes to obtain the cache length of the gateway queue; A data processing module for, when the cache length is greater than a preset length threshold, performing processing using a dynamic load balancing algorithm to obtain a migrated data stream; A data analysis module for analyzing according to the migrated data stream to obtain the timestamp and sequence number of the data stream; A data measurement module for measuring according to the timestamp and the sequence number to obtain the transmission delay and packet loss rate of the primary and backup links; A link switching module for evaluating according to the transmission delay and the packet loss rate, and when the evaluation result is that the primary link fails, performing link dynamic switching to obtain a backup transmission path; A data calculation module for calculating according to the backup transmission path using a distributed routing protocol to obtain a first data transmission path; A data monitoring module for monitoring according to the first data transmission path to obtain monitoring data; A data prediction module for predicting according to the monitoring data using a supervised learning algorithm to obtain an adjustment result; A data optimization module for optimizing according to the adjustment result to obtain an optimal data transmission path for data transmission.

9. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the method for remotely operating a data center based on the analysis of a collection gateway as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored computer program. When the computer program runs, it controls the device where the computer-readable storage medium is located to execute the method for remotely operating a data center based on the analysis of a collection gateway as described in any one of claims 1 to 7.

Images