A data alignment method in a latency-sensitive network

Abstract

The data alignment method in the time delay sensitive network provided by the application is to segment the flow according to the links in the path, each link is a segment, and each segment includes the leaving time of the data frame, the arriving time at the next node, and the scheduling time at the next node; the sending time of the data frame is the leaving time of the first segment; the arriving time of the data frame is the sum of the scheduling time of the last segment and the propagation delay of the last link; the leaving time, the arriving time and the scheduling time of the data frame are controlled through the gate control list (GCL); the data alignment method in the time delay sensitive network can make the data transmitted through different paths arrive at the destination node at the same time on the basis of guaranteeing the data transmission delay requirement.

Description

Technical Field

[0001] This invention belongs to the field of network data alignment, and particularly relates to a data alignment method in latency-sensitive networks. Background Technology

[0002] Time-Sensitive Networking (TSN) is a valuable technology in Industry 4.0, providing deterministic, low-latency, and highly reliable data transmission services for critical flows. TSN enables Ethernet technology to be used in critical systems with strict latency constraints. For example, in automotive environments, TSN protocols can precisely guarantee the time determinism of critical automotive Ethernet signals, and automotive Ethernet will play a crucial role in next-generation vehicle network architectures.

[0003] The problems considered in the design of traditional Ethernet are very different from the current network requirements. Traditional Ethernet provides a best-effort service for data transmission, which can no longer meet the stringent quality of service (QoS) requirements of emerging services. This poses a challenge to the use of Ethernet because it lacks support for deterministic behavior. However, the Time-Sensitive Networking (TSN) standard provides a solution, bringing real-time, low-latency, and deterministic communication to the Ethernet ecosystem.

[0004] Time-Sensitive Networking (TSN) defines a set of standards for adding deterministic low-latency services to Ethernet, with high reliability being a key characteristic. The IEEE 802.1CB protocol introduces the Frame Replication and Elimination for Reliability (FRER) mechanism to TSN networks. It transmits duplicate data frames through two disjoint paths, eliminating a duplicate at the destination node to improve service reliability. However, finding two disjoint paths between the source and destination nodes is not an easy task.

[0005] The frame copying and elimination function in TSN networks requires data frames to be transmitted through different paths. When the number of hops on the paths is equal, the difference in path delay is relatively small. When the number of hops on the paths is unequal, the difference in path delay is relatively large. For zero-jitter services and services with high jitter requirements, when a transmission path fails, the original frame copying and elimination function cannot meet the service jitter requirements. So, under the premise of meeting the service delay requirements, how can we ensure that data frames reach the destination node simultaneously in paths with equal or unequal hop counts? This patent proposes a data alignment method in a delay-sensitive network. Summary of the Invention

[0006] Objective: To address the above problems, this invention proposes a data alignment method for latency-sensitive networks. This method calculates the disjoint paths in the Frame Replication and Elimination for Reliability (IEEE 802.1CB) protocol based on traffic quality of service requirements, such as... Figure 1 The network structure diagram shows that it includes 6 TSN switches (switch1, switch2a, switch2b, switch23b, switch3a, switch3b), 1 synSource node, and 1 destination node. The synSource node uses the Sequence encode function to encode sequence numbers into data packets, and uses the Stream splitting function to copy and distribute data packets. Redundant data is transmitted through path 1: synSource-switch1-switch2a-switch3a-destination and path 2: synSource-switch1-switch2b-switch23b-switch3b-destination. At the destination node, the Sequence decode function is used to decode data packets, and the Sequence recovery function is used to recover and eliminate duplicates of data packets.

[0007] Technical Solution: To achieve the objectives of this invention, a data alignment method for delay-sensitive networks is proposed, comprising the following steps:

[0008] (1) Connect the underlying network devices to the controller, and the controller builds a global view of the underlying network based on the device information;

[0009] (2) When a data stream is being sent in the network, the sending node sends the traffic request information to the controller. The controller combines the traffic request information and the underlying network global view to calculate two non-overlapping stream transmission paths from the sending node to the receiving node.

[0010] (3) The two stream transmission paths calculated by the controller are segmented according to the links, with each link being a stream segment, such as... Figure 2 As shown, the stream segment includes scheduling time, arrival time, and scheduling time information. The connection rule between stream segments is that the scheduling time of the previous stream segment is equal to the sending time of the next stream segment.

[0011] (4) Add the flow information and flow segment information to the solver Z3, and add the data flow collision-free constraint, end-to-end delay constraint, and data frame alignment constraint to the solver Z3. Use the solver Z3 to solve the constraints and obtain the time slot information required when the flow segment forwards data.

[0012] (5) The controller obtains the scheduling time, arrival time and scheduling time of the data stream in the corresponding stream segment by calling the solution results of the Z3 solver. The controller configures the gating list of the underlying data forwarding switch according to the solution results through the netconf protocol.

[0013] (6) The data stream forwards data along the data path and in the corresponding time slot in the gating list setting. The receiving node simultaneously receives multiple replica data packets sent through different forwarding paths.

[0014] Furthermore, the method for step (1) is as follows:

[0015] Start the centralized controller that configures network resources in the Internet and configure the underlying TSN network forwarding devices, so that the network forwarding devices are connected to the controller. The controller obtains a global view of the underlying network through the LLDP protocol, denoted as G = (V, E), where V represents the set of nodes, the edges connecting nodes in the network are represented by the set E, and the edges connecting nodes include data links in two directions, denoted by the set L.

[0016]

[0017] In equation (1), (v1, v2) represents the undirected link connection between nodes, [v1, v2] represents the directed data link from node v1 to node v2, and [v2, v1] represents the directed data link from node v2 to node v1.

[0018] Furthermore, the method for step (2) is as follows:

[0019] The set of data streams is denoted by F, and the data streams in the network are denoted by f. i ∈F represents, where i represents the data stream number, and the description information of the data stream is represented as:

[0020] f i ={f i .period,f i .offset,f i .length, f i .limitedDelay} (2)

[0021] In equation (2), f i .period represents the data packet sending period, f i.offset represents the sending offset of the data packet, f i .length represents the size of the data packet to be sent, f i .limitedDelay represents the delay requirement for data packets;

[0022] The data stream transmission path is represented by a set P, and the data stream f i Path f ∈ ​​F i (path1),f i (path2) is calculated by the controller based on traffic information and the underlying network topology. i (path1) represents the data stream f i The first path ∈ F, f i (path2) represents the data stream f i The second path ∈F, path f i (path1),f i (path2) consists of multiple directed links, collectively represented as follows:

[0023] {[v1, v2], [v2, v3], ..., [v x-1 v x ],...,[v n-1 v n ]} (3)

[0024] In equation (3), [v x-1 v x ]∈L,[v x-1 v x ] indicates from v x-1 node to v x There are directed data links between them, and multiple directed links constitute the transmission path of the service data flow.

[0025] Furthermore, the method for step (3) is as follows:

[0026] (3.1) When aligning data streams, the streams first need to be segmented according to link information. A schematic diagram of the stream segments is shown below. Figure 2 As shown, each data flow path is divided into multiple flow segments. In a flow segment, dt represents the departure time of a data packet, at represents the arrival time of a data packet, and st represents the scheduling time of a data packet. The set of data flow segments is denoted by SE. Represents data stream f i ∈F in link [v x-1 v x If a stream segment ∈ L is a stream segment, then the stream segment is a stream segment. The departure time is The time to reach the next node is The time it takes for the service data packet to be scheduled on the next node is

[0027] (3.2) Connections between service flow segments are achieved by setting the scheduling time of the previous node to equal the departure time of the next node. Any data flow f i In the link between two adjacent wafers [v m v x ]、[v x v y When transferring data, the stream segments need to be interconnected:

[0028]

[0029] In equation (4), Represents data stream f i ∈F in link [v m v x Stream segments on ]∈L Represents data stream f i ∈F in link [v x v y Stream segments on ]∈L Represented as a stream segment The scheduling time, Represented as a stream segment departure time, This means that the scheduling time of the first stream segment is equal to the departure time of the second stream segment in two adjacent stream segments;

[0030] Furthermore, the method for step (4) is as follows:

[0031] (1) No collision constraint

[0032] Two different data streams are represented by f i and f j This means that on any link Only one data frame from a data stream is allowed to be sent at a time; otherwise, a collision will occur.

[0033]

[0034] In equation (5), v m Indicates the head node of the link, v x f represents the tail node of the link. i .period represents the data stream f i Data packet sending cycle, using Represents data stream f i ∈F via link [v m v x ], Represents data stream f i Data packets in link [v m v x The sending offset, f i .length represents the data stream f i The size of the data packet, f j .period represents the data stream f j Data packet sending cycle, using Represents data stream f j After passing through link [v m v x ], Represents data stream f j Data packets in link [v m v x The sending offset, f j .length represents the data stream f j The packet size is F.period, which represents the period of all traffic. LCM(F.period) is the least common multiple of all service data flow periods. α and β represent the number of flow periods included in the least common multiple of all service data flow periods.

[0035] (2) End-to-end delay constraints

[0036] Any data stream f i ,use Represents data stream f i The last stream segment tail node to data stream f i destination node f i The link propagation delay of the destination, with end-to-end delay constraints, is as follows:

[0037]

[0038] In equation (6), Represents data stream f i The first stream segment, Represents data stream f i The departure time of the first stream segment, Represents data stream f i The last stream segment, Represents data stream f i The scheduling time of the last stream segment, For data stream f i The last stream segment The tail node, f i .destination is the data stream fi The destination node, f i .limitedDelay represents the data stream f i End-to-end latency requirements;

[0039] (3) Data frame alignment constraints

[0040] The data frame alignment constraint from the sending node to the destination node is:

[0041]

[0042] In equation (7), Path(f i ) represents the data stream f i All transmission paths, including f i (path1),f i (path2) two paths, f i (path q ) represents the data stream f i The q-th path from the sender to the destination, where q = 1 or 2;

[0043] Represents data stream f i The tail node of the last flow segment in the q-th path, f i .destination is the data stream f i The destination node Represents data stream f i The tail node of the last stream segment in the q-th path to data stream f i destination node f i The propagation delay of the .destination link. Represents data stream f i The scheduling time of the last stream segment in the q-th path.

[0044] Beneficial effects: Compared with the prior art, the technical solution of the present invention has the following beneficial technical effects:

[0045] The frame duplication and elimination mechanism of the IEEE 802.1CB protocol in latency-sensitive networks has been enhanced, so that copies of data frames transmitted through different paths arrive at the receiving node simultaneously. When a transmission path fails and the received data path is switched, zero jitter can still be guaranteed. This enables latency-sensitive networks to support zero-jitter services or services with high jitter requirements, further enhancing the reliability of latency-sensitive networks. Attached Figure Description

[0046] Figure 1 This is a schematic diagram of the network structure;

[0047] Figure 2 This is a schematic diagram of a flow segment. Detailed Implementation

[0048] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0049] This invention proposes a data alignment method in delay-sensitive networks, which includes the following steps:

[0050] (1) Connect the underlying network devices to the controller, and the controller builds a global view of the underlying network based on the device information;

[0051] (2) When a data stream is being sent in the network, the sending node sends the traffic request information to the controller. The controller combines the traffic request information and the underlying network global view to calculate two non-overlapping stream transmission paths from the sending node to the receiving node.

[0052] (3) The two stream transmission paths calculated by the controller are segmented according to the links, with each link being a stream segment, such as... Figure 2 As shown, the stream segment includes scheduling time, arrival time, and scheduling time information. The connection rule between stream segments is that the scheduling time of the previous stream segment is equal to the sending time of the next stream segment.

[0053] (4) Add the flow information and flow segment information to the solver Z3, and add the data flow collision-free constraint, end-to-end delay constraint, and data frame alignment constraint to the solver Z3. Use the solver Z3 to solve the constraints and obtain the time slot information required when the flow segment forwards data.

[0054] (5) The controller obtains the scheduling time, arrival time and scheduling time of the data stream in the corresponding stream segment by calling the solution results of the Z3 solver. The controller configures the gating list of the underlying data forwarding switch according to the solution results through the netconf protocol.

[0055] (6) The data stream forwards data along the data path and in the corresponding time slot in the gating list setting. The receiving node simultaneously receives multiple replica data packets sent through different forwarding paths.

[0056] Furthermore, the method for step (1) is as follows:

[0057] Start the centralized controller that configures network resources in the Internet and configure the underlying TSN network forwarding devices, so that the network forwarding devices are connected to the controller. The controller obtains a global view of the underlying network through the LLDP protocol, denoted as G = (V, E), where V represents the set of nodes, the edges connecting nodes in the network are represented by the set E, and the edges connecting nodes include data links in two directions, denoted by the set L.

[0058]

[0059] In equation (1), (v1, v2) represents the undirected link connection between nodes, [v1, v2] represents the directed data link from node v1 to node v2, and [v2, v1] represents the directed data link from node v2 to node v1.

[0060] Furthermore, the method for step (2) is as follows:

[0061] The set of data streams is denoted by F, and the data streams in the network are denoted by f. i ∈F represents, where i represents the data stream number, and the description information of the data stream is represented as:

[0062] f i ={f i .period,f i .offset,f i .length, f i .limitedDelay} (2)

[0063] In equation (2), f i .period represents the data packet sending period, f i .offset represents the sending offset of the data packet, f i .length represents the size of the data packet to be sent, f i .limitedDelay represents the delay requirement for data packets;

[0064] The data stream transmission path is represented by a set P, and the data stream f i Path f ∈ ​​F i (path1),f i (path2) is calculated by the controller based on traffic information and the underlying network topology. i (path1) represents the data stream f i The first path ∈ F, f i (path2) represents the data stream f i The second path ∈F, path f i (path1),f i (path2) consists of multiple directed links, collectively represented as follows:

[0065] {[v1, v2], [v2, v3], ..., [v x-1 v x ],...,[v n-1 v n ]} (3)

[0066] In equation (3), [v x-1 v x ]∈L,[v x-1 v x ] indicates from v x-1 node to v x There are directed data links between them, and multiple directed links constitute the transmission path of the service data flow.

[0067] Furthermore, the method for step (3) is as follows:

[0068] (3.1) When aligning data streams, the streams first need to be segmented according to link information. A schematic diagram of the stream segments is shown below. Figure 2 As shown, each data flow path is divided into multiple flow segments. In a flow segment, dt represents the departure time of a data packet, at represents the arrival time of a data packet, and st represents the scheduling time of a data packet. The set of data flow segments is denoted by SE. Represents data stream f i ∈F in link [v x-1 v x If a stream segment ∈ L is a stream segment, then the stream segment is a stream segment. The departure time is The time to reach the next node is The time it takes for the service data packet to be scheduled on the next node is

[0069] (3.2) Connections between service flow segments are achieved by setting the scheduling time of the previous node to equal the departure time of the next node. Any data flow f i In the link between two adjacent wafers [v m v x ]、[v x v y When transferring data, the stream segments need to be interconnected:

[0070]

[0071] In equation (4), Represents data stream f i ∈F in link [v m v x Stream segments on ]∈L Represents data stream f i ∈F in link [v x v y Stream segments on ]∈L Represented as a stream segment The scheduling time, Represented as a stream segment departure time, This means that the scheduling time of the first stream segment is equal to the departure time of the second stream segment in two adjacent stream segments;

[0072] Furthermore, the method for step (4) is as follows:

[0073] (1) No collision constraint

[0074] Two different data streams are represented by f i and f j This means that on any link Only one data frame from a data stream is allowed to be sent at a time; otherwise, a collision will occur.

[0075]

[0076] In equation (5), v m Indicates the head node of the link, v x f represents the tail node of the link. i .period represents the data stream f i Data packet sending cycle, using Represents data stream f i ∈F via link [v m v x ], Represents data stream f i Data packets in link [v m v x The sending offset, f i .length represents the data stream f i The size of the data packet, f j .period represents the data stream f j Data packet sending cycle, using Represents data stream f j After passing through link [v m v x ], Represents data stream f j Data packets in link [v m v x The sending offset, f j .length represents the data stream f j The packet size is F.period, which represents the period of all traffic. LCM(F.period) is the least common multiple of all service data flow periods. α and β represent the number of flow periods included in the least common multiple of all service data flow periods.

[0077] (2) End-to-end delay constraints

[0078] Any data stream f i ,use Represents data stream f i The last stream segment tail node to data stream f i destination node f i The link propagation delay of the destination, with end-to-end delay constraints, is as follows:

[0079]

[0080] In equation (6), Represents data stream f i The first stream segment, Represents data stream f i The departure time of the first stream segment, Represents data stream f i The last stream segment, Represents data stream f i The scheduling time of the last stream segment, For data stream f i The last stream segment The tail node, f i .destination is the data stream f i The destination node, f i .limitedDelay represents the data stream f i End-to-end latency requirements;

[0081] (3) Data frame alignment constraints

[0082] The data frame alignment constraint from the sending node to the destination node is:

[0083]

[0084] In equation (7), Path(f i ) represents the data stream f i All transmission paths, including f i (path1),f i (path2) two paths, f i (path q ) represents the data stream f i The q-th path from the sender to the destination, where q = 1 or 2;

[0085] Represents data stream f i The tail node of the last flow segment in the q-th path, f i.destination is the data stream f i The destination node Represents data stream f i The tail node of the last stream segment in the q-th path to data stream f i destination node f i The propagation delay of the .destination link. Represents data stream f i The scheduling time of the last stream segment in the q-th path.

Claims

1. A data alignment method in a time-delay-sensitive network, characterized in that, The method includes the following steps: (1) Connect the underlying network devices to the controller, and the controller constructs a global view of the underlying network based on the device information; (2) When a data stream is being sent in the network, the sending node sends the traffic request information to the controller. The controller combines the traffic request information and the underlying network global view to calculate two non-intersecting stream transmission paths from the sending node to the receiving node. (3) The two stream transmission paths calculated by the controller are segmented according to the links. Each link is a stream segment. The stream segment includes scheduling time, arrival time, and scheduling time information. The connection rule between stream segments is that the scheduling time of the previous stream segment is equal to the sending time of the next stream segment. (4) Add the flow information and flow segment information to the Z3 solver, and add the data flow collision-free constraint, end-to-end delay constraint, and data frame alignment constraint to the Z3 solver. Use the Z3 solver to solve the constraints and obtain the time slot information required when the flow segment forwards data. (5) The controller obtains the scheduling time, arrival time and scheduling time of the data stream in the corresponding stream segment by calling the solution results of the Z3 solver. The controller configures the gating list of the underlying data forwarding switch according to the solution results through the netconf protocol. (6) The data stream forwards data along the data path and in the corresponding time slot in the gating list setting, and the receiving node simultaneously receives multiple replica data packets sent through different forwarding paths; The method for step (2) is as follows: the data stream set is represented by F, and the data streams in the network are represented by... This indicates that, where i represents the data stream number, the description information for the data stream is represented as follows: （2） In equation (2), Represents the data packet sending period. Represents the sending offset of the data packet. Represents the size of the data packet sent. Represents the latency requirement for data packets; The data stream transmission path is represented by a set P, and the data stream... path It is obtained by the controller combining traffic information and the underlying network topology. Represents data stream The first path, Represents data stream The second path, and each path consists of multiple directed links, each of which is represented as follows: （3） In equation (3), , indicating from Node to Directed data links between them; multiple directed links constitute the transmission path of the service data stream. The method for step (3) is as follows: (3.1) When aligning data streams, the streams are first segmented according to link information. Each data stream path is divided into multiple stream segments. In a stream segment, dt represents the departure time of the data packet, at represents the arrival time of the data packet, and st represents the scheduling time of the data packet. The set of data stream segments is represented by SE. use Represents data stream In the link The stream fragment on the above, then the stream fragment The departure time is The time to reach the next node is The service data packet is scheduled to be on the next node in a time interval of 1000 minutes. ; (3.2) Connections between service flow segments are achieved by setting the scheduling time of the previous node to be equal to the departure time of the next node. Any data flow In the link between two adjacent wafers , When data is transferred, the stream segments are interconnected; （4） In equation (4), Represents data stream In the link The stream segment on the above, Represents data stream In the link The stream segment on the above, Represented as a stream segment The scheduling time, Represented as a stream segment departure time, This means that the scheduling time of the first stream segment is equal to the departure time of the second stream segment in two adjacent stream segments.

2. The data alignment method in a time-delay sensitive network according to claim 1, characterized in that, The method for step (1) is as follows: Start the centralized controller that configures network resources in the Internet, and configure the underlying TSN network forwarding device, so that the network forwarding device is connected to the controller. The controller obtains a global view of the underlying network through the LLDP protocol. Let V represent the set of nodes, E represent the set of edges connecting nodes in the network, and L represent the data links in both directions of the edges connecting nodes. （1） In equation (1) This represents an undirected link connection between nodes. Indicates from node To the node Directed data links, Indicates from node To the node Directed data links.

3. The data alignment method in a time-delay sensitive network according to claim 1, characterized in that, The method for step (4) is as follows: (1) No collision constraint Two different data streams and This means that on any link Only one data frame from a data stream is allowed to be sent at a time; otherwise, a collision occurs. （5） In equation (5), Indicates the head node of the link. Indicates the tail node of the link. Represents data stream Data packet sending cycle, using Represents data stream via link , Represents data stream Data packets in the link The sending offset, Represents data stream Data packet size, Represents data stream Data packet sending cycle, using Represents data stream via link , Represents data stream Data packets in the link The sending offset, Represents data stream Data packet size, Indicates the period of all traffic. It is the least common multiple of the cycles of all service data streams; (2) End-to-end delay constraints Any data stream ,use Represents data stream The last stream segment tail node To data stream destination node Given the link propagation delay, the end-to-end delay constraint is: （6） In equation (6), Represents data stream The first stream segment, Represents data stream The departure time of the first stream segment, Represents data stream The last stream segment, Represents data stream The scheduling time of the last stream segment, For data stream The last stream segment The tail node, For data stream The destination node Represented as a data stream End-to-end delay; (3) Data frame alignment constraints The data frame alignment constraint from the sending node to the destination node is: （7） In equation (7), Represents data stream All transmission paths, including Two paths, Represents data stream From sender to destination There are 2 paths, q=1 or 2; Represents data stream The The tail node of the last flow segment of the path, For data stream The destination node Represents data stream In the first The tail node of the last flow segment in the path To data stream destination node The link propagation delay, Represents data stream In the first The scheduling time of the last stream segment in the path.

Images