Time-sensitive network frame preemption optimization method combined with time-aware shaper

Abstract

In order to solve the problem of high-speed frame delay caused by preemptible frames when the frame preemption conditions are not met, a time-sensitive network frame preemption optimization method combined with a time-aware shaper is provided. The core is to obtain the current time through the global clock of the time-aware shaper TAS, and obtain the opening time of the time window corresponding to the next time-sensitive frame to be transmitted through the static gating list to calculate the size of the transmittable frame. Based on the transmittable frame size, the transmission of the next preemptible frame is dynamically planned to ensure that when the preemption conditions are not met, the preemptible frame does not affect the transmission of the time-sensitive frame, that is, the high-speed frame.

Description

Technical Field

[0001] The invention belongs to the field of network communication, and in particular relates to a time-sensitive network frame preemption optimization method combined with a time-aware shaper. Background Art

[0002] The industry has proposed a variety of exclusive network protocols based on standard Ethernet, such as TTEthernet, EtherCAT, PROFINET, SERCOIII, etc., which are called industrial real-time Ethernet, and have gradually become the mainstream bus protocols for industrial control networks. However, the incompatibility of these protocols has led to problems such as difficult device compatibility, poor interoperability, difficulty in porting, and high development, deployment, and operation and maintenance costs. With the development of the industrial Internet, driven by the growing demand for interconnection and deterministic network standardization, IEEE 802 upgraded the audio and video bridging (AVB) working group, which was originally committed to meeting the needs of bandwidth reservation services, to the time-sensitive network (TSN) working group, and proposed a series of standards and specifications for link layer enhancement mechanisms and traffic strategies, mainly including time synchronization, traffic scheduling, reliable transmission, and network management. TSN follows the standard Ethernet protocol system and naturally has better interconnection advantages. It can realize open link layer forwarding while providing deterministic delay and bandwidth guarantee. In recent years, TSN, as a new generation of real-time Ethernet technology, has been widely used in many fields such as industrial Internet, avionics networks, vehicle networks, professional audio and video, and has received continuous attention from academia and industry.

[0003] Time Sensitive Shaper (TAS) is a traffic shaping mechanism for time-sensitive traffic in IEEE 802.1Qbv. All kinds of output frames (including time-sensitive frames and non-time-sensitive frames, and the flow consists of frames) are stored in the 8 queues of the output port according to the priority PCP introduced by IEEE802.1Q. Each queue is associated with an output gate. The switch of the gate is controlled by the static gating list. Each line in the static gating list represents the state of the output gate at that time. The gate state is represented by 8 bits, corresponding to 8 queues respectively. 0 means that the output gate associated with the queue is closed, and 1 means that the output gate is open. When the time-sensitive frame is transmitted, the output gates corresponding to the queues where the other non-time-sensitive frames are located are closed. In order to prevent the non-time-sensitive frames from affecting the time-sensitive frames, a guard band mechanism is introduced. The size of the guard band is the time required to transmit the maximum Ethernet frame (1542 bytes) based on the current port bandwidth. The guard band ensures that the time-sensitive frame is not affected by the non-time-sensitive frame during transmission. However, because other frames are not allowed to be transmitted in the guard band, it will lead to bandwidth waste. The combination of time sensor and frame preemption can shorten the guard band to 127 bytes, which improves the problem of bandwidth waste to a certain extent.

[0004] IEEE802.1Qbu and IEEE802.3br standards jointly define the TSN classic frame preemption model, in which IEEE802.3br introduces two MAC independent sublayers, namely the preemptible frame MAC sublayer pMAC and the high-speed frame MAC sublayer eMAC. The frames assigned to pMAC are called preemptible frames, and the frames assigned to eMAC are called high-speed frames (express frames). The frame preemption logic is executed in the MAC merging sublayer. The preemptible frames currently being transmitted will be sliced, and after encapsulating the frame preemption information, they will wait for an inter-frame gap time (IFG) before transmitting the high-speed frame, which includes the frame preemption overhead. After waiting for the high-speed frame transmission to end, the subsequent fragments of the preemptible frame will continue to be transmitted. The high-speed frame is generally a control flow with a certain delay constraint, requiring end-to-end low delay and low jitter.

[0005] The classic preemption model requires that high-speed frames can preempt preemptible frames, and frames of the same category cannot preempt each other. It also requires that the payload (Data) of the first frame fragment of the transmitted preemptible frame is at least 42 bytes, and the payload of the subsequent fragments of the preemptible frame is at least 60 bytes. The first frame fragment and the subsequent fragments of the preemptible frame must meet the Ethernet minimum frame size (84 bytes), so the minimum payload of the preemptible frame is 102 bytes. With the addition of additional encapsulation information including the interframe gap IFG, the maximum non-preemptible frame size can be calculated to be 143 bytes. This means that in the worst case, when the preemption condition is not met, the high-speed frame has to wait for a maximum of 143 bytes to be transmitted before it can start transmitting, which will bring a large delay to the high-speed frame.

[0006] Therefore, there is an urgent need for a frame preemption optimization method to solve the problem of high-speed frame delay caused by preemptible frames when the frame preemption conditions are not met. Summary of the invention

[0007] The purpose of the present invention is to provide a time-sensitive network frame preemption optimization method combined with a time shaper, mainly to solve the problem of high-speed frame delay caused by preemptible frames when the frame preemption conditions are not met.

[0008] A time-sensitive network frame preemption optimization method combined with a time-aware shaper comprises the following steps:

[0009] S1: Set the value of the frame priority PCP in the frame structure of each frame to be transmitted;

[0010] S2: Allocate the frame to be transmitted to a queue of corresponding priority according to the value of the frame priority PCP;

[0011] The queue refers to: a first-in first-out queue set at the output port of a network communication node where a time-aware shaper TAS is deployed;

[0012] Each queue corresponds to a transmission gate, and the switch of the transmission gate is driven by the static gate list of the time-aware shaper;

[0013] S3: The frame to be transmitted with the highest frame priority PCP value is encapsulated as a time-sensitive frame by the high-speed MAC sublayer eMAC in the MAC layer defined in the IEEE802.3br standard into a high-speed frame, and the other frames in the frame to be transmitted except the time-sensitive frame are encapsulated into a preemptible frame by the preemptible MAC sublayer pMAC in the MAC layer defined in the IEEE 802.3br standard;

[0014] S4: Obtain the current time through the global clock of the time-aware shaper TAS, obtain the opening time of the time window corresponding to the next time-sensitive frame to be transmitted, calculate the transmittable frame size AllowedTransmitSize, and perform high-speed frame preemption based on the transmittable frame size AllowedTransmitSize.

[0015] Preferably, the S4 specifically includes the following sub-steps:

[0016] S41: When the MAC merging sublayer monitors that there is a preemptible frame to be transmitted or the frame buffer pool in the MAC merging sublayer is not empty, and the output port of the current network communication node is idle, the current time is obtained through the global clock of the time-aware shaper TAS, and the opening time of the time window corresponding to the next time-sensitive frame to be transmitted is obtained through the static gating list for planning the time window corresponding to the time-sensitive frame, and the transmittable frame size AllowedTransmitSize is calculated. The calculation method is as follows:

[0017] AllowedTransmitSize=(T next_window -T current )*PortTransmitRate

[0018] PortTransmitRate is the port bandwidth, AllowedTransmitSize is the transmittable frame size, and T next_window is the opening time of the next time-sensitive frame time window, T current is the current time;

[0019] The MAC Merge Subslayer is defined in the IEEE 802.3br standard.

[0020] S42: Obtain the size of the preemptible frame to be transmitted currently, where the preemptible frame to be transmitted currently is the preemptible frame monitored or the cached frame stored in the frame cache pool in the MAC merging sublayer;

[0021] If the transmittable frame size AllowedTransmitSize is not less than the preemptible frame to be transmitted, execute S43, otherwise execute S44;

[0022] S43: Transmit the preemptible frame currently to be transmitted;

[0023] S44: If the current transmittable frame size AllowedTransmitSize is not less than 84 bytes, the preemptible frame to be transmitted is transmitted. During the transmission of the preemptible frame, if the MAC merging sublayer monitors a frame preemption request, S45 is executed; if the current transmittable frame AllowedTransmitSize is less than 84 bytes, the preemptible frame to be transmitted is placed in the frame buffer pool of the MAC merging sublayer as a buffer frame, and S46 is executed;

[0024] S45: Slice the preemptible frame to obtain the first frame fragment, including a 7-byte preamble, a 1-byte preemptible frame start delimiter SMD-Sx, a 6-byte MAC source address MAC SA and a destination address MAC DA, a 2-byte Ethernet frame type EtherType, and a 0 to 1500-byte payload Data;

[0025] Afterwards, a 4-byte frame segment check code mCRC is added to the above segment for transmission, and then an inter-frame gap IFG time is waited for. The high-speed frame time window is opened by the static gating list, and the high-speed frame transmission begins. The remaining preemptible frame segments are placed in the frame buffer pool in the MAC merging sublayer as buffer frames. If the size of the remaining preemptible frame segments is less than 84 bytes, S46 is executed, otherwise S47 is executed.

[0026] The remaining subsequent segments of the preemptible frame include a middle segment of the preemptible frame and a tail segment of the preemptible frame. If there is only one preemption, the segment is a tail segment of the preemptible frame. If there are multiple preemptions, the segment is a middle segment of the preemptible frame or a tail segment of the preemptible frame.

[0027] S46: Determine whether the cached frame in the cache pool is not less than 84 bytes. If the cached frame in the cache pool is not less than 84 bytes, wait for the high-speed frame transmission to be completed and then return to S41. Otherwise, execute S47.

[0028] S47: Fill the cache frame to 84 bytes and return to S41.

[0029] Preferably, in S47, the cached frame is filled in the following manner: the cached frame includes two types of frames, one type is a preemptible frame that has not been sliced, and its frame structure includes a 7-byte preamble, a 1-byte preemptible frame start delimiter SMD-Sx, a 6-byte MAC source address MAC SA and a destination address MAC DA, a 2-byte Ethertype, a 0 to 1500-byte payload Data, and a 4-byte frame check mark FCS;

[0030] The other type is the subsequent fragments of the sliced ​​preemptible frame, including the middle fragment and the tail fragment of the preemptible frame. Its frame structure includes a 6-byte preamble, a 1-byte preemptible frame subsequent fragment delimiter SMD-Cx, a 1-byte fragment count Frag Count, 0 to 1500 bytes of payload Data, and a 4-byte frame fragment check mark mCRC or a frame check mark FCS;

[0031] The checksum in the frame structure is any one of the following two: a frame segment checksum mCRC and a frame checksum FCS. For the first frame segment obtained by preemptible frame slicing or the middle segment of the preemptible frame, mCRC is used as the checksum.

[0032] For the tail frame fragment obtained by slicing the preemptible frame, the unsliced ​​preemptible frame, and the high-speed frame, the FCS is used as the checksum;

[0033] A Padding field is set between the payload Data and the check character;

[0034] The size of the Padding filling field is between 0 and 42 bytes; cache frame filling is achieved by filling preset content into the Padding filling field.

[0035] The overhead of the frame preemption process and the frame gap (IFG) of the present invention are already included in the calculated transmittable frame, and the threshold of 84 bytes is the minimum frame size allowed to be transmitted by Ethernet.

[0036] The padding field of the present invention is a newly added field in the subsequent frame fragmentation structure based on the original IEEE802.3 standard, and the data range is 0 to 42 bytes. Figure 4 shown.

[0037] The time-aware shaper of the present invention complies with the IEEE802.1Qbv standard definition, and the frame preemption process complies with the IEEE802.1Qbu standard definition.

[0038] The beneficial effect of the present invention is that: by combining the time shaper and the frame preemption mechanism, the static gating list and the global clock are used to obtain the current time and the opening time of the next time-sensitive network frame time window, and the current transmittable frame size is calculated. Then, through the above processing flow, when the preemption condition is not met, the high-speed frame is prevented from being affected by the preemptible frame. At the same time, the time sensor combined with the frame preemption mechanism also reduces the problem of bandwidth waste caused by the protection band to a certain extent. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] In order to more clearly illustrate the specific implementation of the present invention or the technical solutions in the prior art, the following will briefly introduce the drawings required for the specific implementation or the prior art description. Some specific embodiments of the present invention will be described in detail in an exemplary but not restrictive manner with reference to the drawings. The same reference numerals in the drawings indicate the same or similar components or parts. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings:

[0040] Figure 1 This is an architectural diagram of the combination of the time sensor and the frame preemption mechanism.

[0041] Figure 2 This is a diagram of the high-speed frame and preemptible frame structure defined in the IEEE802.3br standard.

[0042] Figure 3 The slice result map of the preemptible frame.

[0043] Figure 4 Optimize the flow chart for frame preemption. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0045] S1: Set the value of the frame priority PCP in the frame structure of each transmission frame.

[0046] The transmission frames include time-sensitive frames and non-time-sensitive frames; the frame priority PCP is a field in the 802.1Q tag, which is introduced in the IEEE802.1Q standard. The frame priority PCP describes the priority corresponding to each transmission frame, and the priority range is 0 to 7. In the present invention, the transmission frames carrying control data in automobile or industrial control applications are regarded as time-sensitive frames, that is, the frame priority PCP corresponding to the transmission frames carrying control data is set to 7, and the transmission frames carrying other non-control data are regarded as non-time-sensitive frames (such as audio streams, video streams, best-effort streams, etc.). The setting range of the frame priority PCP corresponding to these different categories of non-time-sensitive frames is between 0 and 6.

[0047] S2: According to the IEEE802.1Qbv standard, each output port of each network communication node (including terminal devices and switches) that has deployed the time sensor TAS contains 8 first-in first-out queues, each queue has a corresponding priority ranging from 0 to 7, and each queue is associated with an output gate, such as Figure 1As shown in the figure, various transmission frames are mapped to queues with specified priorities. For example, if the priority of a time-sensitive frame is 7, then this type of transmission frame will be assigned to the queue with priority 7.

[0048] S3: According to the IEEE802.1Qbv standard, each network communication node that has deployed the time sensor TAS has a synchronized global clock, so the opening and closing of the output gate gate associated with each queue is controlled by a static gating list. Each static gating list contains the status of the output gate gate associated with each queue at different times. For example, 10000000 means that the output gate gate associated with queue 7 where the time-sensitive frame is located is opened, and the output gate gate associated with the queues where the other categories of non-time-sensitive frames are located is closed. If the output status is 01100000, it means that the output gate gate associated with queue 7 where the time-sensitive frame is located is closed, and the output gate gate associated with queues 6 and 5 is opened. At this time, there will be multiple non-time-sensitive frames of different categories, which need to be scheduled through the strict priority algorithm of the transmission selection layer.

[0049] S4: The time-sensitive frame is encapsulated into a high-speed frame by the high-speed MAC sublayer eMAC in the MAC layer, that is, the frame start delimiter SFD is modified to the high-speed frame start delimiter SMD-E. The non-time-sensitive frame is encapsulated into a preemptible frame by the preemptible MAC sublayer pMAC in the MAC layer, that is, the frame start delimiter SFD is modified to the preemptible frame start delimiter SMD-Sx, such as Figure 2 shown.

[0050] S5: If it is detected that there is a preemptible frame to be transmitted or the frame buffer pool in the MAC merging sublayer is not empty (if the frame buffer pool is not empty, the buffered frame is transmitted first), and the output port of the current network communication node is idle, then the current time is obtained through the global clock of the time-aware shaper TAS, and the time window corresponding to the next time-sensitive frame to be transmitted is obtained through the static gating list (such as Figure 1 The open time T1 or T2 shown in the figure is used to calculate the size of the transmittable frame AllowedTransmitSize. The calculation method is as follows, where PortTransmitRate is the port bandwidth, AllowedTransmitSize represents the transmittable frame size, and T next_window is the opening time of the next time-sensitive frame time window, T current is the current time

[0051] AllowedTransmitSize=(T next_window -T current )*PortTransmitRate

[0052] S6: Get the size of the preemptible frame to be transmitted (including the preemptible frame monitored or the cached frame stored in the frame buffer pool in the MAC merging sublayer). If the transmittable frame AllowedTransmitSize is not less than the preemptible frame to be transmitted, execute S7, otherwise execute S8.

[0053] S7: indicates that the preemptible frame currently to be transmitted will not affect the high-speed frame, so it is transmitted directly.

[0054] S8: indicates that the preemptible frame to be transmitted will affect the high-speed frame. If the AllowedTransmitSize of the current transmittable frame is not less than 84 bytes, it will be directly transmitted. During the transmission process, the preemption signal of the high-speed frame will be monitored and the frame preemption operation will be performed. The specific frame preemption operation is shown in S9. If the AllowedTransmitSize of the current transmittable frame is less than 84 bytes, S10 will be executed;

[0055] S9: Execute frame preemption operation, the preemptible frame is sliced, the first frame fragment includes 7 bytes of preemble, 1 byte of preemptible frame start delimiter SMD-Sx, 6 bytes of MAC source address MAC SA and destination address MAC DA, 2 bytes of Ethernet frame type EtherType, 0 to 1500 bytes of payload packet Data, followed by 4 bytes of frame fragment check mCRC and then transmitted, then wait for a frame gap IFG time, the high-speed frame time window is opened by static gating drive, and the high-speed frame is transmitted, and the remaining preemptible frame fragments are placed in the frame buffer pool in the MAC merging sublayer. If the fragment is less than 84 bytes, execute S10, otherwise execute S11;

[0056] S10: Put the preemptible frame or subsequent fragment of the preemptible frame to be transmitted into the frame buffer pool of the MAC merging sublayer, wait for the high-speed frame transmission to be completed, enter a new round of process and return to S5. Here, before entering a new round of process, it is necessary to determine whether the cached frame in the buffer pool is not less than 84 bytes. If the condition is met, no processing is performed, indicating that the cached frame size meets the Ethernet minimum frame requirement, otherwise execute S11.

[0057] S11: indicates that the cache frame size does not meet the Ethernet minimum frame requirement, so the cache frame is padded to 84 bytes through the Padding field between the payload Data and the frame fragment check CRC or frame check FCS in the frame structure. The size of the Padding field is between 0 and 42 bytes, and is 0 bytes by default. Figure 3 shown.

[0058] The remaining subsequent fragments of the preemptible frame include the middle fragment of the preemptible frame or the end fragment of the preemptible frame. If there is only one preemption, the fragment is the end fragment of the preemptible frame. If multiple preemptions occur, it may be the middle fragment of the preemptible frame or the end fragment of the preemptible frame.

[0059] The cached frames include two types of frames, one type is an unsliced ​​preemptible frame, whose frame structure includes a 7-byte preamble, a 1-byte preemptible frame start delimiter SMD-Sx, a 6-byte MAC source address MAC SA and a destination address MAC DA, a 2-byte Ethertype, 0 to 1500 bytes of payload Data, and a 4-byte frame check mark FCS; the other type is a sliced ​​preemptible frame subsequent fragment including a preemptible frame middle fragment and a tail frame fragment, whose frame structure includes a 6-byte preamble, a 1-byte preemptible frame subsequent fragment delimiter SMD-Cx, a 1-byte fragment count Frag Count, 0 to 1500 bytes of payload Data, and a 4-byte frame fragment check mark mCRC or a frame check mark FCS.

[0060] The checksum in the frame structure is any one of the following two: a frame segment checksum mCRC and a frame checksum FCS. For the first frame segment or the middle segment of the preemptible frame obtained by slicing the preemptible frame, mCRC is used as the checksum. For the last frame segment obtained by slicing the preemptible frame, the unsliced ​​preemptible frame, and the high-speed frame, FCS is used as the checksum.

[0061] The above time-aware shaper combined with frame preemption architecture is as follows Figure 1 As shown, each frame fragment after the frame slice can be preempted as follows Figure 3 The above description is only part of the specific implementation of the present invention, but the protection scope of the present invention is not limited thereto, and any changes or substitutions that can be easily thought of by any person familiar with the art within the technical scope disclosed by the present invention should be covered within the protection scope of the present invention.

Claims

1. A time-sensitive network frame preemption optimization method combined with a time-aware shaper, It is characterized in that The steps include: S1: Set the value of the frame priority PCP in the frame structure of each frame to be transmitted; S2: Allocate the frame to be transmitted to a queue of corresponding priority according to the value of the frame priority PCP; The queue refers to: a first-in first-out queue set at the output port of a network communication node where a time-aware shaper TAS is deployed; Each queue corresponds to a transmission gate, and the switch of the transmission gate is driven by the static gate list of the time-aware shaper; S3: The frame to be transmitted with the highest frame priority PCP value is encapsulated as a time-sensitive frame by the high-speed MAC sublayer eMAC in the MAC layer defined in the IEEE802.3br standard into a high-speed frame, and the other frames in the frame to be transmitted except the time-sensitive frame are encapsulated into a preemptible frame by the preemptible MAC sublayer pMAC in the MAC layer defined in the IEEE 802.3br standard; S4: obtaining the current time through the global clock of the time-aware shaper TAS, obtaining the opening time of the time window corresponding to the next time-sensitive frame to be transmitted, calculating the transmittable frame size AllowedTransmitSize, and performing high-speed frame preemption of the preemptible frame based on the transmittable frame size AllowedTransmitSize; The S4 specifically includes the following sub-steps: S41: When the MAC merging sublayer monitors that there is a preemptible frame to be transmitted or the frame buffer pool in the MAC merging sublayer is not empty, and the output port of the current network communication node is idle, the current time is obtained through the global clock of the time-aware shaper TAS, and the opening time of the time window corresponding to the next time-sensitive frame to be transmitted is obtained through the static gating list for planning the time window corresponding to the time-sensitive frame, and the transmittable frame size AllowedTransmitSize is calculated. The calculation method is as follows: AllowedTransmitSize=(T next_window -T current )*PortTransmitRate PortTransmitRate is the port bandwidth, AllowedTransmitSize is the transmittable frame size, and T next_window is the opening time of the next time-sensitive frame time window, T current is the current time; The MAC Merge Subslayer is defined in the IEEE 802.3br standard. S42: Obtain the size of the preemptible frame to be transmitted currently, where the preemptible frame to be transmitted currently is the preemptible frame monitored or the cached frame stored in the frame cache pool in the MAC merging sublayer; If the transmittable frame size AllowedTransmitSize is not less than the preemptible frame to be transmitted, execute S43, otherwise execute S44; S43: Transmit the preemptible frame currently to be transmitted; S44: If the current transmittable frame size AllowedTransmitSize is not less than 84 bytes, the preemptible frame to be transmitted is transmitted. During the transmission of the preemptible frame, if the MAC merging sublayer monitors a frame preemption request, S45 is executed; if the current transmittable frame AllowedTransmitSize is less than 84 bytes, the preemptible frame to be transmitted is placed in the frame buffer pool of the MAC merging sublayer as a buffer frame, and S46 is executed; S45: Slice the preemptible frame to obtain the first frame fragment, including a 7-byte preamble, a 1-byte preemptible frame start delimiter SMD-Sx, a 6-byte MAC source address MAC SA and a destination address MAC DA, a 2-byte Ethernet frame type EtherType, and a 0 to 1500-byte payload Data; Afterwards, a 4-byte frame segment check code mCRC is added to the above segment for transmission, and then an inter-frame gap IFG time is waited for. The high-speed frame time window is opened by the static gating list, and the high-speed frame transmission begins. The remaining preemptible frame segments are placed in the frame buffer pool in the MAC merging sublayer as buffer frames. If the size of the remaining preemptible frame segments is less than 84 bytes, S46 is executed, otherwise S47 is executed. The remaining subsequent segments of the preemptible frame include a middle segment of the preemptible frame and a tail segment of the preemptible frame. If there is only one preemption, the segment is a tail segment of the preemptible frame. If there are multiple preemptions, the segment is a middle segment of the preemptible frame or a tail segment of the preemptible frame. S46: Determine whether the cached frame in the cache pool is not less than 84 bytes. If the cached frame in the cache pool is not less than 84 bytes, wait for the high-speed frame transmission to be completed and then return to S41. Otherwise, execute S47. S47: Fill the cache frame to 84 bytes, and return to S41; In the S47, the cached frame is filled in the following manner: the cached frame includes two types of frames, one type is a preemptible frame that has not been sliced, and its frame structure includes a 7-byte preamble, a 1-byte preemptible frame start delimiter SMD-Sx, a 6-byte MAC source address MAC SA and a destination address MAC DA, a 2-byte Ethertype, a 0 to 1500-byte payload Data, and a 4-byte frame check mark FCS; The other type is the subsequent fragments of the sliced ​​preemptible frame, including the middle fragment and the tail fragment of the preemptible frame. Its frame structure includes a 6-byte preamble, a 1-byte preemptible frame subsequent fragment delimiter SMD-Cx, a 1-byte fragment count Frag Count, 0 to 1500 bytes of payload Data, and a 4-byte frame fragment check mark mCRC or a frame check mark FCS; The checksum in the frame structure is any one of the following two: a frame segment checksum mCRC and a frame checksum FCS. For the first frame segment obtained by preemptible frame slicing or the middle segment of the preemptible frame, mCRC is used as the checksum. For the tail frame fragment obtained by slicing the preemptible frame, the unsliced ​​preemptible frame, and the high-speed frame, the FCS is used as the checksum; A Padding field is set between the payload Data and the check character; The size of the Padding filling field is between 0 and 42 bytes; cache frame filling is achieved by filling preset content into the Padding filling field.

Images