Resource scheduling method and system for traffic integrated controller, and computer storage medium

Abstract

The invention discloses a traffic integrated controller resource scheduling method and system and a computer storage medium, and the method comprises the steps: carrying out the protocol recognition of an input instruction data packet, so as to extract a multi-dimensional feature; reducing the instruction data packet into a complete traffic control instruction by adopting a flow recombination algorithm based on a sliding window; analyzing the semantics of the traffic control instruction, and determining the initial priority of the traffic control instruction according to the multi-dimensional features; calculating a dynamic priority corresponding to the traffic control instruction according to the initial priority; and triggering a multi-level resource preemption strategy according to the dynamic priority.

Description

Technical Field

[0001] The embodiments of this specification relate to the field of traffic control technology, and more particularly to a method and system for resource scheduling of a centralized traffic controller and a computer storage medium. Background Art

[0002] With the rapid development of traffic control systems, modern traffic control systems are evolving towards intelligence, centralization, and collaboration. Regional-level centralized traffic control centers are required to process massive amounts of heterogeneous traffic control commands and data, each with varying levels of urgency, timeliness, and resource requirements. Traditional traffic control systems typically address resource utilization through hardware resource stacking and expansion. A few systems consider designing resource scheduling solutions, but these solutions still present numerous challenges.

[0003] First, resource allocation is static and rigid. Using a static resource configuration strategy, it is impossible to dynamically adjust the resource allocation ratio based on real-time traffic conditions and system load. During peak traffic hours or emergencies, this static allocation method is prone to resource contention and system overload. Second, the priority determination mechanism is simple and crude. Most priority determination methods are based on fixed rules, considering only simple timestamp information and failing to deeply analyze the characteristics of the instruction content itself and the importance of the business. Third, the protocol has poor adaptability and limited ability to identify different traffic control equipment protocols. It is unable to accurately parse the protocol-specific priority indication information, resulting in low priority determination accuracy. Summary of the Invention

[0004] In view of this, the embodiments of this specification provide a traffic centralized controller resource scheduling method, system and computer storage medium to solve the problems of static and rigid resource allocation, simple and rough priority judgment mechanism, and poor protocol adaptability in traffic control systems in the prior art.

[0005] The embodiments of this specification adopt the following technical solutions: This embodiment of the present invention provides a method for scheduling resources of a centralized traffic controller, the method comprising: Perform protocol recognition on the input command data packet to extract multi-dimensional features; Using a flow reassembly algorithm based on a sliding window to restore the instruction data packet into a complete traffic control instruction; parsing the semantics of the traffic control instruction and determining an initial priority of the traffic control instruction based on the multi-dimensional features; Calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority; A multi-level resource preemption strategy is triggered according to the dynamic priority.

[0006] The embodiment of this specification further provides a traffic centralized controller resource scheduling system, the traffic centralized controller resource scheduling system comprising: The recognition module performs protocol recognition on the input command data packet to extract multi-dimensional features; A restoration module, which uses a flow reassembly algorithm based on a sliding window to restore the instruction data packet into a complete traffic control instruction; a parsing module, which parses the semantics of the traffic control instruction and determines the initial priority of the traffic control instruction based on the multi-dimensional features; a calculation module, calculating a dynamic priority corresponding to the traffic control instruction according to the initial priority; The preemption module triggers a multi-level resource preemption strategy according to the dynamic priority.

[0007] The embodiments of this specification also provide a computer storage medium including a program for use in conjunction with an electronic device, the program being executable by a processor to perform the following steps: Perform protocol recognition on the input command data packet to extract multi-dimensional features; Using a flow reassembly algorithm based on a sliding window to restore the instruction data packet into a complete traffic control instruction; parsing the semantics of the traffic control instruction and determining an initial priority of the traffic control instruction based on the multi-dimensional features; Calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority; A multi-level resource preemption strategy is triggered according to the dynamic priority.

[0008] At least one of the above technical solutions adopted in the embodiments of this specification can achieve the following beneficial effects: Through the automatic implementation process of protocol identification, semantic analysis, dynamic priority, and resource preemption, and by identifying communication protocols and transmission data characteristics, the priority of traffic control instructions is calculated and optimized, and resource allocation and complex balanced scheduling are carried out based on priority. Through multi-level and multi-dimensional instruction analysis and dynamic adaptive resource allocation strategies, dynamic resource scheduling of massive heterogeneous traffic instructions is achieved, ensuring real-time response to high-priority instructions. It solves the three major technical problems of rigid static resource allocation, rough priority determination, and poor protocol adaptability, and improves the response speed of high-priority instructions by 300% in traffic evacuation scenarios of 10,000 people. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The drawings described herein are used to provide a further understanding of the embodiments of this specification and constitute a part of the embodiments of this specification. The illustrative embodiments and descriptions of this specification are used to explain this application and do not constitute an improper limitation on this application. In the drawings: Figure 1A flow chart of a method for resource scheduling of a centralized traffic controller provided in an embodiment of this specification; Figure 2 A schematic diagram of a risk transmission mechanism corresponding to a resource scheduling method for a centralized traffic controller provided in an embodiment of this specification; Figure 3 A schematic diagram of a three-level processing architecture corresponding to a traffic centralized controller resource scheduling method provided in an embodiment of this specification; Figure 4 A schematic diagram of the structure of a traffic centralized controller resource scheduling system provided in an embodiment of this specification; Figure 5 A schematic diagram of the structure of a computer storage medium corresponding to a traffic centralized controller resource scheduling method provided in an embodiment of this specification. DETAILED DESCRIPTION

[0010] To make the objectives, technical solutions, and advantages of this application more clear, the technical solutions of this application will be clearly and completely described below in conjunction with the specific embodiments of this specification and the corresponding drawings. Obviously, the embodiments described are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this application.

[0011] The technical solutions provided by the embodiments of this specification are described in detail below with reference to the accompanying drawings.

[0012] like Figure 1 FIG. 1 is a flow chart of a method for resource scheduling of a centralized traffic controller provided in an embodiment of this specification.

[0013] In the embodiment of this specification, the traffic centralized controller resource scheduling method may specifically include the following steps: S101: performing protocol recognition on the input command data packet to extract multi-dimensional features; S103: using a sliding window-based stream reassembly algorithm to restore the instruction data packet into a complete traffic control instruction; S105: parsing the semantics of the traffic control instruction and determining the initial priority of the traffic control instruction based on the multi-dimensional features; S107: Calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority; S109: triggering a multi-level resource preemption strategy according to the dynamic priority.

[0014] In the embodiments of this specification, the resource scheduling method of the traffic centralized controller adopts an automatic implementation process of protocol identification → semantic analysis → dynamic priority → resource preemption. Through multi-level and multi-dimensional instruction analysis and dynamic adaptive resource allocation strategies, dynamic resource scheduling of massive heterogeneous traffic instructions is achieved to ensure real-time response of high-priority instructions.

[0015] As an application embodiment of this specification, in step S101, protocol identification is performed on the input instruction data packet to extract multi-dimensional features, which may specifically include: Performing protocol identification on the instruction data packet using an improved Wu-Manber algorithm, wherein the improved Wu-Manber algorithm uses an FNV-1a rolling hash function to calculate a hash value of a pattern string and skips invalid characters based on a dynamic skip strategy; Extract multi-dimensional features from the identified protocol packets.

[0016] In the embodiments of this specification, the original traffic control instructions issued by the application layer are input as individual instruction packets. A complete traffic control instruction may span multiple instruction packets, specifically TCP (Transmission Control Protocol) fragmented packets. Extracting multi-dimensional features during the packet fragmentation stage can reduce subsequent workload and provide foundational support for subsequent resource scheduling.

[0017] In the prior art, the existing Wu-Manber algorithm is an efficient multi-pattern string matching algorithm that optimizes matching efficiency through character block jumping, hash filtering, and parallel processing.

[0018] The improved Wu-Manber algorithm significantly reduces the protocol recognition time through the FNV-1a rolling hash function and dynamic jump strategy, can achieve high-performance protocol recognition, identify private protocols or standard protocols of different types and models of devices, and solve the compatibility problem of private protocols.

[0019] Furthermore, the improved Wu-Manber algorithm is used to perform protocol identification on the instruction data packet, which may specifically include: Calculate the hash value of the instruction data packet using the FNV-1a rolling hash algorithm; matching the hash value with historical hash values ​​stored in a historical database; When the hash value matches successfully but the character comparison fails, the dynamic jump strategy is executed; Dynamically parse the variable-length field based on the successfully matched protocol type.

[0020] The dynamic jump strategy may specifically include: Calculating a first jump distance based on the character table; If the failure position is at the end of the pattern string, the protocol header feature of the next data block is detected to calculate the second jump distance; The minimum value of the first jump distance and the second jump distance is taken as the final jump step length.

[0021] In the embodiments of this specification, using FNV-1a rolling hash can reduce conflicts, and repeated scanning can be avoided by performing fast incremental hash calculations on protocol signatures.

[0022] In addition, in actual application, common protocols can be marked, and hash values ​​can be pre-calculated and stored as a historical database, which can effectively avoid repeated calculations and shorten protocol identification time. For example, NTCIP (National Transportation Communications for ITS Protocol) can be marked with a hash value of "1203", and private protocols can be marked with a hash value of "#EMG!".

[0023] The dynamic jump strategy can adapt to the protocol characteristics, dynamically adjust the jump step size according to the protocol characteristics, and reduce invalid matches.

[0024] In the embodiments of this specification, the multi-dimensional features may specifically include: basic features, business features and semantic features.

[0025] Among them, basic features may include packet length, source / destination address, protocol version, etc., business features may include traffic control instruction type, control object ID, timeliness mark, etc., and semantic features may include instruction issuance scope, expected execution time, etc., which are not specifically limited here.

[0026] As an application embodiment of this specification, in step S103, a stream reassembly algorithm based on a sliding window is used to restore the instruction data packet to a complete traffic control instruction, which may specifically include: A flow reassembly algorithm based on a sliding window is adopted to reassemble the fragmented instruction data packets into a complete data stream according to the sequence number and confirmation number of the instruction data packet, and restore the complete traffic control instruction.

[0027] In the embodiments of this specification, since the original traffic control instructions issued by the application layer are input in the form of individual instruction data packets, the subsequent implementation of resource scheduling requires complete traffic control instructions, that is, instructions with complete business semantics, rather than network layer fragments. Otherwise, due to field fragmentation, business characteristics will be distorted (fields across fragments cannot be parsed), semantic associations will be broken (the logic between instructions cannot be established, for example, "turning off the red light" and "turning on the green light" are misjudged as unrelated isolated events), erroneous parsing and security risks, etc., which will affect resource scheduling.

[0028] Among them, the stream reassembly algorithm is mainly used in network communications, especially in network intrusion detection systems (IDS), to reassemble TCP streams in order to better analyze and detect abnormal behaviors in network traffic.

[0029] By restoring traffic control instructions, the problem of disordered fragmentation of instruction data packets can be solved, ensuring that instructions are input into the analysis module in the logical order of the application layer. A complete event chain can also be constructed to identify continuous instructions, thereby supporting the priority pre-classification of traffic control instructions.

[0030] As an application embodiment of this specification, in step S105, parsing the semantics of the traffic control instruction and determining the initial priority of the traffic control instruction based on the multi-dimensional features may specifically include: Performing correlation analysis based on the context of the traffic control instructions to obtain temporal relationships, causal relationships, and semantic relationships between the traffic control instructions; Parsing the semantics of the traffic control instruction to obtain a parsing result; The priority of the traffic control instruction is pre-classified according to the analysis result and the multi-dimensional feature to obtain the initial priority.

[0031] In the embodiment of this specification, after the traffic control instruction is restored, it is necessary to further determine whether the business characteristics and semantic characteristics of the traffic control instruction are correct, and whether the instructions are related or continuous.

[0032] Among them, the temporal relationship between the traffic control instructions specifically refers to the timestamp sequence of the instructions, which must be sorted according to the occurrence time of the complete instructions; the causal relationship specifically refers to the logical correlation of the instruction content, for example, instruction A (turning on the green light on the main road) triggers instruction B (turning off the red light on the branch road); the semantic relationship specifically refers to the semantic tags across instructions, which are used to identify the "regional collaborative scheduling" instruction group.

[0033] Furthermore, the priority of the traffic control instruction is pre-classified according to the analysis result and the multi-dimensional features to obtain the initial priority. Specifically, the initial priority of the traffic control instruction can be calculated according to the following formula: The initial priority of the traffic control instruction is calculated according to the following formula: Formula (1).

[0034] in, is the protocol base weight, is the packet feature analysis score, is the urgency of semantic content, a, b, c are their corresponding weight coefficients, which can be preset according to experimental experience, a+b+c=1.

[0035] The protocol basic weight can be obtained based on the static protocol features of the multi-dimensional features, the packet feature analysis score can be obtained based on the packet basic features, and the semantic content urgency can be obtained based on the semantic features.

[0036] By pre-classifying the priorities of traffic control instructions, an initial priority benchmark can be provided for the decision-making layer of resource scheduling, which is a coarse-grained priority screening.

[0037] As an application embodiment of this specification, in step S107, calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority may specifically include: The dynamic priority is calculated using a four-dimensional dynamic priority model, wherein the four-dimensional dynamic priority model is calculated according to the following formula: Formula (2).

[0038] in, 、 、 、 is the dynamic weight coefficient, + + + =1, default α=0.4, β=0.3, γ=0.2, δ=0.1, is the instruction retention time (unit: ms), τ is the system time constant (configurable, default 5000ms), is the current amount of available resources (normalized value 0-1), is the total resource (normalized to 1.0), is the jth semantic feature value (normalized to 0-1), is the feature weight (dynamically adjusted through online learning), is the safety verification result (0-1), , =2.5, =0.6 is an empirical parameter.

[0039] Specifically, =2.5, =0.6 is obtained based on 100,000 sets of training data.

[0040] τ is a dynamically configurable system time constant. After verification with 100,000 training data, a value of 5000ms increased the convergence speed by 60%, so it is used as the default value.

[0041] Similarly, 、 、 、 The default value of The standardized values ​​of are obtained based on a large amount of training data and are not specifically limited here.

[0042] In the embodiment of this specification, the four-dimensional dynamic priority model calculates the dynamic priority of the transmission instruction from four aspects: time, resources, semantics and security. ,and The value is the final execution basis for guiding the controller resource scheduling. In multi-dimensional features, semantic features dominate the priority rather than protocol fields.

[0043] pass Quantify the importance of multi-dimensional features and avoid strong dependence on protocol standards. and Indicates real-time system status, Indicates the safety verification result. is the semantic feature weight.

[0044] The semantic feature term in the above formula (2) Directly inherited from The semantic analysis results of ), by The extracted semantic features (e.g. instruction type, impact range, etc.) are directly input into of Items can avoid repeated parsing, but Dynamic weighted optimization was performed.

[0045] exist Based on the static characteristics of , three new dynamic factors are added: Time decay term ( ), used for response timeliness; Resource margin item ( ), used to indicate system load adaptation; Security check items ( ), used to indicate instruction safety performance.

[0046] in, The time decay term can be used to solve Inherent defects: If a high The instruction is stuck due to system overload. The value automatically degrades over time to prevent low-priority tasks from being unable to execute.

[0047] By adding independent security check items, illegal instructions (such as forged signal instructions or malicious instructions, etc.) can be prevented even if High value, The value will still be suppressed, and the dynamic priority of the illegal instruction will still not be high, so it will not be executed.

[0048] Furthermore, after calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority, the method further includes: Dynamically adjust weight parameters based on reinforcement learning mechanism 、 、 , the weight parameters are updated at predetermined time intervals.

[0049] In a specific application scenario, the weight parameters are updated according to the following formula: Formula (3) Formula (4) Formula (5) in, is the learning rate (default 0.05), is the actual average response delay, is the target delay (configurable), is the current resource utilization, is the target resource utilization (default 0.7), is the accuracy of the most recent N priority decisions.

[0050] Specifically, η is a dynamically adjustable learning rate parameter, is the dynamically adjustable target resource utilization rate, verified by 100,000 training data, and η is set to 0.05. When 0.7 is selected, the convergence speed increases by 40%, so it is used as the default value.

[0051] Further, the security verification results It can be calculated according to the following formula: Formula (6) in, The pass rate of digital signature verification is used to detect the legitimacy of the instruction source. is the instruction behavior risk probability, used for abnormal operation pattern recognition and detection, Score content compliance for protocol field validity verification.

[0052] The security function of digital signature verification is that, on the one hand, it can defend against forgery attacks and ensure that the instructions come from the authorized control center; on the other hand, it can prevent tampering, and any modification of the instructions will cause the signature to become invalid.

[0053] The risk probability of instruction behavior can be calculated specifically from the following detection dimensions: The frequency is abnormal, and the sliding window command count is greater than the threshold. For example, 50 red light commands are received within 1 second. Sequence violations can be calculated by calculating the hidden Markov model (HMM) state transition probability, for example, a green light instruction is not followed by a yellow light instruction; Spatiotemporal conflict, geo-fence crossing detection, for example, the same intersection is operated by different controllers simultaneously; Resource mutations, abnormal CPU / memory usage derivatives, for example, a single instruction triggers 80% CPU usage.

[0054] It should be noted that the above detection dimensions are only examples, and the specific detection dimensions can also be changed according to the actual application scenario and are not specifically limited here.

[0055] The content compliance score can be calculated according to the following rules: Protocol version compatibility; field value range check; reserved bit detection.

[0056] The weight distribution in the above formula (6) is based on 60% for basic identity authentication, 30% for advanced persistent threat detection, and 10% for protocol compliance. The risk transmission mechanism diagram is as follows: Figure 2 As shown in the figure, when any sub-item fails completely, the maximum value of the formula = 0.6×1 + 0.3×1 + 0.1×0 = 0.9, which can effectively avoid a complete collapse caused by a single point failure.

[0057] By securely authenticating traffic control instructions and utilizing the triple protection of cryptographic verification + behavioral analysis + protocol compliance, a multi-layered defense is formed, covering the security chain from identity authentication to content integrity. The innovative use of S-type functions converts discrete behavioral characteristics into continuous risk probabilities for dynamic risk quantification. The 60 / 30 / 10 weight ratio has been verified through attack and defense tests, and an engineered weight design has been implemented to achieve the best balance between security and performance. The upper limit constraint of the formula (≤0.9) and the graded response mechanism can also prevent single point failures from causing system paralysis.

[0058] As an application embodiment of this specification, in step S109, triggering a multi-level resource preemption strategy according to the dynamic priority may specifically include: Set full preemption threshold and partial preemption threshold; When the dynamic priority is greater than the full preemption threshold, starting the resource full preemption mode; When the dynamic priority is greater than the partial preemption threshold and less than or equal to the full preemption threshold, starting the resource partial preemption mode; When the dynamic priority is less than or equal to the partial preemption threshold, the resource negotiation preemption mode is started.

[0059] In the embodiments of this specification, the resource full preemption mode specifically means immediately terminating the currently processed low-priority task and reclaiming all its resources, and then allocating these resources to the high-priority task. For example, the emergency evacuation instruction Pi = 0.92, which is greater than the full preemption threshold of 0.8, or a security incident, < 0.3.

[0060] For example, in a traffic accident scenario, an emergency traffic control command is issued: a traffic accident at an intersection requires an emergency all-green light channel. In this case, the data analysis service is immediately terminated (reclaiming the 4-core CPU + 8GB of memory), and all resources are allocated to the signal control service to execute the emergency traffic control command with a time delay of <15ms (FPGA acceleration).

[0061] Partial resource preemption mode specifically means reclaiming a certain percentage of resources used by low-priority tasks (for example, 30%-70%) and allocating the reclaimed resources to high-priority tasks. Low-priority tasks can continue to run, but with reduced resources.

[0062] For example, during the morning rush hour, traffic control instructions are issued to coordinate green light waves at 10 intersections. In this case, 40% of CPU resources are reclaimed from the video analysis service (originally occupying 8 cores → retaining 4.8 cores) and allocated to the signal optimization service, improving computing throughput. As a result, the signal optimization latency is reduced from 120ms to 45ms.

[0063] The resource negotiation preemption mode specifically refers to sending a resource release request to the low-priority task, allowing the low-priority task to release some resources on its own within a timeout period. If the low-priority task fails to release the resources within the timeout period, the system will forcibly reclaim them.

[0064] For example, when issuing a traffic control instruction: update map data during off-peak hours (Pi=0.58), the scheduler requests to release 2 CPU cores, and the map service responds: 1 core can be released (because key road data is being generated). The scheduler reclaims 1 core and allocates it to the log analysis service.

[0065] To ensure real-time resource preemption, the embodiment of this specification has a built-in high-performance state management engine when preempting resources. It uses write-time copy technology to complete the preservation and recovery of task status within 5ms, and cooperates with the resource recovery mechanism based on FPGA acceleration to ensure that the preemption delay is controlled within 15ms.

[0066] In specific application scenarios, the typical value of the full preemption threshold can be set to 0.8, which is the emergency response threshold and can be applied to traffic accident emergency instructions. The typical value of the partial preemption threshold can be set to 0.6, which is the high priority threshold and can be applied to green light coordination instructions during peak hours.

[0067] In addition, you can set the interval buffer zone threshold to 0.2 to prevent priority jitter from triggering frequent preemption.

[0068] In actual application, the specific values ​​of the full preemption threshold and the partial preemption threshold are dynamically adjusted.

[0069] For example, when the system load is >90%, the threshold is automatically lowered (for example, the full preemption threshold becomes 0.75, the partial preemption threshold becomes 0.55) to enhance the preemption strength. The embodiments of this specification provide a method for resource scheduling of a centralized traffic controller. This method calculates and optimizes the priority of traffic control instructions through an automatic implementation process of protocol identification, semantic analysis, dynamic priority, and resource preemption. By identifying communication protocols and transmission data characteristics, it performs resource allocation and load balancing scheduling based on priority. Through multi-level and multi-dimensional instruction analysis and dynamic adaptive resource allocation strategies, it achieves dynamic resource scheduling of massive heterogeneous traffic instructions, ensures real-time response to high-priority instructions, and solves the three major technical problems of rigid static resource allocation, rough priority determination, and poor protocol adaptability. In a traffic evacuation scenario with 10,000 people, it improves the response speed of high-priority instructions by 300%.

[0070] It should be noted that the above-mentioned specific traffic centralized controller resource scheduling method is only a specific application embodiment and does not limit the scope of the embodiments of this specification. It may also include other specific embodiments, which will not be described one by one here.

[0071] Based on the same inventive concept, the embodiments of this specification also provide specific application embodiments of the above-mentioned traffic centralized controller resource scheduling method.

[0072] like Figure 3 , which is a schematic diagram of a three-level processing architecture corresponding to a traffic centralized controller resource scheduling method provided in an embodiment of this specification.

[0073] In the embodiments of this specification, a three-level intelligent resource scheduling architecture is proposed to solve the problem of difficulty in coordinating massive heterogeneous traffic control instructions and data resources caused by the simultaneous access of all traffic control devices to the centralized controller. The architecture includes a DPI pre-processing layer, a priority decision layer, and a resource scheduling layer. Each layer works together to achieve efficient traffic control resource management and solve the problem of resource occupation of massive heterogeneous traffic data faced by regional-level traffic control centralized control centers / centralized controllers.

[0074] In the specific implementation process, the DPI preprocessing layer receives traffic control instructions, performs protocol recognition through the improved Wu-Manber algorithm, extracts instruction features, and performs flow reorganization and context association to perform instruction priority preprocessing.

[0075] The priority decision layer performs dynamic priority calculation and weight adaptive adjustment based on the instruction priority, and performs security verification on the instruction to generate a control instruction with accurate priority information.

[0076] The resource scheduling layer performs resource allocation and task scheduling according to the control instructions with priority information through a refined resource preemption module and a complex balancing module to ensure that high-priority instructions are processed first.

[0077] Among them, the DPI preprocessing layer: The DPI pre-processing layer uses an improved Wu-Manber multi-pattern matching algorithm and protocol feature analysis technology to achieve in-depth analysis of traffic control instructions.

[0078] The DPI preprocessing layer can specifically include the following modules: (1) Protocol identification module The improved Wu-Manber algorithm is used to achieve high-performance protocol recognition, identifying proprietary or standard protocols for different types and models of devices. The algorithm significantly reduces protocol recognition time and increases recognition speed by 40% through the FNV-1a rolling hash function and dynamic jump strategy.

[0079] The performance comparison data of the traditional Wu-Manber algorithm and the improved Wu-Manber algorithm provided in the embodiments of this specification are shown in Table 1 below: Table 1 Performance comparison of the traditional Wu-Manber algorithm and the improved Wu-Manber algorithm index Traditional Wu-Manber algorithm Improved Wu-Manber algorithm Improvement effect Average matching time 4.2 μs / instruction 2.5 μs / instruction ↑40% Hash collision rate 18% 5% ↓72% Emergency command recognition delay 15 ms 8 ms ↓47% Test environment: 100,000 real traffic instructions (including NTCIP / private protocol mixed traffic).

[0080] (2) Instruction feature extraction module Multi-dimensional features are extracted from the identified protocol data packets, including basic features such as packet length, source / destination address, protocol version, business features such as traffic control instruction type, control object ID, timeliness mark, and semantic features such as instruction issuance scope and expected execution time.

[0081] (3) Stream reassembly and context association module A sliding window-based stream reassembly algorithm is used to process fragmented data packets. Based on the TCP sequence number and acknowledgment number, the segmented TCP data packets are reassembled into a complete data stream to restore the complete instruction. Context-based correlation analysis is performed to analyze the temporal, causal, and semantic relationships between instructions, parse the semantics of the transmitted traffic control instructions, and perform priority pre-classification to obtain the initial priority of the traffic control instructions. .

[0082] For example, command data packet fragment 1 [ID=123, Cmd=Gree] + command data packet fragment 2 [n_Light=ON,Dur=30s] are restored to the complete command: {ID:123, Command:"GreenLight_ON", Duration:30}.

[0083] Priority decision-making layer: The precise priority of traffic control instructions is calculated through multi-dimensional feature fusion and dynamic weight adjustment.

[0084] The priority decision layer can specifically include the following modules: (1) Dynamic priority calculation module The dynamic priority Pi of the transmission instruction is calculated from four aspects: time, resources, semantics and security.

[0085] (2) Weight adaptive adjustment module A weight adjustment strategy based on reinforcement learning is adopted. The weight parameters (α, β, γ) are dynamically adjusted based on reinforcement learning, and the weight is updated every Δt time (default 200ms).

[0086] (3) Security verification module The security verification module performs multi-level security checks on instructions to ensure their legality, integrity, and security, including ECDSA-based digital signature verification, anomaly detection based on instruction behavior analysis, and content security policy checks.

[0087] Resource scheduling layer: The resource scheduling layer implements dynamic and precise resource allocation based on the calculated priorities. The resource scheduling layer design adopts a dual-module collaborative architecture. Through the organic cooperation of the refined resource preemption module and the intelligent load balancing module, it realizes the efficient management and dynamic allocation of traffic control system resources.

[0088] The resource scheduling layer can specifically include the following modules: (1) Refined resource preemption module The refined resource preemption module uses an improved Min-Min algorithm, combined with instruction priority and resource status, to achieve refined resource preemption. The specific steps are as follows: 1. Task sorting: Sort tasks according to the dynamic priority of the instructions, with tasks with higher priority at the front.

[0089] 2. Resource evaluation: Evaluate the expected completion time of each resource processing task 3. Task allocation: Assign tasks to the resource that can complete them the earliest. If the resource is currently processing a low-priority task, preempt the resource.

[0090] The refined resource preemption module employs a multi-level response strategy, automatically selecting full, partial, or negotiated preemption based on priority. The module's built-in high-performance state management engine utilizes copy-on-write technology to save and restore task states within 5ms. Combined with an FPGA-accelerated resource recovery mechanism, preemption latency is kept below 15ms.

[0091] The implementation of the copy-on-write technology can be memory page marking + FPGA dirty page tracking, so that the state is saved in <5ms.

[0092] (2) Load balancing module The load balancing module adopts a layered design architecture. The entry layer uses an optimized Nginx cluster to achieve traffic distribution at 100,000 queries per second. The microservice layer integrates an intelligent load prediction algorithm that dynamically adjusts instance weights based on six metrics, including CPU and memory. An incremental synchronization mechanism maintains resource consistency between the two nodes, ensuring failover latency is less than 200ms.

[0093] For example, when the CPU utilization is greater than 85%, the resource margin weight β is automatically increased (Δβ>0); When the security threat probability is >30%, increase the safety factor δ to 0.3.

[0094] The resource scheduling method designed in the embodiments of this specification is applicable to a centralized control platform with a dual-node deployment. Nodes A and B are deployed in active-active mode. Heartbeat detection is used between the two nodes to achieve fault detection, and the resource database uses semi-synchronous replication to ensure state consistency. The two core modules achieve deep collaboration through a unified resource database. When the system detects insufficient resources, the load balancing module triggers the preemption process and intelligently selects the victim based on service priority.

[0095] For example, in a traffic security scenario for a major event, a large passenger flow needs to be evacuated in a short period of time, which involves the linkage of traffic control equipment at five intersections in the area, mainly the coordinated linkage of traffic signals.

[0096] The resource scheduling process is as follows: 1. DPI preprocessing layer First, the Wu-Manber algorithm is used to identify the instruction information sent by the current upper-layer application platform to the signal machine, and key features are extracted, including protocol type, urgency, duration, and impact range. Based on the priority pre-classification, it is determined to be a high-priority instruction.

[0097] For example, the DPI layer identifies the semantic features of a traffic light instruction: {Instruction type: emergency green light extension, Impact range: 5 intersections, Timeliness: 2 seconds}.

[0098] 2. Priority decision-making layer The dynamic priority calculation module performs priority calculation to obtain the Pi value. When it detects that the CPU occupancy rate exceeds the set target value, it is input to the weight adaptive adjustment module to automatically adjust the β weight.

[0099] For example, the priority decision layer calculates Pi = 0.92 (because the resource utilization rate reaches 90%, the β weight increases to 0.5).

[0100] 3. Resource Scheduling Layer Control instructions with instruction priority information trigger the partial preemption mechanism of the resource scheduling layer. The refined resource preemption module reclaims 30% of CPU resources from lower-priority services such as data analysis, while allocating dedicated CPU cores and memory resources to signal control services. The load balancing module adjusts traffic distribution in real time, directing 80% of requests to nodes with lighter loads, ultimately achieving high load and high availability under centralized control.

[0101] For example, the scheduling layer triggers partial preemption: 40% of CPU resources are recovered from the data analysis service and allocated to the signal control service, and the load balancer directs 90% of the traffic to idle nodes.

[0102] The specific implementation process of the embodiments of this specification can refer to the corresponding implementation steps of the above embodiments, which will not be repeated here.

[0103] Based on the same inventive concept, the embodiment of this specification also provides a traffic centralized controller resource scheduling system. Figure 4 FIG. 1 is a schematic diagram of the structure of a traffic centralized controller resource scheduling system provided in an embodiment of this specification.

[0104] The centralized traffic controller resource scheduling system may specifically include: Identification module 401 performs protocol identification on the input instruction data packet to extract multi-dimensional features; Restoration module 402, using a sliding window-based stream reassembly algorithm to restore the instruction data packet into a complete traffic control instruction; The parsing module 403 parses the semantics of the traffic control instruction and determines the initial priority of the traffic control instruction based on the multi-dimensional features; A calculation module 404 calculates a dynamic priority corresponding to the traffic control instruction according to the initial priority; The preemption module 405 triggers a multi-level resource preemption strategy according to the dynamic priority.

[0105] based on Figure 4 The present specification also provides some specific implementation plans of the system, which are described below.

[0106] Furthermore, protocol recognition is performed on the input command data packet to extract multi-dimensional features, including: Performing protocol identification on the instruction data packet using an improved Wu-Manber algorithm, wherein the improved Wu-Manber algorithm uses an FNV-1a rolling hash function to calculate a hash value of a pattern string and skips invalid characters based on a dynamic skip strategy; Extract multi-dimensional features from the identified protocol packets.

[0107] Furthermore, a flow reassembly algorithm based on a sliding window is used to restore the instruction data packet into a complete traffic control instruction, including: A flow reassembly algorithm based on a sliding window is adopted to reassemble the fragmented instruction data packets into a complete data stream according to the sequence number and confirmation number of the instruction data packet, and restore the complete traffic control instruction.

[0108] Furthermore, parsing the semantics of the traffic control instruction and determining the initial priority of the traffic control instruction based on the multi-dimensional features includes: Performing correlation analysis based on the context of the traffic control instructions to obtain temporal relationships, causal relationships, and semantic relationships between the traffic control instructions; Parsing the semantics of the traffic control instruction to obtain a parsing result; The priority of the traffic control instruction is pre-classified according to the analysis result and the multi-dimensional feature to obtain the initial priority.

[0109] Furthermore, the priority of the traffic control instruction is pre-classified according to the analysis result and the multi-dimensional features to obtain the initial priority, including: The initial priority of the traffic control instruction is calculated according to the following formula: in, is the protocol base weight, is the packet feature analysis score, is the semantic content urgency, , b, c are the corresponding weight coefficients, which can be pre-set according to experimental experience. +b+c=1.

[0110] Furthermore, calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority includes: The dynamic priority is calculated using a four-dimensional dynamic priority model, wherein the four-dimensional dynamic priority model is calculated according to the following formula: in, 、 、 、 is the dynamic weight coefficient, + + + =1, default α=0.4, β=0.3, γ=0.2, δ=0.1, is the instruction retention time (unit: ms), τ is the system time constant (configurable, default 5000ms), is the current amount of available resources (normalized value 0-1), is the total resource (normalized to 1.0), is the jth semantic feature value (normalized to 0-1), is the feature weight (dynamically adjusted through online learning), is the safety verification result (0-1), , =2.5, =0.6 is an empirical parameter.

[0111] Furthermore, after calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority, the system further includes: Dynamically adjust weight parameters based on reinforcement learning mechanism 、 、 , the weight parameters are updated at predetermined intervals, and the weight parameters are updated according to the following formula: in, is the learning rate (default 0.05), is the actual average response delay, is the target delay (configurable), is the current resource utilization, is the target resource utilization (default 0.7), is the accuracy of the most recent N priority decisions.

[0112] Furthermore, a multi-level resource preemption strategy is triggered according to the dynamic priority, including: Set full preemption threshold and partial preemption threshold; When the dynamic priority is greater than the full preemption threshold, starting the resource full preemption mode; When the dynamic priority is greater than the partial preemption threshold and less than or equal to the full preemption threshold, starting the resource partial preemption mode; When the dynamic priority is less than or equal to the partial preemption threshold, the resource negotiation preemption mode is started.

[0113] The embodiments of this specification provide a traffic centralized controller resource scheduling system. Through the automatic implementation process of protocol identification, semantic analysis, dynamic priority, and resource preemption, it calculates and optimizes the priority of traffic control instructions by identifying communication protocols and transmission data characteristics, and performs resource allocation and complex balanced scheduling based on priority. Through multi-level and multi-dimensional instruction analysis and dynamic adaptive resource allocation strategies, it realizes dynamic resource scheduling of massive heterogeneous traffic instructions, ensures real-time response of high-priority instructions, and solves the three major technical problems of rigid static resource allocation, rough priority determination, and poor protocol adaptability. In the traffic evacuation scenario of 10,000 people, the response speed of high-priority instructions is improved by 300%.

[0114] Based on the same inventive concept, an embodiment of this specification further provides an electronic device, including at least one processor and a memory, wherein the memory stores a program and is configured to execute the following steps by the at least one processor: Perform protocol recognition on the input command data packet to extract multi-dimensional features; Using a flow reassembly algorithm based on a sliding window to restore the instruction data packet into a complete traffic control instruction; parsing the semantics of the traffic control instruction and determining an initial priority of the traffic control instruction based on the multi-dimensional features; Calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority; A multi-level resource preemption strategy is triggered according to the dynamic priority.

[0115] Among them, other functions of the processor can also refer to the contents recorded in the above embodiments, which will not be repeated here.

[0116] Based on the same inventive concept, an embodiment of this specification further provides a computer-readable storage medium, including a program for use in conjunction with an electronic device, which can be executed by a processor to perform the following steps: Perform protocol recognition on the input command data packet to extract multi-dimensional features; Using a flow reassembly algorithm based on a sliding window to restore the instruction data packet into a complete traffic control instruction; parsing the semantics of the traffic control instruction and determining an initial priority of the traffic control instruction based on the multi-dimensional features; Calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority; A multi-level resource preemption strategy is triggered according to the dynamic priority.

[0117] Among them, other functions of the processor can also refer to the contents recorded in the above embodiments, which will not be repeated here.

[0118] like Figure 5 As shown, the embodiment of this specification also provides a structural schematic diagram of a computer storage medium.

[0119] The systems, devices, modules, or units described in the above embodiments may be implemented by computer chips or entities, or by products having certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0120] For the convenience of description, the above devices are described as being divided into various units according to their functions. Of course, when implementing this application, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0121] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0122] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0123] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0124] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0125] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0126] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0127] Computer-readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tape storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media (transitory media), such as modulated data signals and carrier waves.

[0128] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of additional identical elements in the process, method, commodity, or apparatus that includes the element.

[0129] The present application may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communications network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0130] The various embodiments in this specification are described in a progressive manner. Similar parts between the various embodiments can be referred to in conjunction with each other. Each embodiment focuses on the differences between the other embodiments. In particular, the system embodiments are generally similar to the method embodiments, so the description is relatively simple. For relevant parts, refer to the description of the method embodiments.

[0131] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should all be included within the scope of the claims of the present application.

Claims

1. A traffic centralized controller resource scheduling method, characterized in that: The traffic centralized controller resource scheduling method includes: Perform protocol recognition on the input command data packet to extract multi-dimensional features; Using a flow reassembly algorithm based on a sliding window to restore the instruction data packet into a complete traffic control instruction; parsing the semantics of the traffic control instruction and determining an initial priority of the traffic control instruction based on the multi-dimensional features; Calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority; A multi-level resource preemption strategy is triggered according to the dynamic priority.

2. The method according to claim 1, wherein Perform protocol recognition on the input command data packets to extract multi-dimensional features, including: Performing protocol identification on the instruction data packet using an improved Wu-Manber algorithm, wherein the improved Wu-Manber algorithm uses an FNV-1a rolling hash function to calculate a hash value of a pattern string and skips invalid characters based on a dynamic skip strategy; Extract multi-dimensional features from the identified protocol packets.

3. The method according to claim 1, wherein A sliding window-based stream reassembly algorithm is used to restore the instruction data packet into a complete traffic control instruction, including: A flow reassembly algorithm based on a sliding window is adopted to reassemble the fragmented instruction data packets into a complete data stream according to the sequence number and confirmation number of the instruction data packet, and restore the complete traffic control instruction.

4. The method according to claim 1, wherein Parsing the semantics of the traffic control instruction and determining the initial priority of the traffic control instruction based on the multi-dimensional features, including: Performing correlation analysis based on the context of the traffic control instructions to obtain temporal relationships, causal relationships, and semantic relationships between the traffic control instructions; Parsing the semantics of the traffic control instruction to obtain a parsing result; The priority of the traffic control instruction is pre-classified according to the analysis result and the multi-dimensional feature to obtain the initial priority.

5. The method according to claim 4, wherein Pre-classifying the priority of the traffic control instruction according to the analysis result and the multi-dimensional features to obtain the initial priority includes: The initial priority of the traffic control instruction is calculated according to the following formula: in, is the protocol base weight, is the packet feature analysis score, is the urgency of semantic content, a, b, c are their corresponding weight coefficients, which are preset according to experimental experience, a+b+c=1.

6. The method according to claim 5, wherein Calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority includes: The dynamic priority is calculated using a four-dimensional dynamic priority model, wherein the four-dimensional dynamic priority model is calculated according to the following formula: in, 、 、 、 is the dynamic weight coefficient, + + + =1; is the instruction retention time, in ms; τ is the system time constant; is the current amount of available resources; is the total resource volume; is the jth semantic feature value, normalized to 0-1; is the feature weight; To verify the safety results; .

7. The method according to claim 6, wherein After calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority, the method further includes: Dynamically adjust weight parameters based on reinforcement learning mechanism 、 、 , the weight parameters are updated at predetermined intervals, and the weight parameters are updated according to the following formula: in, is the learning rate, is the actual average response delay, For the target delay, is the current resource utilization, is the target resource utilization, is the accuracy of the most recent N priority decisions.

8. The method according to claim 1, wherein The multi-level resource preemption strategy is triggered according to the dynamic priority, including: Set full preemption threshold and partial preemption threshold; When the dynamic priority is greater than the full preemption threshold, starting the resource full preemption mode; When the dynamic priority is greater than the partial preemption threshold and less than or equal to the full preemption threshold, starting the resource partial preemption mode; When the dynamic priority is less than or equal to the partial preemption threshold, the resource negotiation preemption mode is started.

9. A traffic centralized controller resource scheduling system, characterized in that: The traffic centralized controller resource scheduling system includes: The recognition module performs protocol recognition on the input command data packet to extract multi-dimensional features; A restoration module, which uses a flow reassembly algorithm based on a sliding window to restore the instruction data packet into a complete traffic control instruction; a parsing module, which parses the semantics of the traffic control instruction and determines the initial priority of the traffic control instruction based on the multi-dimensional features; a calculation module, calculating a dynamic priority corresponding to the traffic control instruction according to the initial priority; The preemption module triggers a multi-level resource preemption strategy according to the dynamic priority.

10. A computer storage medium comprising a program for use in conjunction with an electronic device, the program being executable by a processor to perform the following steps: Perform protocol recognition on the input command data packet to extract multi-dimensional features; Using a flow reassembly algorithm based on a sliding window to restore the instruction data packet into a complete traffic control instruction; parsing the semantics of the traffic control instruction and determining an initial priority of the traffic control instruction based on the multi-dimensional features; Calculating the dynamic priority corresponding to the traffic control instruction according to the initial priority; A multi-level resource preemption strategy is triggered according to the dynamic priority.

Images