Intelligent paging system

Abstract

The invention relates to the technical field of communication network switching and control, and discloses an intelligent paging system, which comprises a state monitoring unit used for measuring and calculating physical transmission time consumption from a core node to a candidate edge access node and determining a logic link time delay value; the strategy arbitration unit is used for calculating the congestion trend gradient of the candidate edge access node and executing arrival situation deduction logic which takes the congestion trend gradient and the logic link time delay value as gain coefficients to carry out product operation so as to generate an arrival prediction load value; and the orthogonal shear selection module is used for locking the candidate edge access node as a direct connection target or dividing the candidate edge access node into a blocking subset according to an arrival predicted load value. Physical time delay is converted into a buffer window for resource release by constructing a product coupling mechanism of physical transmission time delay and a logic congestion gradient; dynamic congestion is avoided; and the network throughput efficiency is improved.

Description

Technical Field

[0001] This invention relates to an intelligent paging system, belonging to the field of communication network switching and control technology. Background Technology

[0002] In the current communication switching network architecture, the core control node establishes connection paths to distributed edge access nodes through a switching matrix, distributing paging signaling and control commands. Existing technologies mostly adopt a load balancing strategy based on state snapshots. The core node periodically collects the command queue depth of the edge nodes through the communication link, and selects the target node based on the principle of minimum load or closest distance according to the static load at the sampling time. The decision logic based on static sampling snapshots is pre-synchronized with the physical execution, ignoring the timing lag caused by the transmission time of the physical link. There is an unavoidable transmission time difference from the time the signaling is sent from the core node to the time it reaches the edge node port. With the increase of switching network throughput and the increase of service burstiness, the queue state of the edge node undergoes high-frequency dynamic evolution, resulting in a phase deviation between the load information obtained by the core node at the time of decision and the actual load state at the time the signaling actually arrives. Increasing the sampling frequency shortens the sensing interval, but it increases the in-band signaling overhead, squeezes the service bandwidth, and cannot eliminate the state slippage during transmission.

[0003] Existing control logic is limited to spatial resource scheduling and ignores dynamic evolution in the temporal dimension. For example, the Chinese invention patent with authorization announcement number CN101998512B discloses a method for load balancing between mobile switching center pools, a mobile switching center and system. The scheme monitors the load rate of the MSC pool. When the load exceeds a predetermined first threshold, it establishes a virtual pool with the lightly loaded MSC pool to achieve cross-regional load sharing. Although such technology introduces an inter-pool cooperation mechanism, it is essentially still a passive response mode based on threshold triggering. The core node's decision-making relies on the load rate information, which is actually the historical statistical data reported by the edge nodes in the preceding period. In the face of a sudden large-scale signaling storm, from the core node initiating a status query and establishing a virtual mapping relationship to the actual transmission of signaling to the target node, there is an unavoidable physical transmission and logical processing delay. The scheme does not take into account the drastic gradient changes in node load during this transit time, resulting in a serious time misalignment between the core node's decision-making basis, i.e., the past idle state, and the actual scenario when the signaling arrives, i.e., the current potential congestion.

[0004] Therefore, the technical problem to be solved by this invention is how to coordinate physical transmission delay and load evolution trend in routing logic, eliminate the time difference between decision and execution, and achieve accurate matching between signaling arrival time and node availability status. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: An intelligent paging system, comprising: The status monitoring unit is used to periodically collect the current instruction queue depth value of multiple candidate edge access nodes via the communication link. Based on the timestamps of heartbeat messages, the physical transmission time from the core node port to each candidate edge access node is analyzed, and the physical transmission time is established as the logical link latency value. ; The policy arbitration unit and connection status monitoring unit are used to perform logical selection for signaling routes, based on the instruction queue depth value of the current sampling period. Instruction queue depth value of the preceding sampling period Calculate the congestion trend gradient of each candidate edge access node and execute the congestion trend gradient With logical link time delay value The arrival status deduction logic, which multiplies the gain coefficients of each other, generates the arrival prediction load value, representing the node load status at the actual arrival time of the signaling. The arrival situation deduction logic follows the following operation rules: ,in, The preset confidence level weighting coefficient; The orthogonal shear selection module is used to receive the arriving predicted load value. It will reach the predicted load value The value is compared with the preset selection threshold, and the predicted load value is reached. When the load falls below the selection threshold, the corresponding candidate edge access node is locked as a pass-through target and included in the pass-through selection subset. This process continues until the predicted load value is reached. When the threshold is higher or equal to the selection threshold, the corresponding candidate edge access node is marked as a congestion risk node and included in the blocking selection subset. The paging execution unit is used to parse the physical port address of each pass-through target in the pass-through selection subset, drive the switching matrix to establish a physical signaling transmission path and distribute paging instructions only for the pass-through targets.

[0006] Preferably, the strategy arbitration unit is used to identify the congestion trend gradient when executing the arrival situation inference logic. Negative and logical link time delay value Under specific operating conditions where the delay exceeds a preset threshold, the strategy arbitration unit utilizes the logical link delay value. As a congestion trend gradient The magnification factor is calculated to be lower than the current instruction queue depth. The predicted load value This allows for the priority selection of remote candidate edge access nodes in the resource release phase within the orthogonal shear selection module. The physical transmission time is used as a buffer window for reducing the instruction queue of the candidate edge access nodes, thereby achieving time-domain alignment between the signaling arrival time and the node idle time.

[0007] Preferably, the orthogonal shearing selection module includes a logical mask generation logic, used to generate a logical mask for candidate edge access nodes included in the blocking selection subset, and to forcibly block the operation of establishing physical signaling transmission paths to candidate edge access nodes in the blocking selection subset during the current paging cycle, until the status monitoring unit detects the arrival prediction load value of the candidate edge access node in a subsequent sampling cycle. It fell back below the selection threshold.

[0008] Preferably, the strategy arbitration unit further includes a timing selection control module, which is used to receive uplink feedback signals from candidate edge access nodes and establish a reverse mapping relationship between downlink distribution logic and uplink feedback signals; the timing selection control module is used to trigger a degradation and flow suppression mechanism when the intensity of the uplink feedback signal exceeds the storm warning threshold, control the generation rate of the signaling distribution token bucket and forcibly insert a silent time gap between consecutive signaling distribution actions, so as to adjust the timing rhythm of signaling through the switching node.

[0009] Preferably, the strategy arbitration unit further includes a range selection module for obtaining the historical topology traversal frequency of the target communication endpoint. It also calculates the mobility inertia factor and uses this factor to determine the time threshold for paging range. Perform inverse weighting correction to generate the corrected time threshold. And based on the corrected time threshold A step-by-step switching selection is performed between single-node routing range, adjacency ring routing range, and global routing range; the range selection module calculates the corrected time threshold. The logic conforms to the following operation rules: ,in, These are the preset inertia weighting coefficients.

[0010] Preferably, the status monitoring unit establishes the logical link time delay value. The method is as follows: During system operation, probe frames are periodically sent to each candidate edge access node, or the timestamp difference of the heartbeat messages reported by each candidate edge access node is parsed. The one-way physical transmission time from the core node port to each candidate edge access node is calculated, and this one-way physical transmission time is directly mapped to the logical link latency value. And store it in the local time delay mapping table.

[0011] Preferably, the strategy arbitration unit includes gradient smoothing operation logic, which is used to collect instruction queue depth values ​​for multiple consecutive sampling periods based on a sliding time window mechanism, and to perform fitting calculation on the instruction queue depth values ​​for multiple consecutive sampling periods using the least squares method to obtain the congestion trend gradient. .

[0012] Preferably, the paging execution unit includes multi-level switching drive logic, which is used to extract the routing identifier of the pass-through target after receiving the pass-through selection subset output by the orthogonal shear selection module, and send circuit establishment instructions to the underlying optoelectronic switching matrix to build a dedicated physical signaling transmission path between the core node port and the pass-through target.

[0013] Preferably, the timing selection control module sets the duration of the silent interval to be positively correlated with the strength of the uplink feedback signal. The timing selection control module is used to set the duration of the silent interval to be greater than or equal to the full signaling processing cycle when the uplink feedback signal indicates that the candidate edge access node is in a fully loaded state.

[0014] Preferably, the intelligent paging system is deployed in the mobile communication switching center or core network control plane network element; the status monitoring unit, policy arbitration unit and paging execution unit are connected through a high-speed internal bus, and each unit is integrated in the same physical rack or distributed in multiple physical cabinets interconnected by optical fiber.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In the intelligent paging system, the product of physical link transmission time and queue change gradient is introduced to construct a node selection mechanism based on arrival situation inference. This changes the static load judgment mode that relies on sampling time and transforms the inherent transmission delay of the physical link into a queue resource release buffer window. For remote nodes in the resource release phase, the transmission path time is used as a resource clearing reserve time, so that the signaling arrival time and the node idle time are matched in the time domain. This eliminates the time difference between the control plane decision time and the forwarding plane execution time, which leads to misselection or omission, and improves the switching system's throughput and connection success rate for dynamic traffic.

[0016] 2. An orthogonal shearing selection strategy is adopted. Based on the predicted load value, the candidate node set is dynamically divided into a pass-through subset and a blocking subset. Nodes in the blocking subset are shielded by logical masks to establish active circuit breaker protection. When a node is detected to have a traffic surge trend, even if the current load is low or an overload occurs at the predicted arrival time, the connection path establishment operation is forcibly blocked. A logical damper is built in the core switching layer to prevent multiple signaling signals from rushing to nodes with instantaneous low load and causing secondary congestion oscillations, thus ensuring the load stability of the core switching matrix and edge access nodes.

[0017] 3. Through a dual adjustment mechanism of timing selection and range selection, the signaling distribution rhythm is dynamically adapted to the target physical characteristics. A reverse mapping is established between the downlink distribution logic and the uplink feedback signal. When the feedback exceeds the limit, the token bucket generation rate is controlled to force insertion into the silent time interval. At the physical level, the timing rhythm of the signaling access switching node is adjusted to avoid the backlog of processing units caused by instantaneous pulse traffic. The mobility inertia factor is calculated using the terminal's historical topology traversal frequency. The time threshold is corrected in reverse to switch the paging range step by step. Based on the inertia range scaling strategy, the routing selection range automatically matches the target terminal's movement rate and trajectory characteristics, ensuring the paging hit rate while reducing the invalid signaling overhead and port resource occupation caused by blind broadcasting across the entire area. Attached Figure Description

[0018] Figure 1 This is a flowchart illustrating the logic control architecture and signal processing of the intelligent paging system of the present invention. Figure 2 This is a trend curve showing the impact of the confidence weighting coefficient of this invention on the key performance indicators of the system. Figure 3 This is a schematic diagram of the system deployment topology and core-edge closed-loop control interaction of the present invention. Detailed Implementation

[0019] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the following embodiments are only used to illustrate the present invention and are not intended to limit the scope of protection of the present invention.

[0020] An intelligent paging system includes a status monitoring unit, a strategy arbitration unit, an orthogonal shearing selection module, and a paging execution unit. These units are interconnected via a high-speed internal data bus or optical backplane, and are integrated into the same physical rack or distributed across multiple physical cabinets interconnected by optical fibers. The status monitoring unit periodically collects the current instruction queue depth values ​​of multiple candidate edge access nodes via communication links. And establish the logical link latency values ​​from the core node port to each candidate edge access node. Establish logical link time delay value The method is as follows: During system operation, probe frames are periodically sent to each candidate edge access node, or the timestamp difference of the heartbeat messages reported by each candidate edge access node is parsed. This timestamp difference represents the one-way physical transmission time required from the core node issuing the instruction to the edge node receiving the instruction. The system directly maps this one-way physical transmission time to a logical link latency value. It is stored in a local time-delay mapping table; the policy arbitration unit connects to the state monitoring unit and performs logical selection for signaling routes, based on the instruction queue depth value of the current sampling period. Instruction queue depth value of the preceding sampling period Calculate the congestion trend gradient for each candidate edge access node. To eliminate sampling jitter, the policy arbitration unit employs gradient smoothing logic, collecting instruction queue depth values ​​for multiple consecutive sampling periods based on a sliding time window mechanism. It then uses the least squares method to fit and calculate the congestion trend gradient from these values. .

[0021] The strategy arbitration unit executes the arrival situation inference logic to generate the arrival prediction load value, which represents the node load status at the actual arrival time of the signaling. The deductive logic follows the operational rules. ,in, This represents the current instruction queue depth. The preset confidence level weighting coefficient; For congestion trend gradient; This is the logical link time delay value; this operation logic utilizes the logical link time delay value. As a time integral variable, the congestion trend gradient Mapped to the expected increase or decrease in load during transmission; the policy arbitration unit identifies the congestion trend gradient. Negative and logical link time delay value In specific operating conditions where the delay exceeds a preset threshold, the strategy arbitration unit utilizes the logical link delay value. As a congestion trend gradient The magnification factor is calculated to be lower than the current instruction queue depth. The predicted load value This mechanism prioritizes remote candidate edge access nodes that are in the resource release phase, using physical transmission time as a buffer window for reducing the command queue of candidate edge access nodes, thus achieving time-domain alignment between signal arrival time and node idle time; the orthogonal shearing selection module receives the arrival predicted load value. The value is then compared with a preset selection threshold, and the predicted load value is reached. When the load falls below the selection threshold, the module locks the corresponding candidate edge access node as a pass-through target and adds it to the pass-through selection subset; when the predicted load value is reached... When the threshold is higher than or equal to the selection threshold, the module marks the corresponding candidate edge access node as a congestion risk node and includes it in the blocking selection subset. For nodes in the blocking selection subset, the orthogonal shear selection module generates a logical mask and forcibly blocks the operation of establishing a physical signaling transmission path within the current paging cycle, until the status monitoring unit detects the arrival prediction load value of the node in a subsequent sampling cycle. It fell back below the selection threshold.

[0022] The paging execution unit parses the physical port addresses of each pass-through target in the pass-through selection subset, drives the switching matrix to establish physical signaling transmission paths only for the pass-through targets, and distributes paging commands. This unit extracts the routing identifier of the pass-through target, sends circuit establishment commands to the underlying optoelectronic switching matrix, and constructs a dedicated physical signaling transmission path between the core node port and the pass-through target. The policy arbitration unit also includes a timing selection control module, which receives uplink feedback signals from candidate edge access nodes and establishes a reverse mapping relationship between downlink distribution logic and uplink feedback signals. When uplink is detected... When the strength of the feedback signal exceeds the storm warning threshold, the timing selection control module triggers a degradation and flow suppression mechanism, controlling the generation rate of the signaling distribution token bucket and forcibly inserting a silent time gap between consecutive signaling distribution actions. The duration of this silent time gap is positively correlated with the strength of the uplink feedback signal. When the uplink feedback signal indicates that the candidate edge access node is under full load, the timing selection control module sets the duration of the silent time gap to be greater than or equal to the complete signaling processing cycle. The policy arbitration unit also includes a range selection module, used to obtain the historical topology traversal frequency of the target communication endpoint. It also calculates the mobility inertia factor and uses this factor to determine the time threshold for paging range. Perform reverse weighting correction to generate the corrected time threshold. Corrected time threshold The calculation conforms to the operation rules. ,in, Based on the time threshold; These are the preset inertia weighting coefficients; Based on the historical topology traversal frequency, the range selection module uses a modified time threshold. Perform a step-by-step switching selection between single-node routing range, adjacency ring routing range, and global routing range.

[0023] Example 1: In application scenarios where this intelligent paging system is deployed in high-density pedestrian areas such as large stadiums or transportation hubs, the core switching network needs to distribute paging signaling to tens of thousands of edge access nodes in a short period of time to cope with the instantaneous traffic surges caused by large-scale events or emergencies. The load status of edge nodes is not static, but fluctuates dramatically with the real-time movement of people and service triggers within the venue. Traditional path selection mechanisms based on load snapshots cannot detect the physical time required for signaling transmission, causing core nodes to often send signaling to nodes that appear idle at the sampling time but are congested due to micro-traffic bursts when the signaling actually arrives, thus causing signaling packet loss, retransmission storms, and even congestion spread throughout the entire switching plane. This example system solves the above challenges through the collaborative operation of a state monitoring unit and a policy arbitration unit. The state monitoring unit periodically collects the instruction queue depth value of each edge node. The logical link latency value is established based on the physical transmission time calculated from the probe frames. Based on this, the strategy arbitration unit does not directly use the current load to make a decision, but instead executes the arrival situation inference logic to calculate the arrival predicted load value. When the load on an edge node in a certain area of ​​the venue increases sharply due to the gathering of people, this is known as the congestion trend gradient. When the value is positive, the deduction logic utilizes the time delay value of the logic link. As a gradient amplifier, it predicts in advance that the node will be in an overloaded state when the signal arrives, generating... It will be higher than the current observed value. The orthogonal shear selection module is based on this Add this node to the blocking selection subset.

[0024] Conversely, for those peripheral nodes around the venue that are in the resource release phase as the crowd disperses, i.e., the congestion trend gradient... For nodes with negative values, the system utilizes the logical link time delay value. The product effect of the negative gradient is used to calculate the predicted load value that is lower than the current load. This enables the orthogonal shear selection module to identify and select nodes that currently appear busy but are expected to become idle again when the signaling arrives after the transmission delay as pass-through targets. The paging execution unit then drives the switching matrix to establish physical paths to these remote nodes with load-bearing potential and distribute signaling.

[0025] Example 2: A network simulation test platform consisting of one core control node and 20 edge access nodes was constructed. The platform is based on the NS-3 network simulator and adopts a star topology. The physical link bandwidth between the core node and the edge nodes is set to 10Gbps, and the link propagation delay is set to a random distribution value ranging from 1.2ms to 48.5ms to simulate the physical environment of a metropolitan area network or wide area network. The test input data uses publicly available mobile communication network signaling traffic datasets, and burst traffic pulses conforming to a Poisson distribution are superimposed on them. To simulate the anti-interference capability under real industrial electromagnetic environment, Gaussian white noise with a signal-to-noise ratio of 20dB is injected into the signal transmission process, and random link jitter with an amplitude of ±18% of the average delay is introduced.

[0026] The experiment included a control group and the sample group of this invention, and confidence weighting coefficients were applied to key parameters. A control group for out-of-range conditions is set up. The control group uses the queue depth value reported only by the edge nodes of the previous sampling period. The static minimum load routing strategy for decision-making, the arrival situation deduction logic of the policy arbitration unit enabled in this invention, is based on the formula... Calculate the predicted load value and perform orthogonal shear selection, where the confidence weighting factor is... The value is set to 0.8, the sampling period is set to 10ms, and the out-of-range control group 1 will... Set to 0.2, the out-of-range control group 2 will The system was set to version 2.0 to verify the boundary effects of parameter values. The test process was divided into three stages: benchmark test, high load stress test lasting 300 seconds, and recovery test. During the stress test, the system monitored and recorded key indicators such as signaling packet loss rate, average queuing delay, and frequency of node congestion.

[0027] Table 1: Measurement Data of Key Performance Indicators During the Stress Testing Phase

[0028] Referring to Table 1, the control group data showed a high packet loss rate and queuing delay, indicating that decisions based on static snapshots cannot adapt to high dynamic traffic, leading to signaling backlog at nodes whose state has deteriorated during transmission. The present invention's sample group introduces a logical link delay value... Congestion trend gradient The product term reduced the packet loss rate to 1.9%, shortened the average queuing latency to 18.9ms, and controlled the congested node ratio to 10.2%. Compared with the out-of-range control group data, when When the value is 0.2, all indicators are close to the control group, indicating insufficient utilization of gradient information and large prediction bias; when When the value is 2.0, although the packet loss rate is better than the control group, the average queuing latency is 6.8ms higher than that of the present invention sample, indicating that the excessive sensitivity leads to the excessive shielding of available path resources, causing unnecessary queuing. In addition, during the recovery test phase of traffic decline, the average queuing latency of the present invention sample decreased at a rate 56% faster than that of the control group.

[0029] Example 3: This example combines Figures 1 to 3 A description of an intelligent paging system, such as... Figure 1As shown, the communication link signal is input to the status monitoring unit, which performs the operations of collecting the command queue depth and calculating the logical link time delay value. Then, the physical transmission time and queue depth data are transmitted to the policy arbitration unit. The policy arbitration unit is connected to the timing selection control module through a reverse mapping relationship, receives the monitoring uplink feedback signal and forcibly inserts the silent time interval. On the other hand, it is responsible for calculating the congestion trend gradient and executing the arrival situation inference logic, thereby generating the arrival prediction load value and sending it to the orthogonal shear selection module. This module divides the nodes into different flow directions based on the threshold value comparison results. For nodes whose arrival prediction load value is higher than or equal to the threshold and there is a risk of congestion, the system classifies them into the blocking selection subset and generates a logical mask or performs a blocking operation. For the direct target whose arrival prediction load value is lower than the selection threshold, the system directs it to the paging execution unit, drives the switching matrix to establish a physical signaling transmission path and finally completes the paging command distribution.

[0030] like Figure 2 As shown, the horizontal axis represents the confidence weighting coefficient. The values ​​range from 0.2 to 2. The vertical axis represents the performance index value, ranging from 0 to 40. The chart contains three curves: the solid line represents the average signaling packet loss rate (%), the dashed line represents the average queuing delay (ms), and the dotted line represents the percentage of congested nodes (%). All three curves show a trend of change with frequency. The value exhibits a nonlinear change characteristic of first decreasing and then increasing, or trending towards flatness; such as... Figure 3 As shown, the core control and decision-making center is deployed within the mobile switching center. It includes a status monitoring unit for sensing transmission time, a strategy arbitration unit for performing situational inference calculations, and an orthogonal shearing module for executing pass-through or blocking decisions. Remote nodes are responsible for transmitting heartbeat and status signals back to the core. After receiving the uplink feedback signal, the status monitoring unit drives subsequent logic. Finally, the paging execution unit and the optoelectronic switching matrix execute the distribution decision. For edge nodes in the resource release phase, the system establishes a physical signaling path for pass-through. For edge nodes with high congestion risk, the system implements logical masking to block them, thereby constructing a closed-loop control network with the core as the central hub, connecting remote and edge nodes.

[0031] Example 4: Targeting the time delay value of logical links in the status monitoring unit The accuracy of the calculation is easily affected by network jitter and the confidence weighting coefficient in the policy arbitration unit. To address the two potential shortcomings of insufficient disclosure regarding the lack of an adaptive adjustment mechanism, this embodiment provides a complete set of parameter calibration and adaptive optimization procedures, specifically for logical link time delay values. The system executes a statistical filtering process based on a sliding time window, and the state monitoring unit does not directly use the round-trip time (RTT) of a single probe frame as the basis for its operation. Instead, it maintains a length of like A sliding window that stores the most recent Physical transmission time sample of the second probe To remove outliers caused by sudden network jitter, the algorithm calculates the standard deviation of the samples within the window. Then remove all values ​​that deviate from the average by more than [a certain amount]. (For example For the sample points, the system calculates a weighted average of the remaining valid samples. The weights decrease exponentially with the freshness of the sample time. This weighted average is then established as the logical link time delay value at the current moment. This procedure ensures It can reflect the long-term trend of link latency and effectively filter out instantaneous jitter noise.

[0032] Secondly, regarding the confidence weighting coefficient in the strategy arbitration unit The system introduces an online self-learning mechanism based on prediction error feedback, which aims to solve the problem of fixed... To address the issue of the system's inability to adapt to different network traffic patterns, at the end of each paging cycle, the system calculates the actual arrival load of all target nodes in the previous cycle. Compared with the predicted arriving load If the mean absolute percentage error (MAPE) between the two values ​​consistently exceeds a preset error threshold, such as 15%, then a trigger will be activated. Correction logic: When This indicates that the system underestimates the congestion trend and needs to be adjusted. To increase the gradient gain; conversely, when When this occurs, it indicates that the system's prediction is overly aggressive and needs to be adjusted. Correct step size The magnitude of MAPE is positively correlated with the system's performance. Through this closed-loop feedback adjustment, the system can dynamically find and maintain the optimal state under the current network environment. This allows for continuous optimization of the accuracy of situational awareness projections without human intervention. Furthermore, to further eliminate the ambiguity of the orthogonal shearing selection threshold setting in the strategy arbitration unit, this embodiment clarifies the dynamic anchoring logic of this threshold. The selection threshold is not a fixed static value, but rather a baseline relative to the real-time processing capabilities of the edge access nodes. The system uses periodic stress tests or historical peak data to calibrate the maximum instruction processing rate of each edge node without packet loss, and defines it as... The threshold for orthogonal shear selection is set to ,in As a safety margin factor, for example, a value of 0.85 to 0.95, this logic ensures that the blocking operation is triggered only when the predicted load approaches the node's physical limit, thus preventing overload circuit breakers and maximizing the effective throughput of the node.

[0033] Example 5: To ensure the congestion trend gradient in the strategy arbitration unit To accurately reflect the real load pressure of nodes with different hardware specifications, this embodiment constructs an offline calibration procedure for the queue depth-processing capacity mapping. In a controlled laboratory environment, for each type of edge access node, the instruction injection rate is gradually increased using a traffic generator until packet loss occurs. During this process, the instruction queue depth at different input rates is recorded. Actual processing delay Nonlinear relationship curve between Based on this curve, the system constructs a lookup table to store the queue depth values ​​collected in real time. Mapped to an equivalent normalized load index, the policy arbitration unit calculates the congestion trend gradient. Instead of directly using the original queue depth, the normalized load index is used. This mechanism eliminates gradient calculation deviations caused by differences in queue capacity of heterogeneous hardware, ensuring the uniformity and accuracy of network-wide congestion situation awareness. In response to the differences in physical link latency characteristics under different network environments, this embodiment provides a standardized on-site deployment pre-calibration procedure.

[0034] Before the system is officially put into operation, a time-delay baseline calibration process must be performed. This process sends bursty high-frequency probe frame sequences to all edge access nodes in the entire network for a duration of no less than 24 hours to cover the complete diurnal traffic fluctuation cycle. The system collects the round-trip time (RTT) data of all probe frames and calculates the average RTT of each node. and standard deviation If a certain node If the link is determined to have a high jitter risk, the system will automatically adjust the logical link delay value of the node. Locked to To introduce additional safety buffer time; for low-jitter nodes, set This procedure ensures that the logical delay parameters can adapt to the physical characteristics of a specific deployment environment, avoiding performance degradation caused by overly aggressive or conservative parameter settings.

[0035] Example 6: To solve the problem of orthogonal shearing selection threshold To address the issue of overly generalized definitions, this embodiment introduces an online procedure for threshold self-calibration. Instead of using a fixed threshold, the system leverages the real-time processing capabilities of edge nodes. Configure dynamically. Defined as the average rate at which a node successfully processes and returns a response within the most recent sliding time window, such as 1 minute, the orthogonal shear selection threshold is set to... ,in For example, with a safety margin factor of 0.9, this procedure introduces real-time processing capability feedback to enable the blocking logic to sense and adapt to performance fluctuations of edge nodes caused by hardware aging, background task preemption, or changes in ambient temperature.

[0036] For the predicted load value reached in the strategy arbitration unit The confidence weighting coefficients used in the calculation To address the engineering black box problem of the lack of standardized calibration methods, this embodiment provides a complete set of pre-system calibration procedures that must be executed during the initial stage of system deployment. The offline calibration process for parameter optimization utilizes historical traffic data playback or a traffic generator to inject simulated traffic patterns with different burst characteristics into the system, covering everything from stable (burst factor) to... From extreme suddenness () The system iterates through the complete range of traffic patterns for each traffic pattern. The value space is set, for example, from 0.1 to 2.0, with a step size of 0.1. The corresponding signaling packet loss rate is recorded and calculated. and average queuing delay By constructing the objective function ,in For the weighting coefficients, determine the... Minimize the optimal The system ultimately generates traffic characteristics based on these values. During the online operation phase, the status monitoring unit calculates the burst factor of the current traffic in real time using the mapping table. And look up the table to get the matching value.

[0037] Example 7: Before calculating the one-way physical transmission time, the status monitoring unit executes a clock synchronization procedure. It locks the system clock phase of the core node and candidate edge access nodes using the IEEE 1588 Precision Time Protocol or Synchronous Ethernet Protocol, controlling the network-wide time synchronization deviation within the microsecond range and eliminating logical link delays caused by clock drift between nodes. To compensate for calculation errors, for edge nodes that do not support hardware timestamps, the status monitoring unit automatically switches the round-trip delay measurement mode, continuously sends multiple sets of probe messages to record the round-trip time, and takes half of the arithmetic average as the equivalent replacement value for one-way physical transmission time, thus maintaining the consistency of the physical meaning of the time delay parameter in heterogeneous network environments.

[0038] The orthogonal shearing selection module determines the selection threshold based on a sliding time window statistical distribution algorithm. It retrieves historical instruction queue depth data from candidate edge access nodes over the past 24 hours, constructs a queue length probability density function, and sets the queue depth value corresponding to a function integral area reaching 90% of the total area as the baseline selection threshold. This threshold is dynamically adjusted based on the actual throughput during the current sampling period. When a node's throughput falls below 80% of its rated processing capacity, the selection threshold is reduced linearly to tighten the blocking criteria under low-load conditions and prevent misjudgments caused by occasional queue fluctuations. The strategy arbitration unit weights the confidence coefficients. Implement closed-loop feedback regulation based on posterior error, and compare the load value at the end of each control cycle to the predicted load value. The root mean square error (RMSE) between the actual load value and the predicted load value is calculated. If the RMS error exceeds the preset tolerance range for three consecutive periods and the actual load value is higher than the predicted value, the error is incremented in steps of 0.1. Value-enhanced congestion trend gradient The gain of the prediction result increases, while the gain decreases by the same step size. The value is maintained until the root mean square error falls back to the tolerance range.

[0039] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0040] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. An intelligent paging system, characterized in that, include: The status monitoring unit is used to periodically collect the current instruction queue depth value of multiple candidate edge access nodes via the communication link. Based on the timestamps of heartbeat messages, the physical transmission time from the core node port to each candidate edge access node is analyzed, and the physical transmission time is established as the logical link latency value. ; The policy arbitration unit and connection status monitoring unit are used to perform logical selection for signaling routes, based on the instruction queue depth value of the current sampling period. Instruction queue depth value of the preceding sampling period Calculate the congestion trend gradient of each candidate edge access node and execute the congestion trend gradient With logical link time delay value The arrival status deduction logic, which multiplies the gain coefficients of each other, generates the arrival prediction load value, representing the node load status at the actual arrival time of the signaling. The arrival situation deduction logic follows the following operation rules: ,in, The preset confidence level weighting coefficient; The orthogonal shear selection module is used to receive the arriving predicted load value. It will reach the predicted load value The value is compared with the preset selection threshold, and the predicted load value is reached. When the load falls below the selection threshold, the corresponding candidate edge access node is locked as a pass-through target and included in the pass-through selection subset. This process continues until the predicted load value is reached. When the threshold is higher or equal to the selection threshold, the corresponding candidate edge access node is marked as a congestion risk node and included in the blocking selection subset. The paging execution unit is used to parse the physical port address of each pass-through target in the pass-through selection subset, drive the switching matrix to establish a physical signaling transmission path and distribute paging instructions only for the pass-through targets.

2. The intelligent paging system according to claim 1, characterized in that, The strategy arbitration unit is used to identify congestion trend gradients when executing arrival situation deduction logic. Negative and logical link time delay value Under specific operating conditions where the delay exceeds a preset threshold, the strategy arbitration unit utilizes the logical link delay value. As a congestion trend gradient The magnification factor is calculated to be lower than the current instruction queue depth. The predicted load value This allows for the priority selection of remote candidate edge access nodes that are in the resource release phase within the orthogonal shear selection module.

3. The intelligent paging system according to claim 1, characterized in that, The orthogonal shearing selection module includes a logical mask generation logic, used to generate logical masks for candidate edge access nodes included in the blocking selection subset. During the current paging cycle, it forcibly blocks the operation of establishing physical signaling transmission paths to candidate edge access nodes within the blocking selection subset, until the status monitoring unit detects the arrival prediction load value of the candidate edge access node in a subsequent sampling cycle. It fell back below the selection threshold.

4. The intelligent paging system according to claim 1, characterized in that, The strategy arbitration unit also includes a timing selection control module, which is used to receive uplink feedback signals from candidate edge access nodes and establish a reverse mapping relationship between downlink distribution logic and uplink feedback signals. The timing selection control module is used to trigger a degradation and flow suppression mechanism when the intensity of the uplink feedback signal exceeds the storm warning threshold. It controls the generation rate of the signaling distribution token bucket and forcibly inserts a silent time gap between consecutive signaling distribution actions to adjust the timing rhythm of signaling through the switching node.

5. The intelligent paging system according to claim 1, characterized in that, The strategy arbitration unit also includes a range selection module for obtaining the historical topology traversal frequency of the target communication endpoint. It also calculates the mobility inertia factor and uses this factor to determine the time threshold for paging range. Perform inverse weighting correction to generate the corrected time threshold. And based on the corrected time threshold Perform a step-by-step switching selection between single-node routing range, adjacency ring routing range, and global routing range; The range selection module calculates the corrected time threshold. The logic conforms to the following operation rules: ,in, These are the preset inertia weighting coefficients.

6. The intelligent paging system according to claim 1, characterized in that, The status monitoring unit establishes the logical link time delay value. The method is as follows: During system operation, probe frames are periodically sent to each candidate edge access node, or the timestamp difference of the heartbeat messages reported by each candidate edge access node is parsed. The one-way physical transmission time from the core node port to each candidate edge access node is calculated, and this one-way physical transmission time is directly mapped to the logical link latency value. And store it in the local time delay mapping table.

7. The intelligent paging system according to claim 1, characterized in that, The strategy arbitration unit includes gradient smoothing logic, which collects instruction queue depth values ​​for multiple consecutive sampling periods based on a sliding time window mechanism, and uses the least squares method to fit and calculate the instruction queue depth values ​​for multiple consecutive sampling periods to obtain the congestion trend gradient. .

8. The intelligent paging system according to claim 1, characterized in that, The paging execution unit includes multi-level switching drive logic, which is used to extract the routing identifier of the pass-through target after receiving the pass-through selection subset output by the orthogonal shear selection module, and send circuit establishment instructions to the underlying optoelectronic switching matrix to build a dedicated physical signaling transmission path between the core node port and the pass-through target.

9. The intelligent paging system according to claim 4, characterized in that, The timing selection control module sets the duration of the silent interval to be positively correlated with the strength of the uplink feedback signal. When the uplink feedback signal indicates that the candidate edge access node is in a fully loaded state, the timing selection control module sets the duration of the silent interval to be greater than or equal to the complete signaling processing cycle.

10. The intelligent paging system according to claim 1, characterized in that, The intelligent paging system is deployed in the mobile communication switching center or core network control plane network element; the status monitoring unit, policy arbitration unit and paging execution unit are connected through a high-speed internal bus, and each unit is integrated in the same physical rack or distributed in multiple physical cabinets interconnected by optical fiber.

Images