A comprehensive optimization method for time-sensitive network transmission based on optimization modulus theory

Abstract

The present invention belongs to the field of network transmission technology, and specifically relates to a comprehensive optimization method for time-sensitive network traffic transmission. The present invention comprehensively optimizes the routing and scheduling aspects in the time-sensitive network, and mainly includes the following steps: first, the input traffic is classified, and the size of their data packets, nodes and other information are obtained, and then the network topology is input, and then the route of each flow is optimized and solved based on this information to achieve the purpose of load balancing of the overall network topology, and then the gating list is optimized and solved, and finally the comprehensive optimization purpose of the overall transmission is completed. The present invention can comprehensively optimize the traffic in the time-sensitive network from routing to scheduling, and can be applied to some industrial Internet scenarios, etc.

Description

Technical Field

[0001] The present invention belongs to the field of network traffic transmission, and in particular relates to a comprehensive optimization method for traffic transmission in a time-sensitive network. Background Art

[0002] Time-Sensitive Networking (TSN) is a subtask of the IEEE 802.1 working group, established in 2012. This Ethernet technology is open, open source, and standardized. Based on the standard Ethernet architecture, it achieves high-quality co-network transmission of multiple service flows by enhancing the flow scheduling capability, and provides highly reliable and bounded low-latency flow transmission services. In some areas with high service quality requirements, a very low-latency network connection is required, and here TSN can play the role of establishing a reliable transmission mechanism to achieve good compatibility between real-time critical data flows and ordinary data flows in the same network, and to achieve the low latency and low jitter required by key services.

[0003] Currently, there are many studies on the generation of gated lists, most of which consider low-priority traffic transportation as the optimization direction, and have not achieved a more comprehensive optimization process. Summary of the invention

[0004] The present invention aims to provide a comprehensive method for optimizing the efficiency of traffic transmission in a time-sensitive network, with the premise of ensuring low latency for high-priority traffic and the improvement goal of improving the transmission effect of low-priority traffic, ultimately achieving an improvement in the overall traffic transmission efficiency.

[0005] The technical solution to achieve the purpose of the present invention is:

[0006] A comprehensive optimization method for traffic transmission in a time-sensitive network, characterized in that it specifically includes the following steps:

[0007] Step (1): classify the traffic input into the time-sensitive network and various information such as the size and arrival time of the data packet of the time-triggered traffic;

[0008] Step (2): Based on the optimization module theory as a solution method, the relevant information of step (1) and the network topology are used as input, and the routing constraints are mathematically expressed to solve the optimal route;

[0009] Step (3): Based on the route obtained in step (2), the gating list is optimized and solved, and finally the comprehensive optimization of the overall transmission is achieved.

[0010] Compared with the existing related technologies, the present invention has the following significant advantages:

[0011] (1) We propose a formulation for the comprehensive routing and scheduling problem based on OMT in TSN.

[0012] (2) We optimize reliability-aware routing, first to satisfy the transmission reliability of different types of traffic, and then to optimize the overall load balancing.

[0013] (3) We propose utility functions for different types of traffic to accurately describe their transmission goals, and solve it to produce more efficient GCL. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 The figure is a flow chart of the comprehensive optimization method of the present invention.

[0015] Figure 2 The network topology.

[0016] Figure 3 Physical structure of the time-aware shaper in TSN.

[0017] Figure 4 is the utility function of different types of traffic. DETAILED DESCRIPTION

[0018] The present invention is further described below in conjunction with the accompanying drawings and specific embodiments, wherein the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions.

[0019] like Figure 1 As shown, the overall flow chart of a comprehensive optimization method and system for traffic transmission in a time-sensitive network according to an embodiment of the present invention includes the following steps:

[0020] Step (1): In the IEEE 802.1Q standard, flows in time-sensitive networks are assigned quality of service attributes. Using this feature, flows are defined by the PCP (priority code point filed) field and the VID (VLAN ID) in the Ethernet packet header. The standard allows switches to identify and classify different types of traffic, and CDT (Control-Data Traffic) traffic is periodic, thereby obtaining relevant information such as packet size and arrival time.

[0021] Step (2): Based on the optimization module theory as a solution method, the relevant information of step (1) and the network topology are used as input, and the routing constraints are mathematically expressed to solve the optimal route;

[0022] Step (201): mathematically represent the routing constraints in the time-sensitive network, including the source point constraint, the end point constraint, the intermediate node constraint and the loop avoidance constraint, which are the basic constraints for routing selection in the network, and the reliability constraint for ensuring the transmission success rate.

[0023] For each flow in the network, it must start from one end node and end at another end node. Therefore, the flow will only leave and not enter the source node. The source point constraint is mathematically expressed as follows:

[0024]

[0025] Where S is the set of all flows, V is the set of all nodes in the network, E is the set of all physical links, and s i .src is the stream i The source node.

[0026] For each flow, its route must reach the destination node. Therefore, similar to the previous constraint, the destination node of the route has only entry and no exit. The mathematical expression of the destination constraint is as follows:

[0027]

[0028] where s i .dest is stream s i The end point.

[0029] In addition to the source and destination nodes, each flow will also pass through intermediate nodes. There should be two cases for intermediate nodes. If the flow passes through an intermediate node, then there should be one input and one output. If the flow does not pass through the node, then there should be no input and no output. The mathematical representation of the intermediate node constraint is as follows:

[0030]

[0031] where v a , v b For flow i Remove the intermediate nodes between the source and destination.

[0032] When looking for the correct route, it goes without saying that a link should not be passed repeatedly and loops should be avoided. Therefore, we add constraints that make it impossible for all flows to be routed through loops. The loop avoidance constraint is mathematically expressed as follows:

[0033]

[0034] In the network, link failures and channel errors are inevitable. Therefore, we define the link failure rate as the probability of data loss in the traffic transmitted through the link. Further, we define the transmission reliability of the traffic as the product of the link efficiency (1 minus the link failure rate) of all the links through which the traffic passes. However, all types of traffic have reliability requirements for transmission, that is, the frame is correctly transmitted from the source node to the destination node, and the transmission reliability rate is greater than a certain value. The reliability constraint is mathematically expressed as follows:

[0035]

[0036] Where [v a , v b ].p is the link [v a , v b ] data loss rate, p i For flow i Reliability requirements should be met.

[0037] Step (202): Taking the load balancing of the overall routing as the optimization goal, the optimization problem is formalized as follows:

[0038] The optimization problem constructed is as follows:

[0039]

[0040] The optimization goal is to make the overall network load more balanced while achieving reliable transmission, so that no single physical link will be overloaded and the overall network will be more balanced.

[0041] Step (203): Based on the optimization module theory as a solution method, the mathematical expression of step (201) and the optimization problem constructed in step (202) are combined, and the optimal route is solved using step (1) as data input.

[0042] Step (3): Based on the satisfiability modulo theory, the relevant information of step (1) is used as input, and the network constraints are mathematically represented to solve the optimal gating list.

[0043] Step (301): mathematically represent the scheduling constraints in the time-sensitive network, including the basic constraints of deterministic Ethernet, such as frame constraints, link constraints, flow transmission constraints and end-to-end constraints, and the specific constraints of the time-sensitive network, such as flow isolation constraints, frame isolation constraints and full constraints for high-criticality flows.

[0044] The frame offset of any frame scheduled in the network must be greater than or equal to 0. In addition, the entire transmission window (offset plus frame duration) must be within the frame period. The frame constraint is mathematically expressed as follows:

[0045]

[0046]

[0047] Where S is the set of all flows, E is the set of all links, [v a , v b ] is the physical link, f is the frame instance, φ is the frame offset, L is the frame transmission duration, and T is the frame period.

[0048] Two frames routed through the same physical link in the network cannot overlap in the time domain. The link constraint is mathematically expressed as follows:

[0049]

[0050]

[0051]

[0052]

[0053]

[0054]

[0055]

[0056] in For i and j super cycle.

[0057] The propagation of frames of a flow must follow the order along the routing path of the flow. The flow transmission constraint is mathematically expressed as follows:

[0058]

[0059]

[0060]

[0061] where d is the physical delay on the transmission medium and δ is the worst-case difference between the local clocks of two synchronized devices.

[0062] The maximum end-to-end latency constraint specifies that the difference between the arrival and delivery times of a flow must be less than or equal to the specified maximum value. The end-to-end constraint is mathematically expressed as follows:

[0063]

[0064] Among them [s i .src,v a] is the first link of the flow transmission, [v b ,s i .dest] indicates the last link before the receiving node, and deadline is the maximum allowed end-to-end waiting time.

[0065] In reality, frame losses may occur, periodic payload sizes may vary over time, or frames may be lost due to, for example, frame loss. In these cases, there may be situations where deterministic behavior on the egress port cannot be guaranteed. Consider a situation where two frames from two different flows are scheduled to arrive at the switch one after the other. Both are placed in the same queue in the required order. At a later point in time, after both are placed in the queue, the timing gate of the queue is opened for the first frame (belonging to the first flow) and only at a later time for the second frame. If the first frame is lost, the frame from the second flow will take a place in the queue and be transmitted in the time slot originally reserved for the first flow, resulting in uncertainty and jitter. To guarantee isolation in real networks, we first consider isolating the flows in the time domain, assuming that they share the same egress queue. Therefore, if a frame of a given flow enters the queue, frames of other flows may not arrive in the queue until all frames of the previous flow have completely arrived. The flow isolation constraint is mathematically expressed as follows:

[0066]

[0067]

[0068]

[0069]

[0070]

[0071]

[0072]

[0073] in is the first frame of the stream, The last frame of the stream.

[0074] A frame from another stream can only enter a queue if an already queued frame of the initial stream has already been serviced. We enforce the following: if two frames come from different streams, one frame can be scheduled to arrive at a shared queue only if another frame has been dispatched from that queue. The frame isolation constraint is mathematically expressed as follows:

[0075]

[0076]

[0077]

[0078]

[0079]

[0080]

[0081]

[0082] The above applies specifically to frames in the same queue.

[0083] Step (302): Under the premise of ensuring low latency of TT traffic, the transmission effect of AVB traffic is improved as the optimization goal, and the optimization problem is formalized as follows:

[0084] The optimization problem constructed is as follows:

[0085]

[0086] The optimization goal is to minimize the effect value. Under the premise of ensuring the delay requirements of high-priority services, this goal can effectively improve the transmission utilization of AVB services.

[0087] Step (303): Based on the optimization module theory as a solution method, the mathematical expression of step (301) and the optimization problem constructed in step (302) are combined, and step (1) is used as the data input to solve the optimal gating list.

Claims

1. A comprehensive optimization method for traffic transmission in a time-sensitive network, characterized in that: The method first obtains traffic-related information, and then optimizes the overall transmission of traffic based on the information, ultimately achieving the purpose of improving transmission efficiency, and specifically includes the following steps: Step 1: Classify the traffic input into the time-sensitive network and obtain the size and arrival time of the data packet of the time-triggered traffic; Step 2: Solve based on the optimization module theory, take the relevant information of step 1 and the network topology as input, and mathematically represent the routing constraints to solve the optimal route; The specific steps include: Step 201: mathematically represent the routing constraints in the time-sensitive network, including source point constraints, end point constraints, intermediate node constraints, loop avoidance constraints, and reliability constraints for ensuring transmission success rate; The mathematical representation of the source point constraint is as follows: Where S is the set of all flows, V is the set of all nodes in the network, E is the set of all physical links, and s i .src is the stream i The source node of The mathematical representation of the endpoint constraint is as follows: where s i .dest is stream s i The end point; The mathematical expression of the intermediate node constraint is as follows: where v a , v b For flow i Remove the intermediate nodes between the source and the destination; The cycle avoidance constraint is mathematically expressed as follows: The reliability constraint is mathematically expressed as follows: Where [v a ,v b ].p is the link [v a ,v b ] data loss rate, p i For flow i Reliability requirements that should be met; Step 202: Taking the load balancing of the overall routing as the optimization goal, the optimization problem is formalized as follows: The optimization problem constructed is as follows: Step 203: Based on the optimization module theory as a solution method, the mathematical expression of step 201 and the optimization problem constructed in step 202 are combined, and the optimal route is solved using step 1 as data input; Step 3: Based on the route obtained in step 2, the gating list is optimized and solved, and finally the comprehensive optimization of the overall transmission is completed, which specifically includes the following steps: Step 301: mathematically represent the scheduling constraints in the time-sensitive network, including frame constraints, link constraints, flow transmission constraints and end-to-end constraints, as well as flow isolation constraints and frame isolation constraints; The frame constraint is mathematically expressed as follows: Where S is the set of all flows, E is the set of all links, [v a ,v b ] is the physical link, f is the frame instance, φ is the frame offset, L is the frame transmission duration, and T is the frame period; The link constraint is mathematically expressed as follows: Among them, hp i j = lcm(s i .T,s j .T), is s i and j The super cycle; The flow transmission constraint is mathematically expressed as follows: Where d is the physical delay on the transmission medium and δ is the worst-case difference between the local clocks of two synchronized devices; The end-to-end constraint is mathematically expressed as follows: Among them [s i .src,v a ] is the first link of the flow transmission, [v b ,s i .dest] indicates the last link before the receiving node, and deadline is the maximum allowed end-to-end waiting time; The mathematical representation of the flow isolation constraint is as follows: in is the first frame of the stream, is the last frame of the stream; The frame isolation constraint is mathematically expressed as follows: Step 302: Under the premise of ensuring low latency of TT traffic, the transmission effect of AVB traffic is improved as the optimization goal, and the optimization problem is formalized as follows: The optimization problem constructed is as follows: Step 303: Based on the optimization module theory as a solution method, the mathematical expression of step 301 and the optimization problem constructed in step 302 are combined, and step 1 is used as the data input to solve the optimal gating list; Finally, the optimization is completed, achieving the goal of improving the overall traffic transmission efficiency.

2. The comprehensive optimization method for traffic transmission in a time-sensitive network according to claim 1, characterized in that: Step 1 further comprises: In the IEEE 802.1Q standard, flows in time-sensitive networks are given quality of service attributes. Using this feature, flows are defined by the priority code point filed in the Ethernet packet header, namely the PCP field and the VLAN ID, namely the VID. The standard allows switches to identify and classify different types of traffic, and the Control-DataTraffic, namely the CDT traffic, is periodic, thereby obtaining information related to the packet size and arrival time.

Images