Short message issuing congestion intelligent processing method and device

Abstract

The invention discloses a short message issuing congestion intelligent processing method and device, and relates to the technical field of communication network flow scheduling and congestion control. An emergency scene is intelligently recognized through a dynamic threshold model, and a processing mode is switched; dynamically setting a short message priority according to the mode and the calling number; a resource allocation method of a fixed dedicated channel and an adjustable public channel is adopted, and a flow rule is predicted by utilizing machine learning so as to dynamically optimize public resource scheduling; differential time window management based on multi-dimensional evaluation and junk short message linkage punishment are carried out on the short messages; designing an intelligent cache library with priority perception capability to perform congestion buffering; in the emergency mode, the retransmission strategy of the short message center platform is actively intervened to cooperate with flow control, fine identification, self-adaptive scheduling and intelligent congestion control of the short message flow are realized, and the guarantee capability of high-priority services, the utilization efficiency of system resources and the overall service reliability are remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of communication network traffic scheduling and congestion control technology, specifically a method and apparatus for intelligent congestion handling during SMS delivery. Background Technology

[0002] With the rapid development of mobile communication technology and the widespread adoption of smart terminals, SMS, as a fundamental communication service, has seen its application scenarios continuously expand and its volume of business significantly increase. Currently, SMS services mainly include point-to-point personal SMS, service-oriented SMS sent by industry organizations (such as bank transaction notifications and identity verification codes), and commercial SP SMS sent by service providers (SPs) (such as news briefings and marketing promotions). Furthermore, emergency notification SMS (such as typhoon warnings and epidemic updates), as a special type of industry SMS involving public safety, places extremely high demands on the timeliness and reliability of delivery.

[0003] However, during statutory holidays, e-commerce promotional periods, major social events, or daily peak business hours, the demand for sending various types of SMS messages surges, easily creating instantaneous traffic spikes on the operator's network side. This leads to a significant increase in processing pressure on the SMS center platform and related network elements, resulting in systemic congestion. Such congestion not only causes delays and increased loss rates in sending ordinary SMS messages, affecting user experience, but may also cause substantial business interruptions and security risks for industries such as finance, transportation, and government that rely on SMS for critical business outreach and security authentication.

[0004] Currently, common solutions are mostly based on static queue scheduling or simple priority adjustment mechanisms. These methods often lack sufficient flexibility and intelligence when facing massive SMS flows that are dynamically changing, highly sudden, and have mixed business types. They are unable to optimize the overall system throughput and efficiently utilize network resources while ensuring the quality of high-priority business services such as emergency notifications and important services. Summary of the Invention

[0005] The purpose of this invention is to provide a method and apparatus for intelligent handling of SMS delivery congestion, which solves the problem that existing technologies lack intelligent and adaptive scheduling capabilities during SMS traffic peaks, cannot guarantee the absolute smooth operation of high-priority services such as emergency notifications, and at the same time optimizes the overall network resource efficiency.

[0006] The objective of this invention can be achieved through the following technical solutions:

[0007] This application provides a method for intelligent congestion handling during SMS delivery, including the following steps:

[0008] S1. Receive pre-delivered SMS messages from the SMS center platform, automatically identify emergency notification SMS messages among them, and determine whether the number of emergency notification SMS messages per unit time exceeds the preset threshold; if it does not exceed the threshold, enter the normal processing mode; if it exceeds the threshold, enter the emergency processing mode.

[0009] S2. Dynamically set the priority of SMS messages based on the entered mode and the caller ID.

[0010] In the normal processing mode, point-to-point SMS is set to high priority, industry SMS to medium priority, and SP SMS to low priority; in the emergency processing mode, emergency notification SMS is set to high priority, point-to-point SMS and other industry SMS to medium priority, and SP SMS to low priority.

[0011] S3. Allocate dedicated channel resources in proportion to SMS messages of different priorities, and set up public channel resources to be used in order of priority; at the same time, collect historical data of SMS messages and use machine learning models to dynamically adjust the usage strategy of the public channel resources.

[0012] S4. For low-priority SP SMS messages, allocate differentiated delivery time windows based on the analysis results of effectiveness, importance and sensitivity, identify SP SMS messages as spam messages, and reduce public channel resources based on the identification results;

[0013] S5. When congestion is detected due to insufficient channel resources, the congested SMS messages are temporarily stored in the SMS cache and released according to the first-in-first-out rule after the congestion is resolved; when the cache storage space is full, the earliest stored low-priority SMS messages are automatically deleted.

[0014] S6. In emergency handling mode, periodically obtain the real-time occupancy rate data of the resend library of the SMS center platform, and when the alarm threshold is exceeded, send an instruction to the SMS center platform to modify the resend strategy so that SMS messages that fail to be sent will no longer be resent.

[0015] This application provides an apparatus for intelligent congestion handling during SMS delivery, applied to a method for intelligent congestion handling during SMS delivery, comprising:

[0016] The mode determination module is used to identify emergency notification SMS messages and switch working modes according to their traffic thresholds;

[0017] The priority setting module is used to set the priority of SMS messages based on the caller ID and the current mode;

[0018] The channel resource control module is used to allocate dedicated channel resources for different priorities and manage public channel resources, and integrates machine learning models to dynamically adjust resource allocation strategies.

[0019] The SP management module is used to configure the delivery time window for SP SMS messages and to identify and penalize spam SMS messages.

[0020] The cache management module is used to store SMS messages during congestion and manage them according to first-in-first-out and priority-aware rules.

[0021] The collaborative intervention module is used to monitor and intervene in the SMS center platform's resend strategy in emergency mode.

[0022] The beneficial effects of this invention are as follows:

[0023] This invention introduces an intelligent scene perception and mode switching mechanism based on a dynamic threshold model, combines caller ID rule matching with rapid content keyword recognition technology based on Trie trees, and implements dynamic priority setting and differentiated channel resource allocation strategies linked to the mode. This effectively solves the key problem that traditional methods cannot accurately identify and absolutely guarantee high-priority services such as emergency notifications under sudden traffic surges. It achieves accurate identification of emergency SMS messages, intelligent switching of system modes, and ensures the priority, reliability, and timely delivery of critical public safety information under congestion conditions.

[0024] By constructing a multi-layered resource pool of dedicated and public channels, and using machine learning models to analyze historical traffic patterns to dynamically predict and adjust the allocation weight of public channel resources, while implementing multi-dimensional evaluation and differentiated time window control based on benefits, importance, and sensitivity for low-priority SP SMS messages, the problem of rigid resource allocation and inability to adapt to dynamic changes in business traffic, resulting in low overall resource utilization efficiency, has been solved in the existing technology. This achieves refined and adaptive scheduling of network channel resources, significantly improving the overall throughput and resource utilization efficiency of the system while ensuring the service quality of high-priority services.

[0025] By designing an intelligent SMS caching library with priority awareness, it stores SMS messages in an orderly manner according to priority during congestion, releases them in an orderly manner according to rules after the congestion is relieved, and executes a priority-aware replacement strategy in extreme cases. At the same time, in emergency mode, it actively intervenes in the SMS center platform's resending strategy through cross-system collaborative interfaces, effectively solving the problems of easy loss of business data, degraded user experience, and potential further deterioration of congestion during congestion. It not only builds an effective congestion buffering and smooth recovery mechanism, reducing SMS loss and delay, and enhancing the system's robustness in the face of sudden traffic, but also directly generates economic benefits such as increased operating revenue and reduced expansion costs by ensuring the delivery of high-value services and reducing unnecessary redundant traffic. Attached Figure Description

[0026] To better understand and implement this application, the technical solution is described in detail below with reference to the accompanying drawings.

[0027] Figure 1 This is a schematic diagram of the normal mode SMS congestion intelligent processing method provided in Embodiment 1 of this application;

[0028] Figure 2 This is a schematic diagram of the emergency mode SMS congestion intelligent processing in a method for SMS congestion intelligent processing provided in Embodiment 1 of this application;

[0029] Figure 3 A schematic diagram of SMS service types and sending channels in a method for intelligent congestion handling during SMS delivery provided in Embodiment 1 of this application;

[0030] Figure 4 This is a schematic diagram of a device for intelligent congestion handling during SMS delivery, provided in Embodiment 2 of this application. Detailed Implementation

[0031] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, exemplary embodiments will be described in detail below, examples of which are illustrated in the accompanying drawings. In the following description, when referring to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application.

[0032] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used herein are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0033] The following detailed description of the specific implementation methods, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided in detail.

[0034] Example 1

[0035] Please see Figures 1-3 This embodiment provides a method for intelligent congestion handling during SMS delivery, including the following steps:

[0036] S1. Receive pre-delivered SMS messages from the SMS center platform, automatically identify emergency notification SMS messages among them, and determine whether the number of emergency notification SMS messages per unit time exceeds the preset threshold; if it does not exceed the threshold, enter the normal processing mode; if it exceeds the threshold, enter the emergency processing mode.

[0037] The emergency notification SMS messages include typhoon notification SMS messages and epidemic notification SMS messages; automatic identification is based on the caller ID characteristics or content keywords of the SMS messages; the threshold is the upper limit of the number of emergency notification SMS messages per unit time, and exceeding this value determines that the emergency handling mode has been entered.

[0038] The preset threshold is managed by a dynamic threshold generation model. This model uses historical emergency SMS traffic data from the same period, the real-time load rate of the current network SMS service, and the public opinion heat index of social events obtained from public information sources as multi-dimensional input features. It analyzes and predicts through a trained machine learning model (such as gradient boosting decision tree), dynamically calculates and outputs the most suitable mode switching trigger threshold. For example, during network off-peak hours or periods of social calm, the model maintains a higher baseline threshold. However, when the system detects an abnormal surge in network traffic in a specific area and public opinion indicates that a sudden event has occurred in that area, even if the absolute number of emergency SMS messages has not yet reached the static threshold, the model will intelligently lower the threshold, thereby triggering the emergency handling mode in advance. This adaptive mechanism enables the system to proactively respond to potential congestion risks and achieve intelligent mode switching from passive response to proactive prediction.

[0039] Furthermore, the automatic identification of emergency notification SMS messages specifically includes:

[0040] The system acquires the caller ID and text content of the SMS message to be identified, and compares the caller ID with a preset emergency number feature database. It then performs keyword matching analysis on the SMS text content using a high-efficiency multi-pattern string matching algorithm based on a Trie tree. Based on the combined results of the number comparison and content keyword matching analysis, if the caller ID is successfully matched or a preset keyword is matched in the content, the SMS message is determined to be an emergency notification SMS, and a corresponding judgment conclusion is generated.

[0041] In this implementation, a high-efficiency multi-pattern string matching algorithm based on Trie trees is used to perform keyword matching analysis on SMS text. A Trie tree is constructed in advance, containing a keyword library of core emergency terms such as typhoon, epidemic, evacuation, and early warning. When the SMS text to be identified arrives, the algorithm uses it as input and performs a single scan traversal starting from the root node of the Trie tree, thus completing the parallel matching detection of all keywords simultaneously. This method effectively avoids the efficiency bottleneck of traditional word-by-word traversal matching and is especially suitable for scenarios with high concurrency processing of massive amounts of SMS messages. It ensures that SMS messages containing key emergency information are accurately identified within milliseconds, providing an efficient and reliable basis for subsequent comprehensive judgment.

[0042] Preset keywords refer to a set of specific words or phrases predefined by the system to identify the characteristics of emergency events (such as typhoons, epidemics, early warnings, evacuations, etc.).

[0043] Specifically, by constructing a dynamic threshold generation model based on multi-dimensional features (historical traffic, real-time load, and public opinion), an intelligent mode switching from passive response to proactive prediction was achieved. Furthermore, a dual-channel identification mechanism combining caller ID feature database comparison with efficient keyword matching based on Trie trees was adopted. This solved the problems of traditional methods failing to accurately and quickly identify emergency notification SMS messages under sudden traffic surges, and failing to proactively adjust system processing strategies. Thus, it achieved millisecond-level accurate identification of emergency information and intelligent pre-emptive entry into emergency preparedness in scenarios with massive SMS concurrency, ultimately ensuring that critical public safety notifications such as those related to typhoons and epidemics can still be delivered preferentially, reliably, and promptly even during congestion.

[0044] S2. Dynamically set the priority of SMS messages based on the entered mode and the caller ID of the SMS message;

[0045] In normal mode, point-to-point SMS is set to high priority, industry SMS to medium priority, and SP SMS to low priority. In emergency mode, emergency notification SMS is set to high priority, point-to-point SMS and other industry SMS to medium priority, and SP SMS to low priority.

[0046] Furthermore, the calling number for point-to-point SMS messages is a mobile phone number; the calling number for industry SMS messages is a service number for banks or government service platforms; the calling number for SP SMS messages is a service provider number; and the calling number for emergency notification SMS messages is a specific number for government emergency platforms.

[0047] Specifically, by establishing intelligent priority mapping rules based on the caller's number that are linked to the system's operating mode (normal / emergency), the key problem of static and fixed priority settings in traditional SMS dispatching, which cannot dynamically adjust the order of business processing according to the actual network scenario (such as whether it is in an emergency), is solved. In normal mode, priority is given to personal point-to-point communication, while in emergency mode, the absolute priority of government emergency notifications is unconditionally guaranteed. This achieves intelligent differentiation and dynamic adaptation of the value of SMS services in different scenarios, and achieves the technical effect of always prioritizing the most urgent communication needs (especially emergency public safety information) in complex network environments, while guiding the rational and orderly allocation of network resources.

[0048] S3. Allocate dedicated channel resources in corresponding proportions for SMS messages of different priorities, and set up public channel resources for use in priority order. Simultaneously, collect historical delivery patterns and use a machine learning model to dynamically adjust the usage strategy of the public channel resources; this dynamic adjustment mechanism is... Figure 2 This is one of the core components in the intelligent processing flow of the normal mode shown.

[0049] In normal mode, the allocation ratio of dedicated channel resources is 30% for high priority, 20% for medium priority, 15% for low priority, and 35% for public channel resources. In emergency mode, high priority accounts for 80%, medium priority accounts for 20%, and low priority is not allocated dedicated channel resources. The machine learning model is used to analyze historical traffic patterns during statutory holidays and e-commerce promotion periods, and dynamically adjust the weight of public channel resources used by each business.

[0050] Furthermore, the use of machine learning models to dynamically adjust the usage strategy of the public channel resources includes: continuously collecting and storing historical SMS delivery data of different priorities to construct a feature dataset containing time, business type, resource allocation weight, and performance indicators; training a time series prediction and multi-objective optimization model based on the feature dataset, which can learn the business traffic patterns in typical scenarios such as statutory holidays and e-commerce promotion periods, and predict the demand for public channel resources for each priority business in future periods; during runtime, the model dynamically outputs optimized public channel resource allocation weights based on real-time system status and prediction results; and the resource scheduler adjusts the public channel scheduling strategy in real time according to the public channel resource allocation weights, thereby realizing dynamic and intelligent adjustment of the public channel resource usage strategy, which can improve overall resource utilization efficiency and system throughput while ensuring the service quality of high-priority businesses.

[0051] The process of predicting the demand for public channel resources for different priority services in the future includes: cleaning and extracting features from historical data to construct a multi-dimensional training sample set with time series as the core, including service type labels, resource allocation records, and performance indicators; training the model using a deep learning architecture that integrates long short-term memory networks and attention mechanisms, enabling the model to automatically learn from the data the periodic and sudden traffic change patterns in typical scenarios such as statutory holidays and e-commerce promotion periods, and the complex mapping relationship between them and the resource demands of different priority services; during the deployment phase, the trained model receives real-time service traffic time series data, current resource utilization, and date type as feature inputs, and continuously predicts the probability distribution of the demand intensity of different priority services for public channel resources in the future set time period, and generates a dynamic resource weight allocation scheme that maximizes the total throughput while satisfying the latency constraints of high-priority services based on a multi-objective optimization algorithm.

[0052] Specifically, by constructing a hybrid resource allocation model that combines dedicated channels with fixed quotas and intelligently scheduled public channels, and by combining a machine learning dynamic optimization mechanism based on time-series prediction and multi-objective optimization, the core problem of static resource allocation in traditional solutions, which cannot adapt to dynamic changes in business traffic and thus makes it difficult to balance the guarantee of high-priority business and the overall efficiency of the system, is solved. The technical solution allocates a fixed proportion of dedicated channels to different priority businesses to ensure their service baseline. At the same time, it establishes a public resource pool and uses deep learning models to analyze historical patterns and predict future demand, thereby dynamically adjusting the competition weight of public resources. This achieves the intelligent and adaptive maximization of the utilization efficiency of public channel resources and the overall throughput of the system while absolutely guaranteeing the service quality of high-priority businesses such as emergency notifications.

[0053] S4. For low-priority SP SMS messages, allocate differentiated delivery time windows based on the analysis results of effectiveness, importance, and sensitivity, and identify SP SMS messages as spam messages, and reduce public channel resources based on the identification results;

[0054] The delivery time window is set based on the business value of SP SMS messages and user impact. High-efficiency SPs are allocated prime time slots, while low-efficiency SPs are restricted to off-peak times. Spam SMS identification is based on feature value threshold judgment. If the threshold is exceeded, the channel resources of that SP SMS message will be halved within 24 hours.

[0055] Furthermore, differentiated delivery time windows are allocated based on the analysis results of effectiveness, importance, and sensitivity. Specifically, this includes: constructing a multi-dimensional evaluation model for each service provider (SP), and calculating a comprehensive score for effectiveness, importance, and sensitivity by quantitatively analyzing its historical SMS revenue contribution rate, whether its business involves key public services (such as bank verification and government notifications), and user complaint and unsubscription rates; then, based on preset scoring thresholds and mapping strategies, SPs are automatically classified into different service levels, and each level is bound to a preset delivery time window strategy. For example, high-effectiveness, high-importance SPs are authorized to send messages during peak user activity periods, while low-effectiveness, low-sensitivity SPs are restricted to sending messages during off-peak periods when the network is relatively idle; finally, the time window configuration is sent to the scheduling module in real time for enforcement, and this allocation strategy can be dynamically adjusted based on the periodic evaluation results of SP business performance.

[0056] The process of constructing the multidimensional evaluation model is as follows: Based on a preset quantitative indicator system, the historical revenue contribution rate and unit traffic value of the SP's SMS service are calculated from the perspective of effectiveness; the business is analyzed from the perspective of importance to determine whether it is associated with key service types such as financial transaction verification and government notifications; and user complaint rate, unsubscription rate, and content compliance score are statistically analyzed from the perspective of sensitivity. The quantitative results of the three dimensions are integrated and calculated through a preset weighted algorithm to generate a comprehensive evaluation score for each SP. Based on the score and the preset level threshold, the service level is automatically classified and categorized.

[0057] The process of calculating the comprehensive scores for effectiveness, importance, and sensitivity is as follows: historical financial data of SPs are obtained from the business support system to calculate their revenue contribution rate and revenue per SMS as the effectiveness score; whether they provide key services such as finance and government affairs are identified from the business registration information database and service type list, and the service weight is quantified as the importance score; and the complaint rate, unsubscription rate, and number of violations are statistically analyzed from the user complaint platform and content review logs and normalized as the sensitivity score.

[0058] Specifically, intelligent grading is achieved by constructing a multi-dimensional SP evaluation model based on effectiveness, importance, and sensitivity. Differentiated delivery time windows are then implemented based on the grading results. A closed-loop penalty mechanism for spam SMS identification and resource reduction is introduced to address the problems of ineffective network resource occupation by low-value or nuisance SMS, inability of high-value services to obtain high-quality channels, and user disruption during rest periods caused by the extensive management of commercial SMS (SP SMS) in traditional SMS scheduling. Through refined value assessment and behavior management of SPs, high-quality network time slots are accurately allocated to high-effectiveness and high-importance services, effectively suppressing spam SMS sending. This achieves the technical effect of significantly improving user experience and purifying the network environment while increasing the operator's network resource monetization efficiency and business revenue.

[0059] S5. When congestion is detected due to insufficient channel resources, the congested SMS messages are temporarily stored in the SMS cache. After the congestion is resolved, they are released according to the first-in-first-out rule. When the cache is full, the earliest stored low-priority SMS messages are automatically deleted.

[0060] The cache library follows a first-in-first-out (FIFO) principle when releasing SMS messages, and prioritizes the elimination of low-priority SMS messages when the storage space is full, so as to ensure that high-priority SMS messages can be cached.

[0061] Furthermore, congested SMS messages are temporarily stored in an SMS cache and released according to a first-in, first-out (FIFO) rule after the congestion is relieved. Specifically, this includes:

[0062] The system monitors the resource utilization and queue backlog status of the SMS delivery channel in real time. When the preset congestion judgment conditions are met, the system automatically starts the caching mechanism and transfers SMS messages that cannot be delivered in time to the SMS cache library.

[0063] Based on the priority identifier of the SMS messages, they are stored in the corresponding high, medium and low priority logical queues in the cache. Each queue is organized according to the order of arrival of the SMS messages, and each SMS message record contains metadata such as priority and entry timestamp.

[0064] Continuously monitor the channel status. When the channel resources are detected to have recovered to a non-congested state, release the SMS messages arranged in the first-in-first-out rule in each queue according to the order of high, medium and low priority queues and resubmit them to the sending channel.

[0065] If the cache storage space is exhausted during the congestion period, a replacement mechanism is activated, prioritizing the deletion of the earliest-stored SMS messages in the low-priority queue. If that queue is empty, the same rules are applied to the medium and high-priority queues in turn, thereby ensuring the cache retention of high-priority SMS messages.

[0066] Specifically, by using an intelligent SMS caching system with priority awareness, a dynamic and adaptive congestion buffering and recovery mechanism is constructed. When channel congestion is detected, SMS messages are distributed and stored in different logical queues according to their priority, ensuring that high-priority information has a retention advantage in the cache. After the congestion is relieved, messages are released in an orderly manner strictly according to priority and combined with the first-in-first-out rule. When the cache space is exhausted, an intelligent replacement strategy based on priority and time is initiated to prioritize the elimination of the least valuable data. This effectively solves the problem that traditional systems are prone to losing critical SMS messages and slow system recovery under instantaneous traffic surges due to a lack of buffering or simple buffering strategies. As a result, the system's resilience, business continuity assurance capabilities, and overall service reliability in the face of sudden congestion are significantly improved.

[0067] S6. In emergency handling mode, periodically obtain the real-time occupancy rate data of the resend library of the SMS center platform, and when it exceeds the alarm threshold, send an instruction to the SMS center platform to modify the resend strategy so that SMS messages that fail to be sent will no longer be resent.

[0068] The alarm threshold is set to 80%. A dedicated monitoring agent actively polls or subscribes to the status interface provided by the SMS center platform every 60 seconds to collect real-time occupancy data of its retransmission database. When the occupancy rate continuously exceeds the alarm threshold for a preset period, such as two consecutive collection periods, a policy adjustment instruction is automatically generated. This instruction is then sent to the SMS center platform via a predefined application programming interface (API) call, specifying policy parameters that include pausing retransmission or significantly reducing the number of retries. Upon receiving the instruction, the SMS center platform dynamically updates the corresponding business or global retransmission rules in its policy engine, thereby achieving cross-system collaborative traffic control and effectively preventing further network congestion caused by repeated retransmission attempts.

[0069] Specifically, in emergency response mode, a dedicated monitoring agent is deployed to periodically collect real-time occupancy data of the SMS center platform's retransmission database. When the occupancy rate continuously exceeds a preset alarm threshold, a policy adjustment command is automatically sent to the platform via a standard interface to dynamically modify the retransmission policy. This invention solves the key problem in traditional systems where failed SMS retransmissions continue during congestion due to the independent operation and lack of coordination among network elements, leading to further exacerbation of secondary congestion and the occupation of high-priority emergency channels by invalid traffic. Through a proactive cross-system collaborative intervention mechanism, redundant retry traffic can be effectively suppressed at the source, directly reducing the invalid service load on key network elements such as the SMS center platform, core network, and wireless base stations. This ensures more sufficient underlying processing capabilities and channel resources for emergency notification SMS messages in extreme congestion scenarios, significantly improving the stability of the entire SMS delivery system in emergency situations and ensuring the absolute smooth operation of critical services.

[0070] In this embodiment, SMS services are divided into five categories according to their source: intra-provincial point-to-point SMS, intra-provincial point-to-point SMS from other operators, point-to-point SMS from other provinces, service provider (SP) SMS, and industry SMS.

[0071] Intra-provincial point-to-point SMS: SMS messages sent by users to target mobile phone numbers via mobile phones. The sending channel is: sending mobile phone → wireless base station → SMS center → target mobile phone;

[0072] Point-to-point SMS messages from other operators within the province: Point-to-point SMS messages forwarded by the SMS gateways of other telecom operators within the province. The sending channel is: SMS gateway of other telecom operators within the province → SMS gateway → SMS center → target mobile phone.

[0073] Point-to-point SMS messages from other provinces: Point-to-point SMS messages forwarded by SMS gateways of telecom operators in other provinces are sent via the following channel: SMS gateway of other provinces → SMS gateway → SMS center → target mobile phone.

[0074] SP SMS: Commercial SMS messages sent by service providers (SPs), with the sending channel being: SP platform → integrated business gateway → SMS gateway → SMS center → target mobile phone;

[0075] Industry SMS: Service-oriented SMS messages sent by industry SMS platforms, with the sending channel being: Industry SMS Platform → SMS Center → Target Mobile Phone; Emergency notification SMS messages are special industry SMS messages.

[0076] In the SMS sending channels of telecom operators, the SMS center platform is a core network element. Improving SMS delivery congestion hinges on optimizing the SMS center platform. See details... Figure 3 A diagram illustrating the types of SMS services and sending channels.

[0077] Example 2

[0078] Please see Figure 4 This embodiment provides an apparatus for intelligent congestion handling during SMS delivery, applied to a method for intelligent congestion handling during SMS delivery, including:

[0079] The mode determination module is used to identify emergency notification SMS messages and switch working modes according to their traffic thresholds;

[0080] The priority setting module is used to set the priority of SMS messages based on the caller ID and the current mode;

[0081] The channel resource control module is used to allocate dedicated channel resources for different priorities and manage public channel resources, and integrates machine learning models to dynamically adjust resource allocation strategies.

[0082] The SP management module is used to configure the delivery time window for SP SMS messages and to identify and penalize spam SMS messages.

[0083] The cache management module is used to store SMS messages during congestion and manage them according to first-in-first-out and priority-aware rules.

[0084] The collaborative intervention module is used to monitor and intervene in the SMS center platform's resend strategy in emergency mode.

[0085] The implementation process of this embodiment is as follows: The SMS center platform of the provincial telecommunications operator sends the pre-sent SMS content to the SMS delivery congestion intelligent processing device. It first automatically identifies emergency notification SMS and special industry SMS, such as typhoon notification SMS and epidemic notification SMS, and determines whether the number of emergency notification SMS within a unit of time exceeds the set threshold. If it does not exceed the threshold, it enters normal mode; if it exceeds the threshold, it enters emergency mode.

[0086] If entering normal mode, the system first accesses the SMS delivery priority setting module. Based on the caller ID of each SMS message, it intelligently and automatically distinguishes the priority level of SMS delivery: point-to-point SMS messages are set to high priority, industry SMS messages to medium priority, and SP SMS messages to low priority. Next, it enters the SMS delivery channel resource control module. High-priority SMS messages occupy more dedicated SMS delivery channel resources, while medium and low-priority SMS messages occupy less. Public channel resources are used sequentially according to priority. Historical patterns of industry SMS delivery, such as statutory holidays and e-commerce promotional periods, are collected, and machine learning models are used to dynamically adjust the use of public channel resources. High-priority point-to-point SMS messages and medium-priority industry SMS messages enter the SMS delivery queue and are sent to the SMS center platform for delivery. Low-priority SP SMS messages enter the SMS delivery time window control module. This module sets delivery time windows for each SP SMS message, avoiding peak SMS delivery times and ensuring that high-efficiency SMS messages are delivered during prime time, preventing disruption to users during rest periods. It also intelligently identifies SP SMS messages as spam; if a message exceeds a spam threshold, the delivery channel resources for that SP SMS message are halved. At this point, low-priority SP SMS messages also enter the SMS delivery queue and are sent to the SMS center platform for delivery. Additionally, when SMS delivery channel resources are insufficient and SMS congestion occurs, the congested SMS messages are sent to the SMS cache. Once the congestion is relieved, SMS messages are released using a first-in, first-out (FIFO) rule. If the cache is full, the earliest-entered low-priority SMS messages are automatically deleted to ensure that high-priority SMS messages are delivered.

[0087] If emergency mode is entered, the system first accesses the SMS delivery priority setting module. Based on the caller ID of each SMS message, it intelligently and automatically distinguishes the priority level of SMS delivery. Emergency notification SMS messages (special industry SMS messages) are set to high priority, point-to-point SMS messages and other industry SMS messages are set to medium priority, and SP SMS messages are set to low priority. Next, the system accesses the SMS delivery channel resource control module. High-priority SMS messages occupy the majority of dedicated SMS delivery channel resources, medium-priority SMS messages occupy a smaller portion of dedicated SMS delivery channel resources, and low-priority SMS messages are not allocated channel resources. The system periodically receives real-time resend library occupancy data from the SMS center platform. When this data is excessively high, the system alarms, calls an interface to notify the SMS center platform, and automatically modifies the SMS center platform's resend policy. SMS messages that fail to be delivered are not resent. High-priority emergency notification SMS messages, medium-priority point-to-point SMS messages, and other industry SMS messages enter the SMS delivery queue and are sent to the SMS center platform for delivery. The SMS cache library function and operating principles are the same as above.

[0088] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention are within the scope of the present invention.

Claims

1. A method for intelligent congestion handling during SMS delivery, characterized in that, Includes the following steps: S1. Receive pre-delivered SMS messages from the SMS center platform, automatically identify emergency notification SMS messages, and determine whether the number of emergency notification SMS messages per unit time exceeds the preset threshold. If the limit is not exceeded, then enter normal processing mode; If the limit is exceeded, emergency response mode will be activated. S2. Dynamically set the priority of SMS messages based on the entered mode and the caller ID. In the normal processing mode, point-to-point SMS is set to high priority, industry SMS to medium priority, and SP SMS to low priority; in the emergency processing mode, emergency notification SMS is set to high priority, point-to-point SMS and other industry SMS to medium priority, and SP SMS to low priority. S3. Allocate dedicated channel resources in proportion to SMS messages of different priorities, and set up public channel resources to be used in order of priority; at the same time, collect historical data of SMS messages and use machine learning models to dynamically adjust the usage strategy of the public channel resources. S4. For low-priority SP SMS messages, allocate differentiated delivery time windows based on the analysis results of effectiveness, importance and sensitivity, identify SP SMS messages as spam messages, and reduce public channel resources based on the identification results; S5. When congestion is detected due to insufficient channel resources, the congested SMS messages are temporarily stored in the SMS cache and released according to the first-in-first-out rule after the congestion is resolved; when the cache storage space is full, the earliest stored low-priority SMS messages are automatically deleted. S6. In emergency handling mode, periodically obtain the real-time occupancy rate data of the resend library of the SMS center platform, and when the alarm threshold is exceeded, send an instruction to the SMS center platform to modify the resend strategy so that SMS messages that fail to be sent will no longer be resent.

2. The method for intelligent congestion handling during SMS delivery according to claim 1, characterized in that, In step S1, the automatic identification of emergency notification SMS messages includes: Obtain the caller ID and text content of the SMS message to be identified; The calling number is compared with a preset emergency number feature database; The SMS content text is analyzed by keyword matching using an efficient multi-pattern string matching algorithm based on Trie trees. The keywords are preset words or phrases used to identify the characteristics of emergency events. Based on the combined results of the caller ID comparison and content keyword matching analysis, if the caller ID is successfully matched or the content matches a preset keyword, the SMS message is determined to be an emergency notification SMS message.

3. The method for intelligent congestion handling during SMS delivery according to claim 2, characterized in that, This paper describes a method for analyzing the keyword matching of SMS text using a high-efficiency multi-pattern string matching algorithm based on Trie trees. The method involves: first, constructing a Trie tree from a keyword database of core terms in the emergency response field; when the SMS text to be identified arrives, using it as the input stream, performing a single scan traversal starting from the root node of the Trie tree, and simultaneously completing parallel matching and detection of all keywords.

4. The method for intelligent congestion handling during SMS delivery according to claim 1, characterized in that, In step S3, the usage strategy for the public channel resources is dynamically adjusted using a machine learning model, including: Continuously collect and store historical delivery data including time, business type, resource allocation weight, and performance indicators to build a feature dataset; Based on the feature dataset, a time series prediction and multi-objective optimization model is trained to predict the demand for public channel resources for each priority business in future periods. Based on the real-time system status and prediction results, the optimized public channel resource allocation weights are dynamically output; the resource scheduler adjusts the public channel scheduling strategy in real time according to the public channel resource allocation weights.

5. The method for intelligent congestion handling during SMS delivery according to claim 1, characterized in that, In step S4, based on the analysis results of effectiveness, importance, and sensitivity, differentiated distribution time windows are allocated, including: A multidimensional evaluation model is constructed for each service provider to quantitatively analyze its historical SMS revenue contribution rate, whether its business involves key public services, and the proportion of user complaints and unsubscriptions, and to calculate a comprehensive score for effectiveness, importance, and sensitivity. Based on preset scoring thresholds and mapping strategies, SPs are automatically categorized into different service levels. Each service level is bound to a preset delivery time window policy, and the delivery of the time window configuration is enforced.

6. The method for intelligent congestion handling during SMS delivery according to claim 5, characterized in that, The construction of the multi-dimensional evaluation model includes: based on a preset quantitative indicator system, calculating the historical revenue contribution rate and unit traffic value of SP SMS services from the perspective of effectiveness; analyzing whether its services are related to key service types from the perspective of importance; and statistically analyzing user complaint rate, unsubscription rate, and content compliance score from the perspective of sensitivity; and merging the quantitative results of the three dimensions through a preset weighted algorithm to generate a comprehensive evaluation score for each SP, and automatically completing the classification and categorization of service levels based on the comprehensive evaluation score and preset level thresholds.

7. The method for intelligent congestion handling during SMS delivery according to claim 1, characterized in that, In step S5, temporarily storing congested SMS messages in the SMS cache and then releasing them specifically includes: When insufficient channel resources are detected, congested SMS messages are stored in the corresponding high, medium and low priority logical queues in the cache according to their priority identifiers. Each queue is organized in order of SMS arrival time. When congestion is detected to be relieved, the SMS messages in each queue, arranged according to the first-in-first-out rule, are released in the order of high, medium, and low priority queues. If the cache is full during the congestion period, a replacement mechanism is activated, prioritizing the deletion of the earliest-stored SMS messages in the low-priority queue.

8. The method for intelligent congestion handling during SMS delivery according to claim 1, characterized in that, In step S6, an instruction is sent to the SMS center platform to modify the retransmission policy, including: Real-time occupancy data of the SMS center platform's resend database is collected at fixed intervals through a dedicated monitoring agent. When the occupancy rate continuously exceeds the alarm threshold for a preset period, a strategy adjustment instruction is automatically generated. The instruction is sent to the SMS center platform via a predefined application programming interface (API) call, so that the SMS center platform can dynamically update the resend rules in its policy engine.

9. An apparatus for intelligent congestion handling during SMS delivery, applied to the method for intelligent congestion handling during SMS delivery as described in any one of claims 1-8, characterized in that, include: The mode determination module is used to identify emergency notification SMS messages and switch working modes according to their traffic thresholds; The priority setting module is used to set the priority of SMS messages based on the caller ID and the current mode; The channel resource control module is used to allocate dedicated channel resources for different priorities and manage public channel resources, and integrates machine learning models to dynamically adjust resource allocation strategies. The SP management module is used to configure the delivery time window for SP SMS messages and to identify and penalize spam SMS messages. The cache management module is used to store SMS messages during congestion and manage them according to first-in-first-out and priority-aware rules. The collaborative intervention module is used to monitor and intervene in the SMS center platform's resend strategy in emergency mode.

Images