Time sensitive network cycle-aware scheduling method and system for cloud computer interlocking

Abstract

The application provides a time-sensitive network cycle-aware scheduling method and system for cloud computer interlocking. First, the application constructs a cloud computer interlocking time-sensitive network topology model and a time-triggered flow model, calculates a super cycle based on a cycle set, and maps a time slot to a congruence remainder class under different cycles; second, the application pre-constructs a cycle-aware pre-allocated time slot set according to the congruence relationship between multi-cycle services, and realizes pre-computation of cross-cycle update through subgroup relationship; then, the application sorts the time-triggered flow, preferentially selects a feasible candidate time slot belonging to the current cycle pre-allocated set on each fixed routing link, and updates the gate control list and each cycle pre-allocated set after successful allocation; finally, the application outputs the gate control list, sending offset, queue, routing and delay results. The application can reduce cross-cycle implicit conflicts, maintain the regularity of the remaining cycle time slot space, and improve the schedulability in the high-density time-triggered flow scenario.

Description

Technical Field

[0001] This invention belongs to the fields of rail transit safety communication, time-sensitive networks, and deterministic network scheduling, and particularly relates to a time-sensitive network periodic sensing scheduling method and system for cloud computer interlocking. Background Technology

[0002] Computerized interlocking systems are safety-critical core systems within railway signaling systems. They continuously process section occupancy, turnout status, route requests, and safety constraints, and output control commands within strict time limits. As railway signaling systems evolve towards cloudification, centralization, and platformization, cloud-based computer interlocking is gradually becoming an important architecture. This architecture deploys interlocking logic, replication nodes, voting nodes, health monitoring, and platform management functions on cloud or data center resources, thereby achieving unified resource management, elastic scaling, and simplified maintenance.

[0003] In cloud computing interlocking architectures, the determinism of network communication becomes a crucial factor affecting security timing. Periodic time-triggered services continuously arise between interlocking application instances, replication nodes, voting nodes, and monitoring and management nodes. These services typically have fixed periods, defined deadlines, low jitter, and bounded latency requirements. Time-Sensitive Networking (TSN) provides deterministic transmission capabilities through time-aware shaping and gating control lists, making it suitable for carrying security-critical periodic services in cloud computing interlocking.

[0004] However, existing time-sensitive network scheduling methods typically treat feasible time slots as equivalent resources, focusing primarily on whether the current flow can find a conflict-free transmission time slot, without fully considering the periodic structure of cloud computing interlocked services. When multiple time-triggered flows (TT) coexist, some time slot allocations feasible for the current flow can disrupt the periodic repetition space required by subsequent flows, causing the remaining time slots to be fragmented into multiple short pieces. Although idle time slots may still exist on the normal timeline, these time slots cannot form a continuous congruent class that satisfies the requirements of subsequent periodic flows, thus rendering subsequent flows unschedulable.

[0005] For cloud computer interlocking systems, schedulability is not only an offline design issue, but also relates to configuration boundary judgment during system expansion, functional addition, monitoring service access, and cloud node expansion. If the scheduling method cannot maintain the regularity of the remaining cycle time slot space, it will prematurely enter an unschedulable state in high-density, multi-cycle business scenarios, limiting system scale and engineering deployment capabilities. Therefore, there is an urgent need for a time-sensitive network scheduling method and system that can perceive the business cycle structure, reduce implicit cross-cycle conflicts, and maintain the continuity of the remaining cycle space. Summary of the Invention

[0006] To address the problems of existing cloud computer interlocking time-sensitive network scheduling methods, such as decreased schedulability of high-density, multi-period time-triggered services, difficulty in controlling implicit conflicts across periods, fragmentation of remaining time slot space, and insufficient efficiency in constructing gating control lists, this invention provides a period-aware scheduling method and system for cloud computer interlocking time-sensitive networks. This method addresses the deterministic communication requirements of periodic safety-critical services in cloud computer interlocking scenarios. It utilizes the periodic structure of time-triggered flows and the congruence relationships between different periods to perform period-aware modeling of time slot resources within the hyper-period. During list scheduling, it prioritizes candidate time slots that maintain the regularity of the periodic structure, thereby reducing implicit conflicts between multi-period services and improving the schedulability of high-density time-triggered flows in time-sensitive networks.

[0007] The present invention adopts the following technical solution.

[0008] A time-sensitive network period-aware scheduling method for cloud computer interlocking includes the following steps.

[0009] Step 1: Construct a cloud computer interlocking time-sensitive network topology model and a time-triggered flow model, representing the cloud computer interlocking time-sensitive network as a directed graph. ,in, Represents a set of time-sensitive network devices. This represents a set of directed physical links, and the time-sensitive network devices include bridging devices and end systems; Represent the time-triggered stream set as ,in This indicates the number of time-triggered streams. Each time-triggered stream... Using a six-tuple model: in, and These represent the sending end and the receiving end, respectively. This indicates the number of data bytes generated in each cycle. Indicates period, Indicates the deadline and satisfies , This indicates a pre-determined fixed route; Time-triggered stream The number of consecutive time slots occupied by a single transmission on a link is denoted as . Its upper bound is denoted as ; A set of scheduling instances is abstracted into a configuration profile: in, Represents network topology. Indicates the number of time-triggered streams. Represents a periodic set. Indicates the deadline rules, This indicates network and service configuration parameters other than the period and deadline; Step 2: Based on the period set in the time-triggered stream set Determine the supercycle And within the supercycle, the time slots are mapped to congruential remainder classes according to different periods; Step 3: Based on the congruence relationship between different periods, pre-construct a period-aware PriA time slot set, and pre-calculate the update baseline of the cross-period PriA time slot set based on the subgroup relationship between different periods; Step 4: Sort the time-triggered streams according to the deadline, period, path length, and transmission length, and allocate time slots link by link along the fixed route of each time-triggered stream; on each link, prioritize the selection of feasible candidate time slots belonging to the current period's PriA time slot set; Step 5: After each successful time slot allocation, update the gating control list on the corresponding link and the PriA time slot set for each period; if all time-triggered streams have been allocated, output the gating control list, transmission offset, queue, routing, and latency results; if there are time-triggered streams that cannot be allocated, output the result that the current scheduling instance is unschedulable.

[0010] Specifically, in step 2, for the time-triggered stream cycle set Calculate the supercycle , which is the least common multiple of all time-triggered flow cycles. The set of time slots within a supercycle is . .

[0011] For any period The congruence remainder class is defined as follows: Define time-triggered stream Number of occurrences within a supercycle for: make ,and If time triggers the flow In the link The starting time slot index within the first cycle is Then time-triggered stream In the link The set of time slots allocated in the previous supercycle is: The In the model In a sense, corresponding to from arrive of A class of continuous congruent remainders.

[0012] Furthermore, if the model The class of distributable remainders is divided into: A maximal continuous interval, the lengths of which are respectively Then time-triggered stream In the link The number of feasible initial remainder classes is represented as follows: like ,but And determine the time-triggered stream In the link It cannot be scheduled.

[0013] Specifically, in step 3, a set of period-aware PriA (PriorAllocated) time slots is constructed based on the congruence relationship between different periods.

[0014] The PriA time slot set is represented as follows: For any , The period is represented as The set of PriA time slots corresponding to the unscheduled time-triggered streams.

[0015] make Indicates time-triggered stream Given the set of remainder classes corresponding to the allocated time slots under different periods, then: in, Indicates in the model The following contains time-triggered streams The set of remainder classes of allocated time slots; in Under the given conditions, Represented as: Time-triggered stream After completing the time slot allocation, the period is... The corresponding PriA time slot set is updated: in, Indicates a time-triggered stream Add to The PriA time slot set in the middle.

[0016] For any The new PriA time slot subset is as follows: when At that time, the newly added PriA time slot subset is: Furthermore, by defining , and This transforms the congruence relationships between different periods into pre-computable subgroup relationships, thereby reducing repeated traversals for cross-period updates.

[0017] For any period Define the module Remainder-class cyclic group: For any two periods Define the period Reference time slot set: and by period In the model Spatially induced congruence subgroups: in, yes subgroups of, and yes The generator.

[0018] During the network initialization phase, let ,mold The occupied remainder class is simplified to: definition As The baseline set; for The subset of the baseline set is: when At that time, a subset of the baseline set is: in, This represents the smallest time slot index in the corresponding remainder class.

[0019] Time-triggered stream After the actual starting time slot is determined, an offset is applied based on the baseline set; for any ,have: Specifically, in step 4, the time-triggered streams are sorted in multiple levels according to the comparison order of deadline, period, path length and transmission length, and periodic-aware time slots are allocated link by link along the fixed route of each time-triggered stream.

[0020] The sorting rules are as follows: in, This indicates that it has a higher priority.

[0021] For time-triggered streams and links First, calculate the set of feasible candidate time slots. Then calculate: in, Indicates time-triggered stream In the link The set of candidate start time slots that simultaneously satisfy both feasibility constraints and periodic awareness priority selection conditions; Indicates time-triggered stream In the link The set of all feasible candidate start time slots that satisfy the scheduling constraints; Indicates the period A pre-maintained set of periodically aware pre-allocated time slots is used to guide the current flow to preferentially select time slots that have a congruent alignment with the allocated time-triggered flow; the symbol " "" indicates a set assignment and update operation.

[0022] like If the set is not empty, the earliest starting time slot will be selected from this set first. like It is empty, but If the value is not empty, it means that there is no feasible time slot on the current link that satisfies the periodicity-aware priority condition. In this case, the process degenerates into selecting the earliest starting time slot from the set of all feasible candidate time slots. like If empty, it indicates a time-triggered stream. Unable to be on the link If the conflict-free transmission and latency constraints are satisfied, the current scheduling instance is determined to be unschedulable.

[0023] Specifically, in step 5, after each successful time slot allocation, the PriA time slot set for the corresponding period is updated according to the following formula: in, This is a deadline-aware filtering function. The default filtering rule is: If all time-triggered streams have been allocated, the output will include the gating control list, send offset, queue, routing, and delay results.

[0024] Specifically, the gating control list generation process includes: assigning the starting time slot of each time-triggered stream to each link. Add to the gating control list and trigger the flow period based on time. In supercycle The inner loop repeats, eventually returning GCL (gated list), Delay, Offset, Queue, and Route.

[0025] This invention also provides a time-sensitive network periodic awareness scheduling system for cloud computer interlocking, used to implement the above-mentioned time-sensitive network periodic awareness scheduling method for cloud computer interlocking, including a topology and flow model construction module, a hyperperiodic and congruent partitioning module, a periodic relationship pre-calculation module, a PriA time slot set maintenance module, a time-triggered flow sorting module, a periodic awareness list scheduling module, and a gating control list generation module.

[0026] The topology and flow model construction module is used to receive the cloud computer interlocking time-sensitive network topology, link rate, time slot granularity, queue number and time-triggered flow configuration, and to establish a directed graph model and a time-triggered flow model.

[0027] The superperiod and congruence partitioning module is used to determine the superperiod based on the period set and map the time slots within the superperiod to congruence remainder classes under different periods.

[0028] The periodic relationship pre-calculation module is used to construct modular addition groups and congruent subgroups based on the greatest common divisor relationship, and to pre-generate congruent mapping relationships between different periods.

[0029] The PriA time slot set maintenance module is used to maintain the PriA time slot set corresponding to each period, and to update the PriA time slot set after the time-triggered stream is successfully allocated.

[0030] The time-triggered stream sorting module is used to determine the scheduling order based on the deadline, period, path length, and transmission length.

[0031] The periodic awareness list scheduling module is used to calculate feasible candidate time slots link by link along the fixed route of the time-triggered flow, and preferentially selects feasible candidate time slots belonging to the current period PriA time slot set.

[0032] The gating control list generation module is used to output gating control lists, sending offsets, queues, routing, and latency results based on the scheduling results.

[0033] The beneficial effects of this invention are as follows.

[0034] This invention maps the time-triggered flow in the cloud computer interlocked time-sensitive network to the congruent remainder class space within the super-period, transforming the repeated transmission constraint of periodic services into a time slot occupancy problem on the congruent structure, enabling the scheduling method to explicitly characterize the structured occupancy relationship of periodic services on time slot resources.

[0035] This invention constructs a modular addition group and a congruent subgroup based on the greatest common divisor relationship between different cycles, and pre-calculates the update baseline of the cross-cycle PriA time slot set based on the subgroup relationship, enabling the scheduling process to quickly identify the congruent alignment relationship between services in different cycles and reduce the overhead of repeated calculations.

[0036] This invention constructs a period-aware PriA time slot set, enabling current services to prioritize feasible candidate time slots that are congruent to existing occupancy, reducing cross-period implicit conflicts caused by local greedy allocation, suppressing the fragmentation of remaining time slot space, and thus improving schedulability in high-density, multi-period time-triggered service scenarios.

[0037] This invention sorts time-triggered streams according to deadline, period, path length, and transmission length, giving priority to services with smaller time margins, more repetitions, stronger path constraints, and greater demand for continuous time slots, which helps improve the overall scheduling success rate.

[0038] This invention can output gating control lists, transmission offsets, queues, routing, and latency results. It is suitable for deterministic communication configurations in time-sensitive networks based on time-aware shaping mechanisms and can provide low-latency, bounded-latency, and conflict-free periodic transmission guarantees for safety-critical services in cloud computer interlocking systems. Attached Figure Description

[0039] Figure 1 This is a schematic diagram of the overall process of the periodic awareness scheduling method of the present invention; Figure 2 This is a schematic diagram of the cloud computer interlocking network topology according to an embodiment of the present invention; Figure 3 This is a comparison of schedulability under a single-cycle configuration in this embodiment of the invention; Figure 4 The results show the schedulability comparison under the harmonic period configuration of the present invention. Figure 5This is a comparison of the schedulability under the non-harmonic periodic configuration of the present invention. Figure 6 This is a general block diagram of the periodic sensing scheduling system according to an embodiment of the present invention. Detailed Implementation

[0040] The time-sensitive network periodic awareness scheduling method and system for cloud computer interlocking provided by the present invention will be further described below with reference to specific embodiments and accompanying drawings. It should be understood that the following embodiments are only used to explain the technical solution of the present invention and are not intended to limit the scope of protection of the present invention. Equivalent substitutions, parameter adjustments, or module combinations made by those skilled in the art based on the content of this specification without departing from the technical concept of the present invention should all fall within the scope of protection of the present invention.

[0041] This embodiment addresses safety-critical deterministic communication services in a cloud-based computer interlocking system. The cloud-based computer interlocking system deploys some interlocking logic, status acquisition, control command issuance, redundant instance synchronization, health monitoring, and platform management functions in the cloud or edge computing nodes. These services are typically transmitted in time-sensitive networks as periodic time-triggered streams, requiring bounded end-to-end latency, low jitter, and no link collisions. Therefore, this embodiment employs a time-aware shaping mechanism to achieve deterministic transmission of time-triggered streams by generating a gating control list for each link.

[0042] Figure 1 The overall flow of the periodic-aware scheduling method of the present invention is shown. The flow includes the construction of a cloud computer interlocking time-sensitive network topology model and a time-triggered flow model, the classification of hyperperiodic and congruent remainder classes, the construction of a periodic-aware PriA time slot set, periodic-aware list scheduling, and the output of the gating control list and scheduling results.

[0043] In this embodiment, the cloud computer interlocking time-sensitive network consists of switching nodes, terminal nodes, and directed communication links. Switching nodes forward time-triggered services and other non-critical services; terminal nodes carry cloud computer interlocking application instances, input / output control units, redundant interlocking instances, status monitoring units, voting units, or edge computing nodes. All time-sensitive network devices maintain a unified time base through a high-precision time synchronization mechanism, enabling gating operations to be executed according to unified time slot boundaries.

[0044] Represent the cloud computer interlocking time-sensitive network as a directed graph: in, Represents a set of time-sensitive network devices. This represents the set of directed physical links. For any directed link... Data from device Send to device All physical links are full-duplex. Each egress port maintains multiple transmission queues, and the central controller is responsible for calculating and issuing the gating control list. All devices perform gating operations based on high-precision time synchronization.

[0045] Figure 2 The cloud computer interlocking network topology used in this embodiment is illustrated. This topology is organized in a modular manner, with the backbone employing a ring-like interconnection structure. Each repeating subnet corresponds to a computing group, and within each group, a local access unit corresponds to a computing node. Application endpoints under each local access unit correspond to computing elements. This structure matches the centralized deployment and redundant communication requirements of cloud computer interlocking services.

[0046] In this embodiment, the periodic safety-critical communication services in the cloud computer interlocking are modeled as a set of time-triggered flows: in, This indicates the number of time-triggered streams. Each time-triggered stream... Using a six-tuple model: in, and These represent time-triggered events. The sending and receiving ends, Indicates time-triggered stream Number of bytes generated in each cycle Indicates time-triggered stream The cycle, Indicates the deadline, and satisfies , Indicates time-triggered stream Fixed routes on a given topology. Time-triggered flow cycle set. .

[0047] To provide a unified description of network and service combinations across different experimental and deployment scenarios, this embodiment further represents a set of scheduling instances as a configuration profile: in, Represents network topology. Indicates the number of time-triggered streams. Represents a periodic set. Indicates the deadline rules, Other configuration parameters include load parameters, number of bridged devices, and link settings.

[0048] All time slot calculations in this paper are based on the time slot index. For each link, the time slot number starts from... Begin. The set of time slots within a supercycle is: The gating control list defines the gating state within one supercycle and repeats it periodically in subsequent supercycles.

[0049] For any period Supercycle The time slots within the module are according to the model The remainder is divided into There are three distinct classes of congruence remainders. Specifically, a class of congruence remainders is represented as: in, To indicate congruence, that is Can be an integer Divisibility. This partition is used to describe the repetitive occupancy structure of a periodic stream within a super-period. For time-triggered streams of the same period, their periodic transmission instances correspond to the same remainder class in the modulo-period sense. Therefore, their repetitive transmission constraints can be described by the remainder class rather than by individual time slots.

[0050] Furthermore, record Time-triggered stream The number of consecutive time slots required to complete one transmission on a single link. (Note: The original text contains some formatting errors and inconsistencies. A more accurate translation would require the full context.) Time-triggered stream In the link The index of the first transmission slot in the first cycle. and Once determined, the time-triggered flow is in the link. The transmission time slots of subsequent periodic instances are arranged according to the period. Repeat. Time-triggered stream. It forwards along its fixed route link by link, so the flow triggered at the same time will be allocated a transmission start time slot on multiple links along its route.

[0051] In this embodiment, for simplicity, it is assumed that all time-sensitive network links have the same link transmission rate and the same time slot granularity. Therefore, a single transmission of the same time-triggered stream on different links occupies the same number of consecutive time slots. can be It is uniquely determined, and its upper bound is denoted as In other implementations, if the transmission rates of different links are different, time-triggered streams can be used. In the link The number of consecutive time slots occupied by the last transmission is denoted as And calculate according to the transmission rate of the link.

[0052] Define time-triggered stream The number of occurrences within a supercycle is ,as follows: make ,in Within a supercycle, time-triggered flow In the link The set of time slots assigned above Represented as: From the above formula, we can see that All time slots After taking the mold, it corresponds exactly to A continuous remainder class, i.e. from arrive Therefore, the stream is triggered at any time. In the link A necessary condition for schedulability is that, in the module... The situation exists There are several consecutive remainder classes, and all congruent time slots in these remainder classes can be allocated to... .

[0053] If the model The class of distributable remainders is divided into: A maximal continuous interval, the lengths of which are respectively Then time-triggered stream In the link Number of feasible initial remainder classes for: like ,but Time-triggered stream In the link The above is not schedulable. If a segment of length is... The continuous distributable remainder intervals are divided into intervals of lengths respectively. and The two sections, of which Then for any ,have: This relationship indicates that further time-slot fragmentation does not increase the number of feasible starting positions. Therefore, the goal of period-aware scheduling is to minimize the spread of new remainder classes and avoid dividing long, contiguous, allocatable intervals into multiple shorter intervals.

[0054] Based on the above analysis, this embodiment introduces the PriA time slot set, where PriA represents the period-aware priority allocation of time slots. For any period PriA time slot set Defined as: the set of candidate starting time slots induced by the congruent mapping relationship between different periods of the already occupied time slots, under the condition that partial time-triggered flow scheduling has been completed. For a period of The subsequent unscheduled time triggers the flow, if it prioritizes... If a feasible candidate time slot is selected, the periodic transmission position of the stream can be aligned congruently with the existing occupied time slots as much as possible, thereby reducing the spread of newly added remainder classes and reserving a more complete periodic space for subsequent time-triggered streams.

[0055] Without causing confusion, the following sets are all for the current link. Define, omitting the link index. Let... This represents the set of time-triggered streams that have completed time slot allocation. For time-triggered streams that have already been scheduled... Its set of allocated time slots on the current link Recorded as: , in, Indicates time-triggered stream In the link The starting time slot index within the first cycle, This indicates the period of the time-triggered stream. This indicates the number of consecutive time slots occupied by the time-triggered stream in a single transmission on a link. This indicates a superperiod determined by the entire set of periods.

[0056] Furthermore, for any period ,Will Mapping to Module The congruent remainder class space yields a time-triggered stream. The allocated time slots in the period The corresponding remainder class set : in, Indicates time slot In the model The class of congruence remainders in the sense of [missing information]. Notation: Then time-triggered stream For the target period The newly added portion of the PriA time slot set is denoted as: Among them, when At that time, there were: when At that time, there were: Therefore, the cycle The corresponding PriA slot set is defined as the union of the sets of newly added PriA slots induced by all scheduled time-triggered flows for this period: The set of PriA slots corresponding to all periods is represented as follows: therefore, It is not a set of already occupied time slots, but rather a set of time slots derived from occupied time slots through congruence relationships, with a period of... The subsequent time-triggered flow preferentially selects the set of candidate start time slots. During the scheduling process, if a candidate start time slot belongs to both the feasible candidate time slot set and the set of candidate start time slots, the subsequent time-triggered flow will be selected based on the set of candidate start time slots. If so, this time slot is selected first to maintain the congruent alignment between multi-cycle time-triggered streams.

[0057] To avoid explicit traversal after each allocation To mitigate the redundant overhead, this embodiment utilizes modular addition groups and subgroup relationships for pre-computation. For any period... Define the module Remainder-type cyclic group : For any two periods Define the period Reference time slot set : and cycle In the model Spatially induced congruence subgroups: in, yes a subgroup of, and yes The generators are thus generated. Therefore, the subgroup relationship between any two cycles can be pre-calculated during the network initialization phase, thereby avoiding repeated traversal of all allocated time slots for each time-triggered stream.

[0058] During the initialization phase, let Then the model The occupied remainder class can be simplified to: Further, define As The baseline set. A subset of this baseline set is represented as: as well as: in, This represents the smallest time slot index in the corresponding remainder class. When the time-triggered stream... Once the actual starting time slot is determined, only an offset needs to be applied to that baseline. For any ,have: Based on the foregoing analysis, PriA slot design, and congruence relation update framework, this embodiment adopts the period-aware list scheduling method PriA-LS. This method first initializes the PriA slot set for each period and pre-calculates the baseline structure; then, it sorts time-triggered flows according to scheduling vulnerability; finally, it incrementally allocates slots along a fixed route for each flow link by link, and updates the PriA slot set after each successful allocation.

[0059] To improve the final scheduling success rate, PriA-LS prioritizes scheduling more vulnerable time-triggered streams. The sorting rule is as follows: The time-triggered streams are sorted in multiple levels based on deadline (ascending order), period (ascending order), path (descending order), and transmission length (descending order), meaning that time-triggered streams with tighter deadlines, shorter periods, longer paths, and larger transmission lengths are allocated priority. This sorting ensures that services more prone to failure receive more structurally complete time slot resources first.

[0060] For sorted time-triggered streams Scheduling is performed link by link along its fixed route. For the current link... First, calculate the set of feasible candidate time slots. .like If not, return "unschedulable". Otherwise, calculate: like Then choose: like Then, backtrack to the set of all feasible candidate time slots and select the earliest one: Time-triggered stream On the link After the allocation is completed, for any Update the PriA slot set according to the following formula: in, The above baseline plus offset formula is used to obtain, This is the deadline-aware filtering function. To avoid adding excessively late time slots to the PriA time slot set, the following default filtering rules are used: in, It is an adjustable parameter. For topology The maximum path length between any two nodes in the array. Represents the set of integers.

[0061] This filtering step preserves PriA time slots that are still meaningful under the deadline constraint, avoiding the problem that candidate time slots, while maintaining congruent alignment, cannot reserve sufficient time margin for subsequent link transmissions. The meaning of this filtering function is: only when the candidate start time slot... Distance-time triggering stream The deadline remains no less than A candidate time slot is allowed to be added to the PriA time slot set only if there is sufficient time margin in the time slot. If it is too late, i.e. the above inequality is not satisfied, then even if the time slot has periodic alignment characteristics in the congruence relation, it will not be added to the PriA time slot set, so as to avoid subsequent scheduling prioritizing time slots that may cause the deadline constraint to fail.

[0062] set up Number of time-triggered streams For different number of cycles, Maximum route length This is an upper bound on the number of candidate start time slots that need to be checked on a single link. Therefore, the overall time complexity of PriA-LS is: This complexity is acceptable for cloud computer interlocking scenarios. On the one hand, PriA-LS avoids the global search overhead of solver methods; on the other hand, congruence relation simplification moves most of the cross-cycle computations forward to the initialization phase, and online updates only require baseline lookup and offset operations. Therefore, this method is suitable for repeatedly conducting schedulability assessments and configuration boundary analyses under different cloud computer interlocking service configurations.

[0063] To verify the feasibility and effectiveness of the proposed method (PriA_LS) in cloud computer interlocking time-sensitive networks, this embodiment performs simulation verification on a cloud computer interlocking topology. The simulation topology uses a small-scale cloud computer interlocking time-sensitive network with 16 bridging devices, and the number of time-triggered flows gradually increases from 10 to 220. Comparison methods include SMT_NW, SMT_WA, LS, LS_TB, and LS_PL. The evaluation metric is the schedulability rate, i.e., the proportion of instances that can successfully complete the scheduling of all time-triggered flows under the same configuration.

[0064] Regarding the periodic configuration, this embodiment employs three types of period sets. The first type is a single-period configuration. The milliseconds (ms) are used to verify the impact of time slot allocation strategies on schedulability in scenarios with weak periodic heterogeneity. The second category is harmonic period configuration. The milliseconds (ms) indicate a clear divisibility relationship between different periods, used to verify the effect of the period-aware congruent alignment mechanism on multi-period services. The third category is non-harmonic period configuration. The alignment between different periods is weaker, and the time slot fragmentation problem is more obvious. This is used to verify the robustness of the present invention in complex heterogeneous periodic business scenarios.

[0065] Figure 3 The results of schedulability comparison under a single-cycle configuration are shown. In this configuration, all time-triggered streams have the same cycle structure, making the scheduling problem relatively simple. As the number of time-triggered streams increases, the schedulability of all compared methods decreases, but the method of this invention decreases more slowly and maintains a higher schedulability over a larger range of time-triggered streams. This demonstrates that even in scenarios where cycle heterogeneity is not significant, cycle-aware candidate time slot selection still helps protect the remaining available time slot space.

[0066] Figure 4 The results show a comparison of schedulability under a harmonic period configuration. In this configuration, different periods are divisible, and multiple periodic flows compete for time slot resources with congruent alignment. Traditional list-based scheduling methods primarily focus on the immediate feasibility of the current link, easily disrupting the continuous congruent remainder class required by subsequent periodic flows. In contrast, the method of this invention maintains a period-aware PriA time slot set, allowing time-triggered flows to preferentially occupy candidate time slots congruently aligned with existing allocations, thereby maintaining a higher schedulability in medium-to-high load regions.

[0067] Figure 5The results show a comparison of schedulability under a non-harmonic periodic configuration. In this configuration, the greatest common divisor between different periods is small, the congruence alignment is weak, cross-period implicit conflicts are more likely to occur, and time slot fragmentation is most severe. Most comparative methods show a rapid decline in schedulability after the increase in service load, while the method of this invention can still maintain a significantly higher schedulability over a wider range of time-triggered flow numbers. This result shows that the present invention can maintain a more regular remaining period time slot structure in weakly aligned heterogeneous periodic scenarios, thereby expanding the deployable area that meets the requirements of deterministic communication.

[0068] comprehensive Figures 3 to 5 It can be seen that the method of this invention can achieve a high schedulability rate in three typical scenarios: single-cycle, harmonic-cycle, and non-harmonic-cycle. This is because the invention does not treat time slot resources as independent discrete points, but rather uses period sets, superperiodicity, congruence remainder classes, and subgroup relationships to characterize the structural connections between multi-cycle time-triggered flows. Based on this, the scheduler guides the selection of candidate time slots through the period-aware PriA time slot set, reducing the disruption of the global periodic structure caused by local greedy allocation.

[0069] This invention also provides a time-sensitive network period-aware scheduling system for cloud computer interlocking, such as... Figure 6 As shown, a time-sensitive network period-aware scheduling method for cloud computer interlocking is used to implement the above-mentioned method. The system includes a topology and flow model construction module, a hyperperiodic and congruent partitioning module, a periodic relationship pre-calculation module, a PriA time slot set maintenance module, a time-triggered flow sorting module, a period-aware list scheduling module, and a gating control list generation module.

[0070] The topology and flow model construction module is used to receive the cloud computer interlocking time-sensitive network topology, link rate, time slot granularity, queue number and time-triggered flow configuration, and to establish a directed graph model and a time-triggered flow model.

[0071] The superperiod and congruence partitioning module is used to determine the superperiod based on the period set and map the time slots within the superperiod to congruence remainder classes under different periods.

[0072] The periodic relationship pre-calculation module is used to construct modular addition groups and congruent subgroups based on the greatest common divisor relationship, and to pre-generate congruent mapping relationships between different periods.

[0073] The PriA time slot set maintenance module is used to maintain the PriA time slot set corresponding to each period, and to update the PriA time slot set after the time-triggered stream is successfully allocated.

[0074] The time-triggered stream sorting module is used to determine the scheduling order based on the deadline, period, path length, and transmission length.

[0075] The periodic awareness list scheduling module is used to calculate feasible candidate time slots link by link along the fixed route of the time-triggered flow, and preferentially selects feasible candidate time slots belonging to the current period PriA time slot set.

[0076] The gating control list generation module is used to output gating control lists, sending offsets, queues, routing, and latency results based on the scheduling results.

[0077] In engineering deployment, the above system can run on a centralized network configuration controller. The centralized network configuration controller generates a gating list based on the cloud computer interlocking service configuration and network topology, and distributes the gating list to the corresponding switching devices. Each switching device executes the gating list according to a unified time synchronization benchmark, thereby providing deterministic transmission guarantees for cloud computer interlocking security-critical services.

[0078] The above description is merely a preferred embodiment of the present invention. The present invention is not limited to the above-described embodiments. Any equivalent transformations or substitutions made to the network size, period set, queue number, time slot granularity, sorting weight, filtering parameters, or gating list generation method within the scope of the technical concepts and claims disclosed in this invention shall fall within the protection scope of this invention.

Claims

1. A time-sensitive network period-aware scheduling method for cloud computer interlocking, characterized in that, Includes the following steps: Step 1: Construct a cloud computer interlocking time-sensitive network topology model and a time-triggered flow model, representing the cloud computer interlocking time-sensitive network as a directed graph. ,in, Represents a set of time-sensitive network devices. This represents a set of directed physical links, and the time-sensitive network devices include bridging devices and end systems; Represent the time-triggered stream set as ,in Indicates the number of time-triggered streams; each time-triggered stream Using a six-tuple model: in, and These represent the sending end and the receiving end, respectively. This indicates the number of data bytes generated in each cycle. Indicates period, Indicates the deadline and satisfies , This indicates a pre-determined fixed route; Time-triggered stream The number of consecutive time slots occupied by a single transmission on a link is denoted as . Its upper bound is denoted as ; A set of scheduling instances is abstracted into a configuration profile: in, Represents network topology. Indicates the number of time-triggered streams. Represents a periodic set. Indicates the deadline rules, This indicates network and service configuration parameters other than the period and deadline; Step 2: Based on the period set in the time-triggered stream set Determine the supercycle And within the supercycle, the time slots are mapped to congruential remainder classes according to different periods; Step 3: Based on the congruence relationship between different periods, pre-construct a period-aware PriA time slot set, and pre-calculate the update baseline of the cross-period PriA time slot set based on the subgroup relationship between different periods; Step 4: Sort the time-triggered streams according to the deadline, period, path length, and transmission length, and allocate time slots link by link along the fixed route of each time-triggered stream; on each link, prioritize the selection of feasible candidate time slots belonging to the current period's PriA time slot set; Step 5: After each successful time slot allocation, update the gating control list on the corresponding link and the PriA time slot set for each period; if all time-triggered streams have been allocated, output the gating control list, transmission offset, queue, routing, and latency results; if there are time-triggered streams that cannot be allocated, output the result that the current scheduling instance is unschedulable. In step 2, for the time-triggered stream cycle set Calculate the supercycle , is the least common multiple of all time-triggered flow cycles; the set of time slots within a supercycle is . ; For any period The congruence remainder class is defined as follows: Define time-triggered stream Number of occurrences within a supercycle for: make ,and If time triggers the flow In the link The starting time slot index within the first cycle is Then time-triggered stream In the link The set of time slots allocated in the previous supercycle is: The In the model In a sense, corresponding to from arrive of A class of consecutive congruent remainders; In step 3, a set of PriA time slots for periodic sensing is constructed based on the congruence relationship between different periods; The PriA time slot set is represented as follows: For any , The period is represented as The set of PriA time slots corresponding to the unscheduled time-triggered stream; make Indicates time-triggered stream Given the set of remainder classes corresponding to the allocated time slots under different periods, then: in, Indicates in the model The following contains time-triggered streams The set of remainder classes of allocated time slots; in Under the given conditions, Represented as: Time-triggered stream After completing the time slot allocation, the period is... The corresponding PriA time slot set is updated: in, Indicates a time-triggered stream Add to The PriA time slot set in the middle; For any The new PriA time slot subset is as follows: when At that time, the newly added PriA time slot subset is: ； By definition , and This transforms the congruence relationships between different periods into pre-computable subgroup relationships, thereby reducing repeated traversals for cross-period updates. For any period Define the module Remainder-class cyclic group: For any two periods Define the period Reference time slot set: and by period In the model Spatially induced congruence subgroups: in, yes subgroups of, and yes Generators; During the network initialization phase, let ,mold The occupied remainder class is simplified to: definition As The baseline set; for The subset of the baseline set is: when At that time, a subset of the baseline set is: in, This represents the smallest time slot index in the corresponding remainder class; Time-triggered stream After the actual starting time slot is determined, an offset is applied based on the baseline set; for any ,have: 。 2. The method according to claim 1, characterized in that, If the model The class of distributable remainders is divided into: A maximal continuous interval, the lengths of which are respectively Then time-triggered stream In the link The number of feasible initial remainder classes is represented as follows: like ,but And determine the time-triggered stream In the link It cannot be scheduled.

3. The method according to claim 1, characterized in that, In step 4, the time-triggered streams are sorted in multiple levels according to the comparison order of deadline, period, path length and transmission length, and periodic-aware time slots are allocated link by link along the fixed route of each time-triggered stream. The sorting rules are as follows: in, This indicates higher priority; For time-triggered streams and links First, calculate the set of feasible candidate time slots. Then calculate: in, Indicates time-triggered stream In the link The set of candidate start time slots that simultaneously satisfy both feasibility constraints and periodic awareness priority selection conditions; Indicates time-triggered stream In the link The set of all feasible candidate start time slots that satisfy the scheduling constraints; Indicates the period A pre-maintained set of periodically aware pre-allocated time slots is used to guide the current flow to preferentially select time slots that have a congruent alignment with the allocated time-triggered flow; symbol " " indicates a set assignment / update operation; like If the set is not empty, the earliest starting time slot will be selected from this set first. like It is empty, but If the value is not empty, it means that there is no feasible time slot on the current link that satisfies the periodicity-aware priority condition. In this case, the process degenerates into selecting the earliest starting time slot from the set of all feasible candidate time slots. like If empty, it indicates a time-triggered stream. Unable to be on the link If the conflict-free transmission and latency constraints are satisfied, the current scheduling instance is determined to be unschedulable.

4. The method according to claim 1, characterized in that, In step 5, After each successful time slot allocation, the PriA time slot set for the corresponding period is updated according to the following formula: in, This is a deadline-aware filtering function; the default filtering rule is: If all time-triggered streams have been allocated, the output will include the gating control list, send offset, queue, routing, and delay results.

5. The method according to claim 4, characterized in that, The gating control list generation process includes: assigning the starting time slot of each time-triggered stream to each link. Add to the gating control list and trigger the flow period based on time. In supercycle The inner loop repeats, eventually returning GCL, Delay, Offset, Queue, and Route.

6. A time-sensitive network periodic-aware scheduling system for cloud computer interlocking, used to implement the method described in any one of claims 1-5, characterized in that, It includes a topology and flow model construction module, a hyperperiodic and congruent partitioning module, a periodic relationship pre-calculation module, a PriA time slot set maintenance module, a time-triggered flow sorting module, a periodic awareness list scheduling module, and a gating control list generation module; The topology and flow model construction module is used to receive the cloud computer interlocking time-sensitive network topology, link rate, time slot granularity, queue number and time-triggered flow configuration, and to establish a directed graph model and a time-triggered flow model. The super-period and congruence partitioning module is used to determine the super-period based on the period set and map the time slots within the super-period to congruence remainder classes under different periods. The periodic relationship pre-calculation module is used to construct a modular addition group and a congruent subgroup based on the greatest common divisor relationship, and to pre-generate congruent mapping relationships between different periods; The PriA time slot set maintenance module is used to maintain the PriA time slot set corresponding to each period, and to update the PriA time slot set after the time-triggered stream is successfully allocated; The time-triggered stream sorting module is used to determine the scheduling order based on the deadline, period, path length, and transmission length. The periodic awareness list scheduling module is used to calculate feasible candidate time slots link by link along the fixed route of the time-triggered flow, and preferentially select feasible candidate time slots belonging to the current period PriA time slot set; The gating control list generation module is used to output gating control lists, sending offsets, queues, routing, and latency results based on the scheduling results.

Images