A scheduling system for real-time network traffic

Abstract

The present invention discloses a scheduling system for real-time network traffic. The TT scheduling module implements the scheduling and transmission of TT traffic in the queue corresponding to the shared cache according to a time schedule table; the RC scheduling module implements the scheduling and transmission of RC traffic in the queue corresponding to the shared cache by using polling and traffic shaping; the BE scheduling module schedules and transmits BE traffic outside and inside the SoC by polling and traffic shaping; the hybrid scheduling module implements the mixed traffic scheduling and transmission of different traffic data frames by using a medium reservation scheduling method according to the scheduling priority: TT traffic > PCF protocol control frame > RC traffic > BE traffic, thereby meeting the low latency of network traffic transmission and the real-time and deterministic transmission of key traffic; in terms of engineering implementation, configurable dual-redundancy or triple-redundancy channel transmission and radiation-resistant reinforcement design are adopted to meet the SEU radiation-resistant standard test and improve the reliability of real-time network traffic transmission.

Description

Technical Field

[0001] This invention relates to the field of real-time communication in computer networks, and in particular to a scheduling system for real-time network traffic. Background Technology

[0002] With the development of next-generation launch vehicles, satellites, spacecraft, large aircraft, and automobiles towards high levels of informatization and intelligence, traditional bus networks, characterized by low bandwidth and lack of flexible networking support, and based on event-triggered communication modes lacking real-time communication mechanisms and with uncertain data arrival times, can no longer meet the demands of intelligent mission development and multi-source network data fusion transmission. To meet the requirements of real-time communication, real-time Ethernet that supports both time-triggered and event-triggered communication has become essential. Therefore, the development of real-time Ethernet end systems with real-time performance, high reliability, and flexible configurable scheduling tables has become a trend.

[0003] Real-time network traffic schedulers are a crucial component of real-time Ethernet end systems, situated at the forefront of these systems. Traditional real-time network traffic scheduling methods employ both preemptive and non-preemptive scheduling. Preemptive scheduling guarantees fixed latency and jitter for TT (transmission time) traffic but increases jitter in ET (transmission duration) traffic and reduces bandwidth utilization. Non-preemptive scheduling avoids the frame fragmentation and excessive jitter issues associated with preemptive scheduling, but introduces variable transmission jitter into TT traffic. In the extreme working environment of aerospace, the overall performance of the real-time network end system will be severely degraded. To ensure the determinism and real-time communication capabilities of real-time Ethernet end systems, improving the data transmission bandwidth, latency, and reliability of real-time network traffic scheduling becomes critical, necessitating the design of a deterministic and highly reliable real-time network traffic scheduling system. Summary of the Invention

[0004] The technical problem solved by this invention is to overcome the shortcomings of the prior art and provide a scheduling system for real-time network traffic that meets the requirements of fixed latency, bounded jitter, and high bandwidth of key data frames in real-time network hybrid scheduling, thereby achieving high reliability, determinism, and real-time communication in real-time networks.

[0005] The technical solution of the present invention is: to provide a scheduling system for real-time network traffic, including: a TT scheduling module, which is used to cache TT traffic, generate TT traffic request scheduling information and initiate TT traffic scheduling requests, transmit TT traffic under the scheduling of the hybrid scheduling module, and feed back a TT traffic transmission end flag to the hybrid scheduling module;

[0006] The RC scheduling module is used to cache RC traffic, generate RC traffic request scheduling information and initiate RC traffic scheduling requests. Under the scheduling of the hybrid scheduling module, it transmits RC traffic and sends an RC traffic transmission end flag back to the hybrid scheduling module.

[0007] The BE scheduling module is used to cache BE traffic, generate BE traffic request scheduling information and initiate RC traffic scheduling requests. Under the scheduling of the hybrid scheduling module, it transmits BE traffic and sends a BE traffic transmission end flag back to the hybrid scheduling module.

[0008] The time synchronization module is used to initiate scheduling requests for PCF protocol control frames, transmit PCF protocol control frames under the scheduling of the hybrid scheduling module, and send a PCF protocol control frame transmission end flag back to the hybrid scheduling module.

[0009] The hybrid scheduling module prioritizes hybrid scheduling services according to the following order: TT traffic > PCF protocol control frames > RC traffic > BE traffic. Based on the requested scheduling information, it responds to each scheduling request using a media reservation method to achieve service traffic scheduling.

[0010] TT traffic and RC traffic are stored using a shared cache, while BE traffic is stored using a separate cache.

[0011] Furthermore, the TT scheduling module, RC scheduling module, and BE scheduling module generate request scheduling information by combining the queue status information in the corresponding caches with the TT traffic transmission time scheduling configuration table, RC traffic BAG table, and BE traffic configurable BAG value, respectively.

[0012] Furthermore, the TT scheduling module and the RC scheduling module share a common cache for caching 2. n The system has several traffic queues, one of which is used to buffer the TT traffic queue and the other is used to buffer the RC traffic queue. It is also used to output the TT data frames in the TT traffic queue or the RC data frames in the RC traffic queue after the hybrid scheduling module responds to the scheduling request. It also updates the queue status information of all traffic queues in the shared buffer in real time according to the input and output of data frames in each traffic queue. After the output is completed, it sends the TT traffic transmission end flag or the RC traffic transmission end flag back to the hybrid scheduling module.

[0013] The TT scheduling module includes a TT traffic transmission time scheduling configuration table and n sub-scheduling modules;

[0014] The TT traffic transmission time scheduling configuration table is used to configure and plan the scheduling information for each TT traffic queue.

[0015] The TT traffic queue is scheduled for transmission according to the TT traffic transmission time scheduling configuration table. One scheduling cycle is split into multiple integrated cycles, and each integrated cycle is further divided into n different time slices. The scheduling of all TT traffic queues is completed by cyclically completing multiple integrated cycles. At the beginning of each integrated cycle, the n sub-scheduling modules jointly read the TT traffic transmission time scheduling configuration table to obtain the queue numbers of the n TT traffic queues and the different trigger time intervals corresponding to each queue number, and trigger the TT traffic scheduling request enable. The TT traffic request scheduling information, namely the queue number of the TT traffic queue and the different trigger time intervals corresponding to each queue number, as well as the queue status information updated by the shared cache, is sent to the hybrid scheduling module for TT traffic scheduling.

[0016] Furthermore, the RC scheduling module includes an RC traffic BAG table, multiple two-level arbitration pollers, and 2 n One flow shaping module;

[0017] The RC traffic BAG table is used to configure and plan the traffic shaping information for each RC traffic queue.

[0018] The two-level arbitration poller polls the queue status information in the shared buffer to obtain the queue number of the RC traffic queue to be scheduled and the frame length information of the RC data frame in the corresponding RC traffic queue.

[0019] The traffic shaping module reads the RC traffic BAG table for the RC traffic queue to be scheduled, obtains the traffic shaping time interval between each RC data frame in the corresponding RC traffic queue, performs traffic shaping on the RC traffic queue to be scheduled, and triggers RC scheduling request enable; it then sends the RC traffic request scheduling information, namely the queue number of the RC traffic queue to be scheduled, the frame length information of the RC data frames in the corresponding RC traffic queue, and the traffic shaping time interval between each RC data frame, to the hybrid scheduling module for RC traffic scheduling.

[0020] Furthermore, the BE scheduling module includes a poller, a traffic shaping module, and a separate FIFO buffer module;

[0021] A separate FIFO cache module is used to cache one external BE traffic queue and one internal BE traffic queue of the SoC;

[0022] The poller polls the queue status information in the FIFO buffer to obtain the queue type of the BE traffic queue to be scheduled and the frame length information of the BE data frames in the corresponding BE traffic queue.

[0023] The traffic shaping module is a configurable BAG value traffic shaping module. It uses the configured traffic shaping BAG value time interval to shape the BE traffic queue to be scheduled and triggers BE scheduling request enable. It sends the BE traffic request scheduling information, namely the queue type of the BE traffic queue to be scheduled, the frame length information of the BE data frames in the corresponding BE traffic queue, and the traffic shaping BAG value time interval between each BE data frame, to the hybrid scheduling module for BE traffic scheduling.

[0024] Furthermore, the TT scheduling module and RC scheduling module add serial number redundancy to the output TT traffic queue and RC traffic queue for transmission; the time synchronization module uses multiple redundant channels to output the output PCF protocol control frames; and the BE scheduling module selects multiple different channel ports for outputting the output BE traffic queue according to the channel number.

[0025] Furthermore, the hybrid scheduling module adopts a request-response approach, responding to corresponding scheduling requests to achieve scheduling of different traffic types, specifically as follows:

[0026] Determine if a TT traffic scheduling request exists; if a TT traffic scheduling request exists, respond to the TT traffic scheduling request; if no TT traffic scheduling request exists, then:

[0027] Determine if a PCF protocol control frame scheduling request exists. If a PCF protocol control frame scheduling request exists, determine if the current output PCF protocol control frame blocks the next TT data frame. If it blocks, do not respond to the PCF frame scheduling request; if it does not block, respond to the PCF frame scheduling request. If no PCF protocol control frame scheduling request exists, then:

[0028] Determine if an RC traffic scheduling request exists. If an RC traffic scheduling request exists, determine if the current output RC traffic is blocking the next TT data frame or the next PCF protocol control frame. If it is blocking, do not respond to the RC traffic scheduling request; if it is not blocking, respond to the RC traffic scheduling request. If no RC traffic scheduling request exists, then:

[0029] Determine if a BE traffic scheduling request exists. If a BE traffic scheduling request exists, check if the current output BE traffic is blocking the next TT data frame, the next PCF protocol control frame, or the next RC data frame. If it is blocking, do not respond to the BE traffic scheduling request. If it is not blocking any of them, respond to the BE traffic scheduling request. If no BE traffic scheduling request exists, the hybrid scheduling module enters an idle state.

[0030] Furthermore, the hybrid scheduling module adopts a media reservation method. Based on the media reservation time, it determines the blocking status of TT data frames when responding to PCF protocol control frame scheduling requests, RC traffic scheduling requests, or BE traffic scheduling requests. The media reservation time is divided into PCF frame scheduling blocking time, RC traffic scheduling blocking time, and BE traffic scheduling blocking time.

[0031] Furthermore, the factors determining the media reservation time include traffic type, data frame length of the corresponding traffic, Ethernet MAC rate, and MAC layer buffer usage;

[0032] PCF frame scheduling blocking time: It is obtained by combining the different trigger time intervals corresponding to each TT traffic queue number, the time when the TT scheduling request is enabled, the TT traffic lookup reservation time, the configurable fixed blocking time of PCF frames, and the PCF frame reservation interval time.

[0033] The TT traffic lookup time is the time from the start of each integration cycle to the completion of reading the TT traffic sending time scheduling configuration table.

[0034] The fixed blocking time of PCF frames is configured according to the Ethernet MAC rate. At gigabit speeds, the fixed blocking time of PCF frames is greater than half the frame length of the PCF protocol control frame. At 100 Mbps speeds, the fixed blocking time of PCF frames is 10 times the configured value of the fixed blocking time of PCF frames at gigabit speeds.

[0035] The PCF frame reservation interval is determined based on the Ethernet MAC rate and the MAC layer buffer usage at the time of the PCF protocol control frame scheduling request. If the MAC layer buffer usage is empty, the PCF frame reservation interval is 0. If the MAC layer buffer usage is not empty, at gigabit speeds, the PCF frame reservation interval is 1 / 4 of the PCF protocol control frame length, and at 100 Mbps speeds, the PCF frame reservation interval is 10 times that at gigabit speeds.

[0036] RC traffic scheduling blocking time: It is obtained by combining the different trigger time intervals corresponding to each TT traffic queue number, the time when the TT scheduling request is enabled, the TT traffic lookup reservation time, the configurable fixed blocking time of RC traffic, and the RC traffic reservation interval time.

[0037] The RC traffic blocking time is fixed and configured according to the Ethernet MAC rate. It is 10 times the configured value for 100 Mbps network speed and 1 Gigabit network speed.

[0038] The RC traffic reservation interval is determined based on the data frame length, Ethernet MAC rate, and MAC layer buffer usage in the current RC traffic queue to be scheduled.

[0039] At gigabit speeds, if the current MAC layer cache usage is less than 1 / 4 of the frame length of the RC data frames in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of the frame length of the RC data frames in the current RC traffic queue to be scheduled, the current MAC layer cache usage, and 1 / 4 of the frame length of the RC data frames in the current RC traffic queue to be scheduled; if the current MAC layer cache usage is not less than 1 / 4 of the frame length of the RC data frames in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of the frame length of the RC data frames in the current RC traffic queue to be scheduled and the current MAC layer cache usage.

[0040] At 100 Mbps network speed, if 10 times the current MAC layer cache usage is less than 1 / 4 of the frame length of the RC data frame in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of 10 times the frame length of the RC data frame in the current RC traffic queue to be scheduled, plus 10 times the current MAC layer cache usage, and 1 / 4 of the frame length of the RC data frame in the current RC traffic queue to be scheduled; if 10 times the current MAC layer cache usage is not less than 1 / 4 of the frame length of the RC data frame in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of 10 times the frame length of the RC data frame in the current RC traffic queue to be scheduled, plus 10 times the current MAC layer cache usage.

[0041] BE traffic scheduling blocking time: It is obtained by combining the different trigger time intervals corresponding to each TT traffic queue number, the time when the TT scheduling request is enabled, the TT traffic lookup reservation time, the configurable fixed blocking time of BE traffic, and the BE traffic reservation interval time.

[0042] The fixed blocking time for BE traffic is configured in the same way as the fixed blocking time for RC traffic.

[0043] The BE traffic reservation interval is determined based on the data frame length, Ethernet MAC rate, and MAC layer buffer usage in the current BE traffic queue to be scheduled. The calculation method is the same as that for RC traffic reservation interval, and the BE data frame length in the BE traffic queue to be scheduled is used in the calculation.

[0044] The advantages of this invention compared to the prior art are:

[0045] (1) This invention designs a TT scheduling module, an RC scheduling module, and a BE scheduling module. Compared with conventional scheduling, it has higher determinism and controllability of each traffic service, while retaining the advantages of flexibility of each traffic service. A hybrid scheduling module is designed, which uses a configurable media reservation scheduling method and determines the scheduling priority according to the nature of each traffic service: TT>PCF>RC>BE for corresponding traffic scheduling and transmission. Through the media reservation scheduling method and the fact that high-priority traffic can be sent first, when the scheduling of PCF frames, RC traffic, and BE traffic switches to TT traffic time slice scheduling, the configurable media reservation time determines the possible conflicts that may occur when MAC layer sends data frames, reduces the transmission delay of TT traffic, ensures the real-time performance of critical traffic, effectively improves the traffic control and scheduling efficiency of real-time networks, maximizes bandwidth utilization, and meets key network requirements such as high bandwidth, low latency, and predictability.

[0046] (2) This invention enables the modification of time planning and other table items and parameters according to specific needs in actual application scenarios through the corresponding TT traffic transmission time scheduling table, RC traffic BAG table and BE traffic configurable BAG value, so that each traffic service can be controlled.

[0047] (3) The present invention designs a configuration method and configuration value for media reservation in the hybrid scheduling module. The media reservation time is determined according to the Ethernet MAC rate, traffic type, corresponding data frame length and MAC layer buffer usage. The design of configurable fixed occlusion time of PCF frame and PCF frame reservation interval time solves the problem of uncertain delay output in MAC layer buffer after PCF frame response scheduling, and ensures that the transmission delay of PCF protocol control frame is fixed and the jitter is bounded.

[0048] (4) At the start of each integration cycle, the n sub-scheduling modules of the present invention jointly read the TT traffic transmission time scheduling table, which improves the efficiency of TT traffic table lookup and reduces the table lookup delay. The offset time of each integration cycle is used as the TT traffic table lookup reservation time. By using the table lookup reservation time as part of the reservation time of the scheduling transmission medium for PCF frames, RC traffic and BE traffic, the TT traffic is scheduled and transmitted near the start time point, which further reduces the TT traffic delay and ensures the real-time performance of TT traffic scheduling.

[0049] (5) The RC scheduling module of the present invention uses a two-level arbitration polling method and fast BAG table lookup to perform traffic shaping, which reduces the jitter of RC traffic scheduling and transmission. The queue number extraction adopts a two-level arbitration polling method, which reduces the consumption of chip resources and improves timing optimization. The scheduling and transmission of external BE traffic and SoC internal BE traffic by polling and configurable traffic shaping BAG value ensures the normal transmission of non-critical business BE traffic.

[0050] (6) This invention provides a shared cache, offering a maximum shared cache size of 2 for TT traffic and RC traffic. n The system buffers traffic data frames in each queue, enabling TT traffic and RC traffic to share the buffer using different queue numbers. It also supports scheduled transmission of TT traffic degraded to RC traffic, ensuring that TT traffic transmission information is not lost and improving the reliability of critical traffic transmission. By sharing the buffer and multiplexing the data flow path between TT and RC services, the chip's consumption of storage resources is reduced.

[0051] (7) In engineering implementation, data transmission adopts configurable dual-redundant or triple-redundant channel transmission and software radiation-resistant triple-modal redundancy reinforcement design to improve the reliability of the scheduling system in actual engineering applications.

[0052] (8) The real-time network traffic scheduling system designed in this invention can not only meet the requirements of the above-mentioned aerospace electronic systems, but also be applicable to any other Ethernet that uses priority service scheduling for transmission. Attached Figure Description

[0053] Figure 1 A schematic diagram of a scheduling system for real-time network traffic services provided by the present invention;

[0054] Figure 2 A schematic diagram illustrating the implementation of hybrid scheduling for TT traffic services provided by this invention;

[0055] Figure 3 This is a schematic diagram of the mixed scheduling system for various service traffic provided in an embodiment of the present invention;

[0056] Figure 4 This is a schematic diagram of the overall simulation results of the mixed scheduling system for various service traffic provided in an embodiment of the present invention. Detailed Implementation

[0057] To better understand the technical solution of the present invention, the embodiments of the present invention will be specifically described below with reference to the accompanying drawings. It should be understood that the specific embodiments described are used to explain the present invention, but do not limit the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0058] Figure 1This invention provides a schematic diagram of a scheduling system for various service traffic in a real-time network, with reference to... Figure 1 The scheduling system includes a TT scheduling module, an RC scheduling module, and a BE scheduling module, which are used to cache TT traffic, RC traffic, and BE traffic, respectively, generate their respective request scheduling information, and initiate scheduling requests. Under the scheduling of the hybrid scheduling module, the corresponding traffic is transmitted, and each traffic transmission ends and sends a signal to the hybrid scheduling module.

[0059] The time synchronization module is used to initiate scheduling requests for PCF protocol control frames, transmit PCF protocol control frames under the scheduling of the hybrid scheduling module, and send a PCF protocol control frame transmission end flag back to the hybrid scheduling module.

[0060] The hybrid scheduling module prioritizes services according to the following order: TT traffic > PCF protocol control frames > RC traffic > BE traffic. Based on the requested scheduling information, it responds to each scheduling request using media reservation to achieve service traffic scheduling.

[0061] TT and RC traffic use a shared cache, while BE traffic uses a separate cache. Specifically, the TT scheduling module and the RC scheduling module share a single shared cache for a maximum of 2... n The shared buffer manages the buffers for each traffic queue. Therefore, it pre-plans which queue numbers belong to TT (Traffic Transaction) or RC (Remote Transaction) traffic. By default, 1 / 4 of the queue numbers are designated for TT traffic, and the rest for RC traffic. The shared buffer is also used to output TT data frames from the TT traffic queue or RC data frames from the RC traffic queue after the hybrid scheduling module responds to the scheduling request. It also updates the queue status information of all traffic queues in the shared buffer in real time (whether the corresponding data frame exists in each traffic queue) based on the input and output of data frames in each traffic queue. After output, it sends a TT traffic transmission end flag or an RC traffic transmission end flag back to the hybrid scheduling module. Furthermore, using a shared buffer for TT and RC traffic allows TT traffic to degenerate into RC traffic for scheduling and transmission, ensuring no information is lost during TT traffic transmission and improving the reliability of critical traffic transmission.

[0062] In one possible implementation, the TT scheduling module includes a TT traffic transmission time scheduling configuration table, n sub-scheduling modules, and a shared buffer TT traffic queue, where n is preferably a multiple of 4.

[0063] The TT traffic transmission time scheduling configuration table is used to configure and plan the scheduling information for each TT traffic queue. In specific implementation, the TT traffic transmission time scheduling configuration table can be planned in advance by making it software configurable.

[0064] A shared buffer TT traffic queue is used, and transmission is scheduled according to the TT traffic transmission time scheduling configuration table. A scheduling period can be split into multiple integrated periods. After the corresponding TT traffic is scheduled in each integrated period, it is re-scheduled. Each integrated period is further divided into n different time slices. These different time slices serve as the transmission scheduling times for n different TT traffic. PCF protocol control frames, RC traffic, and BE traffic are not transmitted in these time slices, while the remaining time slices are scheduled to transmit PCF protocol control frames, RC traffic, and BE traffic according to priority.

[0065] There are n sub-scheduling modules, see [link to relevant documentation] Figure 2 As shown, within each integration cycle, at the beginning of the integration cycle, the n sub-scheduling modules jointly read the TT traffic transmission time scheduling configuration table to obtain the queue numbers of the n TT traffic queues and the different trigger time intervals corresponding to each queue number, triggering the TT traffic scheduling request enable; and send the TT traffic request scheduling information, i.e., the queue number of each TT traffic queue and the different trigger time intervals corresponding to each queue number, as well as the queue status information of the shared cache update, to the hybrid scheduling module for TT traffic scheduling.

[0066] Upon receiving a response from the hybrid scheduling module to the TT traffic scheduling request, the shared buffer outputs the TT traffic with the corresponding queue number. After the output is completed, it sends a TT traffic transmission end flag back to the hybrid scheduling module. Preferably, the output TT traffic is serialized and transmitted through three redundant channels.

[0067] Preferably, the shared cache provides a total of 512KB of cache space with 8192 cache units (each cache unit is 64B in size), which can complete the individual configuration of the private space cache unit size for each traffic queue, providing a maximum of 256 queues (n=8) for TT traffic and RC traffic, of which 64 are configured as TT traffic queues and 192 are RC traffic queues according to the nature of the service traffic; the TT scheduling module includes a TT traffic transmission time scheduling configuration table, 8 sub-scheduling modules and 64 shared cache TT traffic queues; the TT traffic transmission time scheduling configuration table is a pre-planned transmission time scheduling configuration table for the 64 TT traffic queues, so it is necessary to read the transmission time scheduling configuration table at least 8 times in the integration cycle to finally obtain the queue number of the 64 TT traffic queues and the different trigger time intervals corresponding to each queue number.

[0068] In one possible implementation, the time synchronization module initiates a PCF protocol control frame scheduling request when a PCF protocol control frame needs to be transmitted; after receiving a response from the hybrid scheduling module to the PCF protocol control frame scheduling request, it outputs the PCF protocol control frame and sends a PCF protocol control frame transmission end flag back to the hybrid scheduling module after the output is completed; preferably, the output PCF protocol control frame is transmitted through three redundant channels.

[0069] In one possible implementation, the RC scheduling module includes an RC traffic BAG table, multiple two-level arbitration pollers, and 2 n A traffic shaping module and a shared cache RC traffic queue.

[0070] The RC traffic BAG table is used to configure and plan the traffic shaping information for each RC traffic queue.

[0071] The two-level arbitration poller polls the queue status information in the shared buffer to obtain the queue number of the RC traffic queue to be scheduled and the frame length information of the RC data frame in the corresponding RC traffic queue.

[0072] The traffic shaping module reads the RC traffic BAG table for the RC traffic queue to be scheduled, obtains the traffic shaping time interval between each RC data frame in the corresponding RC traffic queue, performs traffic shaping on the RC traffic queue to be scheduled, and triggers RC scheduling request enable; it then sends the RC traffic request scheduling information, namely the queue number of the RC traffic queue to be scheduled, the frame length information of the RC data frames in the corresponding RC traffic queue, and the traffic shaping time interval between each RC data frame, to the hybrid scheduling module for RC traffic scheduling.

[0073] Upon receiving a response from the hybrid scheduling module to the RC traffic scheduling request, the shared buffer outputs the RC traffic with the corresponding queue number. After the output is completed, it sends an RC traffic transmission end flag back to the hybrid scheduling module. Preferably, the output RC traffic is serialized and transmitted through three redundant channels.

[0074] Preferably, the RC scheduling module includes 16 two-level arbitration pollers, an RC traffic BAG table, 256 traffic shaping modules, and 192 shared buffer RC traffic queues. In the RC traffic BAG table, the traffic shaping time interval (with an accuracy of 10µs) for the 192 RC traffic queues is pre-planned via the AHB bus. Using the 192 traffic shaping modules, the corresponding traffic shaping time interval is obtained, and traffic shaping is performed on the RC traffic queues to be scheduled, ensuring that the RC data frames for each RC traffic queue number adhere to the required bandwidth interval.

[0075] In one possible implementation, the BE scheduling module includes a poller, a traffic shaping module, and a separate FIFO buffer module.

[0076] A separate FIFO cache module is used to cache one external BE traffic queue (BEEXIT) and one internal BE traffic queue (BEINN) of the SoC; the scheduling system is a component of the SoC, the internal BE traffic queue of the SoC contains data from other components of the SoC, and the external BE traffic queue of the SoC contains data from outside the SoC.

[0077] The poller polls the queue status information in the FIFO buffer to obtain the queue type of the BE traffic queue to be scheduled and the frame length information of the BE data frames in the corresponding BE traffic queue.

[0078] The traffic shaping module is a configurable BAG value traffic shaping module. It uses the configured traffic shaping BAG value time interval to shape the BE traffic queue to be scheduled and triggers BE scheduling request enable. It sends the BE traffic request scheduling information, that is, the queue type of the BE traffic queue to be scheduled (i.e., whether it is external BE traffic or internal BE traffic of SoC), the frame length information of the BE data frames in the corresponding BE traffic queue, and the traffic shaping BAG value time interval between each BE data frame, to the hybrid scheduling module for BE traffic scheduling.

[0079] Upon receiving the response from the hybrid scheduling module to the BE traffic scheduling request, the individual FIFO buffer module outputs BE traffic of the corresponding queue type. After the output is completed, it sends a BE traffic transmission end flag back to the hybrid scheduling module. Preferably, the output BE traffic can be output through three different network ports.

[0080] Preferably, the BE scheduling module includes one poller, one traffic shaping module, and two FIFO buffers. The two FIFO buffers are used to store the external BE traffic queue and the internal BE traffic queue of the SoC, respectively; in the traffic shaping module, the time interval of the traffic shaping BAG value of BE traffic is configured via the AHB bus (with an accuracy of 10µs).

[0081] The hybrid scheduling module uses media reservation for traffic scheduling. The media reservation time is determined based on the Ethernet MAC rate, traffic type, data frame length, and MAC layer buffer usage. The traffic is scheduled and transmitted according to the priority of service traffic: TT traffic > PCF protocol control frames > RC traffic > BE traffic.

[0082] Furthermore, the hybrid scheduling module adopts a request-response approach. First, each scheduling module generates request scheduling information and initiates a scheduling request. The hybrid scheduling module then determines the hybrid scheduling priority based on this information and responds to traffic that meets the priority requirements. The following describes the hybrid scheduling workflow for various service traffic within the hybrid scheduling module. Figure 3 As shown, the details are as follows:

[0083] a. The hybrid scheduling module starts in an idle state and determines whether TT traffic needs to be sent. If TT traffic needs to be sent, it responds to the TT traffic scheduling request. After responding to the TT traffic scheduling request, if there is a TT traffic queue in the corresponding queue number in the shared buffer, the shared buffer outputs the TT traffic queue for the corresponding queue number and sends a TT traffic transmission end flag back to the hybrid scheduling module. Then, it enters the TT traffic protection interval time. After the protection interval time is completed, the scheduling of this TT traffic ends, and the hybrid scheduling module enters the idle state again. The TT traffic scheduling is complete. If there is no TT traffic queue in the corresponding queue number in the shared buffer, it waits for a fixed time and then enters the TT traffic protection interval time. After the protection interval time is completed, the scheduling of this TT traffic ends, and the hybrid scheduling module enters the idle state again. The TT traffic scheduling is complete. The fixed waiting time and the protection interval time are both fixed values.

[0084] If the shared buffer times out sending TT frames, the hybrid scheduling module will return to the idle state and end the scheduling of the current TT traffic.

[0085] b. When there is no TT traffic to send, determine whether a PCF protocol control frame needs to be sent based on the request scheduling information. If a PCF protocol control frame needs to be sent, determine whether the current PCF protocol control frame will block the scheduling of the next TT traffic frame based on the medium reservation time. If so, the hybrid scheduling module returns to the idle state, waits for the scheduling of the next TT traffic frame, and then responds by sending a PCF protocol control frame. If the current PCF protocol control frame does not block the scheduling of the next TT traffic frame, the time synchronization module sends a PCF protocol control frame and feeds back the PCF protocol control frame end flag to the hybrid scheduling module. The PCF protocol control frame sending ends, the hybrid scheduling module returns to the idle state, and the current PCF protocol control frame scheduling is complete.

[0086] If the time synchronization module times out sending the PCF protocol control frame, the hybrid scheduling module will return to the idle state and end the scheduling of the current PCF protocol control frame.

[0087] c. When no PCF protocol control frame needs to be sent, determine whether RC traffic needs to be sent. If RC traffic needs to be sent, determine whether the current RC data frame being sent blocks the scheduling of the next TT traffic frame based on the request scheduling information and media reservation time. If yes, the hybrid scheduling module returns to the idle state, waits for the scheduling of the next TT traffic frame, and then responds by sending an RC data frame. Otherwise, determine whether the current RC data frame being sent blocks the scheduling of the next PCF protocol control frame. If yes, the hybrid scheduling module returns to the idle state, waits for the scheduling of the next PCF frame, and then responds by sending an RC data frame. If the current RC data frame being sent does not block the transmission of the next TT traffic frame and PCF protocol control frame, request to send an RC data frame, wait for the shared buffer to send an RC data frame, and send a feedback RC data frame end flag to the hybrid scheduling module. If the hybrid scheduling module does not receive the feedback RC data frame end flag, it waits for the shared buffer to send the RC data frame end flag. If the hybrid scheduling module receives the feedback RC data frame end flag, the current RC data frame transmission ends, the hybrid scheduling module returns to the idle state, and the current RC traffic scheduling is completed.

[0088] If the wait for the shared buffer to send RC data frames times out, the hybrid scheduling module will directly return to the idle state and end the scheduling of the current RC traffic.

[0089] d. When there is no RC traffic to send, determine whether BE traffic needs to be sent. If yes, determine whether the current BE data frame being sent blocks the scheduling of the next TT traffic frame based on the request scheduling information and media reservation time. If yes, the hybrid scheduling module returns to the idle state, waits for the scheduling of the next TT traffic frame, and then responds by sending a BE data frame. Otherwise, determine whether the current BE data frame being sent blocks the scheduling of the next PCF protocol control frame. If yes, the hybrid scheduling module returns to the idle state, waits for the scheduling of the next PCF protocol control frame, and then responds by sending a BE data frame. Otherwise, determine whether the current BE data frame being sent blocks the scheduling of the next RC data frame. If yes, the hybrid scheduling module returns to the idle state, waits for the scheduling of the next RC data frame, and then responds by sending a BE data frame. If the current BE data frame being sent does not block the scheduling of the next TT traffic frame, PCF protocol control frame, and RC data frame, i.e., during the idle time, send a BE data frame, waits for separate buffering to send the current BE data frame, and sends a BE data frame sending end flag to the hybrid scheduling module. The BE data frame sending ends, the hybrid scheduling module returns to the idle state, and the current BE traffic scheduling is completed. Waiting to begin responding to the next different traffic scheduling.

[0090] If the wait for a separate buffer to send a BE data frame times out, the hybrid scheduling module will directly return to the idle state and end the scheduling of the current BE traffic.

[0091] Furthermore, considering the different processing clocks of the hybrid scheduling module and the subsequent MAC layer, the media reservation time is divided into PCF frame scheduling obstruction time, RC traffic scheduling obstruction time, and BE traffic scheduling obstruction time:

[0092] (1) PCF frame scheduling blockage time: It is obtained by combining the different trigger time intervals corresponding to the n TT traffic queue numbers, the time when the TT scheduling request is enabled, the TT traffic lookup table reservation time, the configurable PCF frame fixed blockage time, and the PCF frame reservation interval time; if a PCF protocol control frame scheduling request is received within the PCF frame scheduling blockage time, the request will wait for a response.

[0093] The TT traffic lookup table reservation time is the time from the start of each integration cycle to the completion of reading the TT traffic sending time scheduling configuration table. Since TT traffic has the highest scheduling priority and is not subject to the blocking authority of other service traffic, time must be reserved for RC, BE traffic and PCF protocol control frames at the start of each integration cycle to read the TT traffic sending time scheduling configuration table.

[0094] The fixed blocking time of PCF frames is configured according to the Ethernet MAC rate. At gigabit speeds, the fixed blocking time of PCF frames is greater than half the frame length of the PCF protocol control frame (in bytes). At 100 Mbps speeds, the fixed blocking time of PCF frames is 10 times the configured value of the fixed blocking time of PCF frames at gigabit speeds.

[0095] The PCF frame reservation interval is determined based on the Ethernet MAC rate and the MAC layer buffer usage at the time of the PCF protocol control frame scheduling request. To ensure that the PCF protocol control frame output still has a fixed delay jitter requirement, it is necessary to determine whether the MAC layer buffer usage is empty at this time. If the MAC layer buffer usage is empty, the PCF frame reservation interval is 0. If the MAC layer buffer usage is not empty, at gigabit speeds, the PCF frame reservation interval is 1 / 4 of the PCF protocol control frame length (in bytes), and at 100 Mbps speeds, the PCF frame reservation interval is 10 times that at gigabit speeds.

[0096] (2) RC traffic scheduling blocking time: It is obtained by combining the different trigger time intervals corresponding to the n TT traffic queue numbers, the time when the TT scheduling request is enabled, the TT traffic lookup reservation time, the configurable fixed blocking time of RC traffic, and the RC traffic reservation interval time; if an RC traffic scheduling request is received within the RC traffic scheduling blocking time, the request will wait for a response.

[0097] The RC traffic blocking time is fixed and configured according to the Ethernet MAC rate. It is 10 times the configured value for 100 Mbps network speed and 1 Gigabit network speed.

[0098] The RC traffic reservation interval is determined based on the data frame length, Ethernet MAC rate, and MAC layer buffer usage in the current RC traffic queue to be scheduled.

[0099] At gigabit speeds, if the current MAC layer cache usage is less than 1 / 4 of the frame length (in bytes) of the RC data frames in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of the frame length of the RC data frames in the current RC traffic queue to be scheduled, the current MAC layer cache usage, and 1 / 4 of the frame length of the RC data frames in the current RC traffic queue to be scheduled; if the current MAC layer cache usage is not less than 1 / 4 of the frame length of the RC data frames in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of the frame length of the RC data frames in the current RC traffic queue to be scheduled and the current MAC layer cache usage.

[0100] At 100 Mbps network speed, if 10 times the current MAC layer cache usage is less than 1 / 4 of the frame length (in bytes) of the RC data frame in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of 10 times the frame length of the RC data frame in the current RC traffic queue to be scheduled, plus 10 times the current MAC layer cache usage, and 1 / 4 of the frame length of the RC data frame in the current RC traffic queue to be scheduled; if 10 times the current MAC layer cache usage is not less than 1 / 4 of the frame length of the RC data frame in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of 10 times the frame length of the RC data frame in the current RC traffic queue to be scheduled, plus 10 times the current MAC layer cache usage.

[0101] (3) BE traffic scheduling blocking time: It is obtained by combining the different trigger time intervals corresponding to the n TT traffic queue numbers, the time when the TT scheduling request is enabled, the TT traffic lookup table reservation time, the configurable fixed blocking time of BE traffic, and the BE traffic reservation interval time; if a BE traffic scheduling request is received within the BE traffic scheduling blocking time, the request will wait for a response.

[0102] The fixed blocking time for BE traffic is configured in the same way as the fixed blocking time for RC traffic.

[0103] The BE traffic reservation interval is determined based on the data frame length, Ethernet MAC rate, and MAC layer buffer usage in the current BE traffic queue to be scheduled. The calculation method is the same as that for RC traffic reservation interval, and the BE data frame length in the BE traffic queue to be scheduled is used in the calculation.

[0104] Figure 4This is a schematic diagram of the overall simulation results of the mixed scheduling system for various service traffic provided in this embodiment of the invention. The results show that the mixed scheduling is carried out according to the traffic scheduling priority: TT>PCF>RC>BE. The configuration integration period is 150us. Within each integration period, TT traffic scheduling for 8 different queue numbers is implemented. The time interval between the TT traffic scheduling for 8 different queue numbers starting at time 0 of the integration period is 15us, including 2 frames of PCF protocol control frames, 8 frames of traffic with the same RC queue number configured with a BAG value of 10us, and 2 frames each of BE external traffic and SoC internal BE inner traffic. Different dashed lines in the figure represent the request scheduling enable of different traffic under the mixed scheduling situation. TT traffic is strictly scheduled and output according to the priority of the sending time scheduling table, which ensures the strict determinism and reliability of TT traffic under the mixed scheduling situation.

[0105] It is understood that this invention has been described through embodiments, and those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of this invention. Furthermore, under the teachings of this invention, these features and embodiments can be modified to adapt to specific circumstances without departing from the spirit and scope of this invention. Therefore, this invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are protected by this invention.

[0106] The contents not described in detail in this specification are common knowledge to those skilled in the art.

Claims

1. A scheduling system for real-time network traffic, characterized in that, include: The TT scheduling module is used to cache TT traffic. It uses the TT traffic transmission time scheduling configuration table and the queue status information in the TT traffic cache to generate TT traffic request scheduling information and initiate TT traffic scheduling requests. Under the scheduling of the hybrid scheduling module, it transmits TT traffic and sends a TT traffic transmission end flag back to the hybrid scheduling module. The RC scheduling module is used to cache RC traffic. It uses the RC traffic BAG table and combines it with the queue status information in the RC traffic cache to generate RC traffic request scheduling information and initiate RC traffic scheduling requests. Under the scheduling of the hybrid scheduling module, it transmits RC traffic and sends an RC traffic transmission end flag back to the hybrid scheduling module. The BE scheduling module is used to cache BE traffic. It uses configurable BAG values ​​and combines them with queue status information in the BE traffic cache to generate BE traffic request scheduling information and initiate RC traffic scheduling requests. Under the scheduling of the hybrid scheduling module, it transmits BE traffic and sends a BE traffic transmission end flag back to the hybrid scheduling module. The time synchronization module is used to initiate scheduling requests for PCF protocol control frames, transmit PCF protocol control frames under the scheduling of the hybrid scheduling module, and send a PCF protocol control frame transmission end flag back to the hybrid scheduling module. The hybrid scheduling module prioritizes services according to the following order: TT traffic > PCF protocol control frames > RC traffic > BE traffic. It responds to each scheduling request based on the corresponding request information to achieve service traffic scheduling. During the response process, based on the media reservation time, it determines the blocking status of TT data frames when responding to PCF protocol control frame scheduling requests, RC traffic scheduling requests, or BE traffic scheduling requests. The media reservation time is divided into PCF frame scheduling obstruction time, RC traffic scheduling obstruction time, and BE traffic scheduling obstruction time. Factors determining media reservation time include traffic type, data frame length of the corresponding traffic, Ethernet MAC rate, and MAC layer buffer usage; TT traffic and RC traffic are stored using a shared cache, while BE traffic is stored using a separate cache.

2. The scheduling system for real-time network traffic according to claim 1, characterized in that: The TT scheduling module and the RC scheduling module share a common buffer for caching 2. n The system has several traffic queues, one of which is used to buffer the TT traffic queue and the other is used to buffer the RC traffic queue. It is also used to output the TT data frames in the TT traffic queue or the RC data frames in the RC traffic queue after the hybrid scheduling module responds to the scheduling request. It also updates the queue status information of all traffic queues in the shared buffer in real time according to the input and output of data frames in each traffic queue. After the output is completed, it sends the TT traffic transmission end flag or the RC traffic transmission end flag back to the hybrid scheduling module. The TT scheduling module includes a TT traffic transmission time scheduling configuration table and n sub-scheduling modules; The TT traffic transmission time scheduling configuration table is used to configure and plan the scheduling information for each TT traffic queue. The TT traffic queue is scheduled for transmission according to the TT traffic transmission time scheduling configuration table. One scheduling cycle is split into multiple integrated cycles, and each integrated cycle is further divided into n different time slices. The scheduling of all TT traffic queues is completed by cyclically completing multiple integrated cycles. At the beginning of each integrated cycle, the n sub-scheduling modules jointly read the TT traffic transmission time scheduling configuration table to obtain the queue numbers of the n TT traffic queues and the different trigger time intervals corresponding to each queue number, and trigger the TT traffic scheduling request enable. The TT traffic request scheduling information, namely the queue number of the TT traffic queue and the different trigger time intervals corresponding to each queue number, as well as the queue status information updated by the shared cache, is sent to the hybrid scheduling module for TT traffic scheduling.

3. The scheduling system for real-time network traffic according to claim 2, characterized in that: The RC scheduling module includes an RC traffic BAG table, multiple two-level arbitration pollers, and 2 n One flow shaping module; The RC traffic BAG table is used to configure and plan the traffic shaping information for each RC traffic queue. The two-level arbitration poller polls the queue status information in the shared buffer to obtain the queue number of the RC traffic queue to be scheduled and the frame length information of the RC data frame in the corresponding RC traffic queue. The traffic shaping module reads the RC traffic BAG table for the RC traffic queue to be scheduled, obtains the traffic shaping time interval between each RC data frame in the corresponding RC traffic queue, performs traffic shaping on the RC traffic queue to be scheduled, and triggers RC scheduling request enable. The RC traffic request scheduling information, namely the queue number of the RC traffic queue to be scheduled, the frame length information of the RC data frames in the corresponding RC traffic queue, and the traffic shaping time interval between each RC data frame, is sent to the hybrid scheduling module for RC traffic scheduling.

4. The scheduling system for real-time network traffic according to any one of claims 1 to 3, characterized in that: The BE scheduling module includes a poller, a traffic shaping module, and a separate FIFO buffer module; A separate FIFO cache module is used to cache one external BE traffic queue and one internal BE traffic queue of the SoC; The poller polls the queue status information in the FIFO buffer to obtain the queue type of the BE traffic queue to be scheduled and the frame length information of the BE data frames in the corresponding BE traffic queue. The traffic shaping module is a configurable BAG value traffic shaping module. It uses the configured traffic shaping BAG value time interval to shape the BE traffic queue to be scheduled and triggers the BE scheduling request to be enabled. The BE traffic request scheduling information, namely the queue type of the BE traffic queue to be scheduled, the frame length information of the BE data frames in the corresponding BE traffic queue, and the time interval of the traffic shaping BAG value between each BE data frame, is sent to the hybrid scheduling module for BE traffic scheduling.

5. The scheduling system for real-time network traffic according to claim 4, characterized in that: The TT scheduling module and RC scheduling module add sequence numbers to the output TT traffic queue and RC traffic queue for redundant channel transmission; the time synchronization module uses multiple redundant channels to output the output PCF protocol control frames. The BE scheduling module selects multiple different channel ports for output based on the channel number of the output BE traffic queue.

6. The scheduling system for real-time network traffic according to claim 4, characterized in that: The hybrid scheduling module adopts a request-response approach, responding to scheduling requests to achieve scheduling of different traffic types. Specifically: Determine if a TT traffic scheduling request exists; if a TT traffic scheduling request exists, respond to the TT traffic scheduling request; if no TT traffic scheduling request exists, then: Determine if a PCF protocol control frame scheduling request exists. If a PCF protocol control frame scheduling request exists, determine if the current output PCF protocol control frame blocks the next TT data frame. If it blocks, do not respond to the PCF frame scheduling request; if it does not block, respond to the PCF frame scheduling request. If no PCF protocol control frame scheduling request exists, then: Determine if there is an RC traffic scheduling request currently in use; If an RC traffic scheduling request exists, determine whether the current output RC traffic is blocking the next TT data frame or the next PCF protocol control frame. If it is blocking, do not respond to the RC traffic scheduling request. If it is not blocking, respond to the RC traffic scheduling request. If there is no RC traffic scheduling request, then: Determine if a BE traffic scheduling request exists. If a BE traffic scheduling request exists, check whether the current output BE traffic is blocking the next TT data frame, the next PCF protocol control frame, or the next RC data frame. If it is blocking, do not respond to the BE traffic scheduling request; if it is not blocking, respond to the BE traffic scheduling request. If there is no BE traffic scheduling request, the hybrid scheduling module enters an idle state.

7. The scheduling system for real-time network traffic according to claim 1, characterized in that: PCF frame scheduling blocking time: It is obtained by combining the different trigger time intervals corresponding to each TT traffic queue number, the time when the TT scheduling request is enabled, the TT traffic lookup reservation time, the configurable fixed blocking time of PCF frames, and the PCF frame reservation interval time. The TT traffic lookup time is the time from the start of each integration cycle to the completion of reading the TT traffic sending time scheduling configuration table. The fixed blocking time of PCF frames is configured according to the Ethernet MAC rate. At gigabit speeds, the fixed blocking time of PCF frames is greater than half the frame length of the PCF protocol control frame. At 100 Mbps speeds, the fixed blocking time of PCF frames is 10 times the configured value of the fixed blocking time of PCF frames at gigabit speeds. The PCF frame reservation interval is determined based on the Ethernet MAC rate and the MAC layer buffer usage at the time of the PCF protocol control frame scheduling request. If the MAC layer buffer usage is empty, the PCF frame reservation interval is 0. If the MAC layer buffer usage is not empty, at gigabit speeds, the PCF frame reservation interval is 1 / 4 of the PCF protocol control frame length, and at 100 Mbps speeds, the PCF frame reservation interval is 10 times that at gigabit speeds. RC traffic scheduling blocking time: It is obtained by combining the different trigger time intervals corresponding to each TT traffic queue number, the time when the TT scheduling request is enabled, the TT traffic lookup reservation time, the configurable fixed blocking time of RC traffic, and the RC traffic reservation interval time. The RC traffic blocking time is fixed and configured according to the Ethernet MAC rate. It is 10 times the configured value for 100 Mbps network speed and 1 Gigabit network speed. The RC traffic reservation interval is determined based on the data frame length, Ethernet MAC rate, and MAC layer buffer usage in the current RC traffic queue to be scheduled. At gigabit speeds, if the current MAC layer cache usage is less than 1 / 4 of the frame length of the RC data frames in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of the frame length of the RC data frames in the current RC traffic queue to be scheduled, the current MAC layer cache usage, and 1 / 4 of the frame length of the RC data frames in the current RC traffic queue to be scheduled; if the current MAC layer cache usage is not less than 1 / 4 of the frame length of the RC data frames in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of the frame length of the RC data frames in the current RC traffic queue to be scheduled and the current MAC layer cache usage. At 100 Mbps network speed, if 10 times the current MAC layer cache usage is less than 1 / 4 of the frame length of the RC data frame in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of 10 times the frame length of the RC data frame in the current RC traffic queue to be scheduled, plus 10 times the current MAC layer cache usage, and 1 / 4 of the frame length of the RC data frame in the current RC traffic queue to be scheduled; if 10 times the current MAC layer cache usage is not less than 1 / 4 of the frame length of the RC data frame in the current RC traffic queue to be scheduled, then the RC traffic reservation interval is the sum of 10 times the frame length of the RC data frame in the current RC traffic queue to be scheduled, plus 10 times the current MAC layer cache usage. BE traffic scheduling blocking time: It is obtained by combining the different trigger time intervals corresponding to each TT traffic queue number, the time when the TT scheduling request is enabled, the TT traffic lookup reservation time, the configurable fixed blocking time of BE traffic, and the BE traffic reservation interval time. The fixed blocking time for BE traffic is configured in the same way as the fixed blocking time for RC traffic. The BE traffic reservation interval is determined based on the data frame length, Ethernet MAC rate, and MAC layer buffer usage in the current BE traffic queue to be scheduled. The calculation method is the same as that for RC traffic reservation interval time, and the calculation utilizes the BE data frame length in the BE traffic queue to be scheduled.

Images