Deterministic network data flow priority control method
Optimizing the transmission of low-priority streams through network topology modeling and integer linear scheduling model, the problem of low bandwidth utilization and dynamic path selection caused by bandwidth reservation strategies of high-priority data streams in deterministic networks is solved, and efficient data streams are achieved Priority control improves the real-time response and resource utilization of the network.
Patent Information
- Application Number
- CN202410913003.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-09
- Publication Date
- 2024-12-03
AI Technical Summary
The bandwidth reservation strategy of high-priority data streams in existing deterministic networks leads to low bandwidth utilization, and it is difficult to adapt to dynamic path selection and data stream scheduling requirements in complex network environments, especially in real-time data transmission scenarios.
Use network topology modeling and time-sensitive flow edge occupation matrix definition, calculate the hop number of time-sensitive flow paths and the maximum number of flows of the paths, establish an integer linear scheduling model, and add multiple constraints to optimize the transmission start time of low-priority flows. Implement dynamic priority scheduling.
It improves the overall operation efficiency and security of the network, ensures priority access and rapid response of key service flows, and realizes efficient bandwidth usage and elastic priority allocation.
Smart Images

Figure CN119071249A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of computer network communication technology, and in particular to a deterministic network data stream priority control method. Background Technology
[0002] Deterministic networks are those that provide predictable network communication services, ensuring that performance metrics such as latency, bandwidth, and packet loss rate meet stringent prerequisites during data transmission. These networks are typically used in scenarios with extremely high real-time and reliability requirements, such as industrial automation, connected vehicles, and real-time audio / video transmission. In deterministic network data flow scheduling, it is necessary to reserve sufficient bandwidth for high-priority data flows.
[0003] A search revealed Chinese Patent Publication No. CN115484205B, entitled "A Deterministic Network Routing and Queue Scheduling Method and Apparatus," belonging to the field of computer technology. The method includes: creating an agent Ar to calculate the forwarding path and an agent Ac to calculate the forwarding period at each node along the path; agents Ar and Ac share a reward, using the global network state as input to evaluate the network and the state value as output, continuously updating the network with the goal of maximizing the expected reward, and selecting the optimal route and the optimal forwarding queue to specify the forwarding path of the deterministic flow and the period offset information at each node along the path.
[0004] This approach employs a fixed bandwidth reservation strategy to reserve necessary bandwidth for high-priority data streams. However, this method has significant limitations. First, high-priority data streams may not continuously occupy all reserved bandwidth. For example, during off-peak data transmission periods, this bandwidth remains idle and is not effectively utilized. In this case, even if low-priority data streams have transmission needs, they cannot use these idle bandwidth resources, resulting in low overall bandwidth utilization.
[0005] Secondly, as networks expand in scale and complexity, with more network nodes and intricate connections, traditional network topology optimization and maintenance strategies face significant challenges. Especially in scenarios requiring real-time data transmission, existing technologies often struggle to meet the demands of dynamic path selection and data flow scheduling. In complex network environments, effective path selection and traffic management become paramount, and the aforementioned methods exhibit clear deficiencies, failing to adapt to rapidly changing network conditions and data flow requirements. Summary of the Invention
[0006] The purpose of this invention is to address the shortcomings of existing technologies by proposing a deterministic network data flow priority control method. Its advantages lie in its applicability to deterministic networks, ensuring priority passage and rapid response for critical service flows, and significantly improving the overall operational efficiency and security of the network.
[0007] To achieve the above objectives, the present invention adopts the following technical solution:
[0008] A deterministic network data stream priority control method includes the following steps:
[0009] Step 1: Network topology modeling and definition of time-sensitive flow edge occupancy matrix;
[0010] Step 1.1: Define the set of switch nodes S, the set of host nodes H, and the set of nodes V in the network topology, and V = (S∪H); model the network topology as a directed graph G = (V,E);
[0011] Step 1.2: Define a time-sensitive flow f, and f = (s f ,d f ,p f ,t f ,dl f ), where s f It is the sending location of f, d f It is the destination of f, p f It is the sending period of f, t f It is the transmission delay of f, dl f It is the deadline for f;
[0012] Step 1.3: Define the edge occ[f][e]∈{0,1} of the time-sensitive flow f on edge e. |F|×|E| If time-sensitive flow f occupies edge e, then let occ[f][e] = 1; otherwise, let occ[f][e] = 0, where F is the set of time-sensitive flows and e ∈ E; define tr[f][e] as the start time of transmission of time-sensitive flow f on edge e;
[0013] Step 2: Calculate the hop count and maximum flow count of the time-sensitive flow path, and their fusion metrics;
[0014] Step 3: Calculate the low-priority flow f low The fusion index of the acyclic path set is used to sort the acyclic paths, resulting in the sorted acyclic path set SP. low ;
[0015] Step 4: Establish an integer linear scheduling model and add multiple constraints;
[0016] Step 4.1: Define and initialize the relevant variables;
[0017] Step 4.2: Add transmission start time constraints to the integer linear scheduling model;
[0018] The set of transmission periods χ of f is defined using the following formula. f :
[0019]
[0020] In the formula, f∈F conf p F [f] represents the transmission period matrix of f, and m is the value of p. F The greatest common divisor of [f]; x is the independent variable, and ¥ is a natural number;
[0021] For any f∈F conf , u[f][e]=1, there exists x∈χ f This makes the following constraint on the start time tr[f][e] of f on edge e hold;
[0022] x×m≤tr[f is [e]≤(x+λ)×mt F [f]
[0023] In the formula, λ is the time slot occupancy parameter, λ∈(0,1], t F [f] is the transmission duration of f;
[0024] Use the following formula to constrain tr[f][e];
[0025] tr[f][e] = tr[f][e] mod p F [f], if tr[f][e]≥p F [f]and occ[f][e]=1
[0026] In the formula, mod represents the modulo operation;
[0027] Step 4.3: Add conflict-free constraints to the integer linear scheduling model;
[0028] Define the first high-priority flow f1∈F conf The second high-priority flow f2∈F conf ;
[0029] Define p F [f1] represents the transmission period matrix of f1, p F [f2] represents the transmission period matrix of f2, and h is the value of p. F [f1] and p F The least common multiple of [f2];
[0030] The set of transmission periods χ of f1 is defined using the following formula. f1 :
[0031]
[0032] The set of transmission periods γ of f2 is defined using the following formula. f2 :
[0033]
[0034] In the formula, y represents another independent variable;
[0035] For any f1∈F conf and f2∈F conf , Define conflict-free constraints using the following formula:
[0036] tr[f1][e]+x·p F [f1]≥tr[f2][e]+y·p F [f2]+t F [f2], if occ[f1][e]+occ[f2][e]≥2
[0037] In the formula, t F [f2] is the transmission duration of f2;
[0038] Step 4.4: Add adjacent edge constraints to the integer linear scheduling model;
[0039] Define any q-th edge and any nth edge And e q ≠e n If edge e q and e n If they are adjacent edges, then let A EE [e q ][e n ] = 1;
[0040] For any f∈F conf A EE [e q ][e n ] = 1, construct adjacent edge constraints using the following formula:
[0041] tr[f][e n ]=t[f][e q ]+h F [f], if occ[f][e q ]+u[f][e n ≥2
[0042] In the formula, h F[f] is the single-hop delay time of f;
[0043] Step 4.5: Add flow deadline constraints to the integer linear scheduling model;
[0044] Step 4.6: Minimize the low-priority flow f low exist Taking the transmission start time on the first edge e1 as the objective, the objective function SST of the dynamic priority scheduling model is constructed using the following formula:
[0045]
[0046] In the formula, tr[f low [e1] represents f low The start time of transmission on the first edge e1;
[0047] Step 5: Solve the model using a solver to obtain f. low The start time of transmission, to ensure that in F conf In the reserved time slots, low-priority stream f low Also available in e z Transmission is performed on the side.
[0048] The present invention is further configured such that, in step 2, the hop count HOP(p) of an acyclic path p is defined using the following formula:
[0049] HOP(p) = len(p) - 1
[0050] In the formula, len(p) represents the total number of nodes in p.
[0051] The present invention is further configured such that, in step 2, the number of flows occupied on edge e, OCC[e], is defined using the following formula:
[0052]
[0053] In the formula, |F| represents the number of data streams in F.
[0054] The present invention is further configured such that, in step 2, the maximum number of flows (OCC) of all edges of the acyclic path p is defined using the following formula. max (p):
[0055]
[0056] In the formula, p i Let i and p represent nodes i on the acyclic path p. i+1 Represents node i+1 on the acyclic path p; (p i ,p i+1 ) represents the edge between node i and node i+1 on the acyclic path p.
[0057] The present invention is further configured such that, in step 2, HOP(p) and OCC are defined using the following formula. max The fusion index HOC(p) of (p):
[0058]
[0059] In the formula, |E| represents the number of edges in E.
[0060] The present invention is further configured such that step 3 includes the following sub-steps:
[0061] Step 3.1: Use a depth-first search algorithm to find a low-priority flow f in G. low Sending location Arrive at the destination The set of all acyclic paths AP low ;
[0062] Step 3.2: Calculate AP low The fusion parameter index for each acyclic path in the dataset, and the AP based on the fusion parameter index. low Sort all acyclic paths in ascending order and output the sorted set of acyclic paths SP. low Define the configuration stream set F conf Initialize F conf =F high .
[0063] The present invention is further configured such that the specific operation steps of step 4.1 are: defining variables and initializing k = 1; defining SP low The k-th acyclic path in the equation is Will Transform into the edge set of the k-th acyclic path Traversal If f low If edge e is occupied, then let occ[f] low [e] = 1, otherwise, occ[f] low [e] = 0, where,
[0064] The present invention is further configured such that the specific operation steps of step 4.5 are as follows: adding a flow cutoff time constraint using the following formula:
[0065] h F [f]×HOP(p f )≤dl F [f]
[0066] In the formula, f∈F conf p f It is an acyclic path of f, HOP(p)f ) represents an acyclic path p f hop count, dl F [f] is the deadline for f;
[0067] To minimize the low-priority flow f low exist Taking the transmission start time on the first edge e1 as the objective, the objective function SST of the dynamic priority scheduling model is constructed using the following formula:
[0068]
[0069] In the formula, tr[f low [e1] represents f low The start time of transmission on the first edge e1.
[0070] The present invention is further configured such that step 5 includes the following sub-steps:
[0071] Step 5.1: Solve the model using the solver. If SST has no integer solutions, then k = k + 1; if k < len(SP) low If SST is an integer, proceed to step 4.1; if SST is an integer, proceed to step 5.2.
[0072] Step 5.2, for the edge set The z-th edge e in z Perform a traversal. And z∈¥; f low Merged into F conf In the set;
[0073] f is obtained using the following formula low Start transmission time tr[f low ][e z ], to ensure in F conf In the reserved time slots, low-priority stream f low Also available in e z Transmission is performed at the edge to achieve priority transmission control;
[0074]
[0075] In the formula, h F [f low ] represents f low Single-hop delay time, f low The transmission period matrix; HC(e z ) represents the z-th edge e z The current hop count;
[0076] Step 5.3: Based on tr[f low ][ez In the edge set Transmission f low Complete f low The deployment.
[0077] The present invention further specifies that the fusion index HOC(p) is defined by the formula: HOC(p) = αgHOP(p) + βgOCC max (p), where α is the weighting factor of HOP(p), and β is the weighting factor of OCC. max (p) weighting factor.
[0078] The beneficial effects of this invention are as follows:
[0079] 1. Improved real-time response: This invention proposes an advanced data flow priority control method to ensure that critical data is processed first in network communication, which greatly improves the system's response speed and accuracy, and provides reliable timing guarantees for network critical services.
[0080] 2. Efficient bandwidth utilization: This invention establishes a dynamic priority scheduling mechanism, which optimizes the transmission time and path selection of data streams with different priorities, thereby achieving optimal allocation of network resources.
[0081] 3. Flexible priority allocation: This invention provides multi-priority traffic management, which enables latency-sensitive and non-latency-sensitive traffic to be effectively allocated according to priority, ensuring priority transmission of urgent data while meeting the transmission requirements of other traffic. Attached Figure Description
[0082] Figure 1 This is a schematic diagram of the algorithm flow for a deterministic network data stream priority control method proposed in this invention;
[0083] Figure 2 This is a schematic diagram of priority scheduling control for a deterministic network data stream priority control method proposed in this invention;
[0084] Figure 3 This is a schematic diagram of time slot resource allocation for a deterministic network data stream priority control method proposed in this invention. Detailed Implementation
[0085] The technical solution of this patent will be further described in detail below with reference to specific embodiments.
[0086] The embodiments of this patent are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this patent, and should not be construed as limiting this patent.
[0087] Example 1
[0088] Reference Figure 1 A deterministic network data stream priority control method includes the following steps:
[0089] Step 1: Network topology modeling and definition of time-sensitive flow edge occupancy matrix;
[0090] Step 1.1: Define the set of switch nodes S, the set of host nodes H, and the set of nodes V in the network topology, and V = (S∪H); model the network topology as a directed graph G = (V,E);
[0091] Step 1.2: Define a time-sensitive flow f, and f = (s f ,d f ,p f ,t f ,dl f ), where s f It is the sending location of f, d f It is the destination of f, p f It is the sending period of f, t f It is the transmission delay of f, dl f It is the deadline for f;
[0092] Step 1.3: Define the edge occ[f][e]∈{0,1} of the time-sensitive flow f on edge e. |F|×|E| If time-sensitive flow f occupies edge e, then let occ[f][e] = 1; otherwise, let occ[f][e] = 0, where F is the set of time-sensitive flows and e ∈ E; define tr[f][e] as the start time of transmission of time-sensitive flow f on edge e;
[0093] Step 2: Calculate the hop count and maximum flow count of the time-sensitive flow path, and their fusion metrics;
[0094] Step 2.1: Define the hop count HOP(p) of an acyclic path p using equation (1):
[0095] HOP(p)=len(p)⁻¹ (1)
[0096] In equation (1), len(p) represents the total number of nodes in p;
[0097] Step 2.2: Define the number of flows occupied on edge e, OCC[e], using equation (2):
[0098]
[0099] In equation (2), |F| represents the number of data streams in F;
[0100] Using equation (3), we define the maximum number of flows (OCC) for all edges of an acyclic path p. max (p):
[0101]
[0102] In equation (3), p i Let i and p represent nodes i on the acyclic path p. i+1 Represents node i+1 on the acyclic path p; (p i ,p i+1 () represents the edge between node i and node i+1 on the acyclic path p;
[0103] Step 2.3: Define HOP(p) and OCC using equation (4). max The fusion index HOC(p) of (p):
[0104]
[0105] In equation (4), |E| represents the number of edges in E;
[0106] HOC(p) is used to balance the number of hops (HOP(p)) of the acyclic path p with the maximum number of time-sensitive flows (OC) across all edges. max (p), and combine them into a value between 0 and 1; the path with the smaller fusion index is more suitable as the selected routing path, here for HOP(p) and OC max (p) Both parameters are normalized to ensure that the path selection is not biased towards either one.
[0107] Step 3: Calculate the low-priority flow f low The fusion index of the acyclic path set is used to sort the acyclic paths, resulting in the sorted acyclic path set SP. low ;
[0108] This invention determines the priority of a time-sensitive stream by identifying the Priority Code Point (PCP) in the stream, and then classifies the stream into high-priority and low-priority categories based on a set priority level threshold. For example... Figure 2 As shown, high-priority flows typically include control flows and critical data flows. These flows are small in size but extremely important and require priority processing. Meanwhile, considering efficiency, some low-priority time-sensitive flows can be scheduled within the same time period as high-priority flows, achieving resource sharing. The system's multi-priority flow management function ensures the priority transmission of critical data while also meeting the transmission needs of other data flows.
[0109] Step 3.1: Use the depth-first search algorithm to find the low-priority flow f in G. low Sending location Arrive at the destination The set of all acyclic paths AP low ;
[0110] Step 3.2, Calculate AP low The fusion parameter index for each acyclic path in the dataset, and the AP based on the fusion parameter index. low Sort all acyclic paths in ascending order and output the sorted set of acyclic paths SP. low Define the configuration stream set F conf Initialize F conf =F high ;
[0111] Step 4: Establish an integer linear scheduling model and add multiple constraints;
[0112] Step 4.1: Define and initialize the relevant variables;
[0113] k = 1; Define SP low The k-th acyclic path in the equation is Will Transform into the edge set of the k-th acyclic path
[0114] Traversal If f low If edge e is occupied, then let occ[f] low [e] = 1, otherwise, occ[f] low [e] = 0, where,
[0115] Step 4.2: Add transmission start time constraints to the integer linear scheduling model;
[0116] The set of transmission periods χ of f is defined using equation (5). f :
[0117]
[0118] In equation (5), f∈F conf p F [f] represents the transmission period matrix of f, and m is the value of p. F The greatest common divisor of [f]; x is the independent variable, and ¥ is a natural number;
[0119] For any f∈F conf , u[f][e]=1, there exists x∈χ f This makes the constraint of equation (6) on the start time tr[f][e] of f on edge e hold;
[0120] x×m≤tr[f is[e]≤(x+λ)×mt F [f] (6)
[0121] Equation (6), where λ is the time slot occupancy parameter, λ∈(0,1], t F [f] is the transmission duration of f;
[0122] Use equation (7) to constrain tr[f][e];
[0123] tr[f][e] = tr[f][e] mod p F [f], if tr[f][e]≥p F [f]and occ[f][e]=1 (7)
[0125] In equation (7), mod represents the modulo operation;
[0126] If the transmission periods of f1, f2, and f3 are 1, 2, and 4 milliseconds, respectively, then m equals 1 millisecond. For f3 with a period of 4 milliseconds, 0 ≤ x < 4. If λ = 0.5, and assuming the transmission duration of f3 is 10 μs, then the transmission start time of f3 is limited to the following ranges: [0, 0.49] milliseconds, [1, 1.49] milliseconds, [2, 2.49] milliseconds, and [3, 3.49] milliseconds.
[0127] Step 4.3: Add conflict-free constraints to the integer linear scheduling model;
[0128] Define the first high-priority flow f1∈F conf The second high-priority flow f2∈F conf ;
[0129] Define p F [f1] represents the transmission period matrix of f1, p F [f2] represents the transmission period matrix of f2, and h is the value of p. F [f1] and p F The least common multiple of [f2];
[0130] The set of transmission periods χ of f1 is defined using equation (8). f1 :
[0131]
[0132] The set of transmission periods γ of f2 is defined using equation (9). f2 :
[0133]
[0134] In equation (9), y represents another independent variable;
[0135] For any f1∈F conf and f2∈F conf , Define conflict-free constraints using equation (10):
[0136] tr[f1][e]+x·p F [f1]≥tr[f2][e]+y·p F [f2]+t F [f2], if occ[f1][e]+occ[f2][e]≥2 (10)
[0138] In equation (10), t F [f2] is the transmission duration of f2;
[0139] Step 4.4: Add adjacent edge constraints to the integer linear scheduling model;
[0140] Define any q-th edge and any nth edge And e q ≠e n If edge e q and e n If they are adjacent edges, then let A EE [e q ][e n ] = 1;
[0141] For any f∈F conf A EE [e q ][e n =1, construct adjacent edge constraints using equation (11):
[0142] tr[f][e n ]=t[f][e q ]+h F [f], if occ[f][e q ]+u[f][e n ]≥2 (11)
[0143] In equation (11), h F [f] is the single-hop delay time of f;
[0144] Step 4.5: Add flow deadline constraints to the integer linear scheduling model;
[0145] Add a flow cutoff time constraint using equation (12):
[0146] h F [f]×HOP(p f )≤dlF [f] (12)
[0147] In equation (12), f∈F conf p f It is an acyclic path of f, HOP(p) f ) represents an acyclic path p f hop count, dl F [f] is the deadline for f;
[0148] To minimize the low-priority flow f low exist Taking the transmission start time on the first edge e1 as the objective, the objective function SST of the dynamic priority scheduling model is constructed using equation (13):
[0149]
[0150] In equation (13), tr[f low [e1] represents f low The start time of transmission on the first edge e1;
[0151] like Figure 3 As shown, when F conf When the time-sensitive flow satisfies the constraints, and when the SST solver has integer solutions, it means f low You can be in F conf The system is deployed in the reserved time slots; this invention designs an efficient data flow priority control algorithm to ensure the priority transmission of critical data flows in network communication, which significantly improves the speed and accuracy of system response.
[0152] Step 5: Solve the model using a solver to obtain f. low The start time of transmission, to ensure that in F conf In the reserved time slots, low-priority stream f low Also available in e z Transmission is performed at the edge;
[0153] Step 5.1: Solve the model using the solver. If SST is an integer, proceed to step 5.2; if SST has no integer solutions, then k = k + 1; if k < len(SP) low Proceed to step 4.1;
[0154] Step 5.2, for the edge set The z-th edge e in z Perform a traversal. And z∈¥; f low Merged into F conf In the set;
[0155] f is obtained using equation (14)low Start transmission time tr[f low ][e z ], to ensure in F conf In the reserved time slots, low-priority stream f low Also available in e z Transmission is performed at the edge to achieve priority transmission control;
[0156]
[0157] In equation (14), h F [f low ] represents f low Single-hop delay time, f low The transmission period matrix; HC(e z ) represents the z-th edge e z The current hop count;
[0158] Step 5.3, based on tr[f low ][e z In the edge set Transmission f low Complete f low The deployment.
[0159] To verify the advantages of the above embodiments over the prior art, the following experiments were conducted:
[0160] Configure a network consisting of three nodes and a central switch. All nodes are connected to the switch via a 100Mbps link. Node 1 acts as an actuator, responsible for generating high-priority streams to send motor control messages to node 2. Node 2 acts as a sensor, generating low-priority streams to send device status information to node 3. Continuously monitor system performance using network monitoring tools to verify the effectiveness of the network configuration and ensure the accuracy and stability of the simulation.
[0161] The experimental parameters and results are shown below:
[0162] Table 1 Experimental parameter settings
[0163]
[0164] Table 2 Comparison of Experimental Performance Results
[0165]
[0166] As can be seen from the test results in Table 2 above, the technical solution disclosed in Example 1 is effective regardless of the f low Average latency, f low Shaking, f high Average latency, fhigh Both jitter and bandwidth utilization are significantly improved compared to existing technologies.
[0167] Example 2
[0168] Reference Figure 1 A deterministic network data stream priority control method, which differs from Embodiment 1 in that the fusion index HOC(p) is defined as follows: HOC(p) = αgHOP(p) + βgOCC max (p), where α is the weighting factor of HOP(p), and β is the weighting factor of OCC. max The weighting factor of (p); the definition formula of this fusion index HOC(p) simplifies the calculation process, facilitates the quick acquisition of calculation results, and improves the efficiency of calculation.
[0169] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.
Claims
1. A deterministic network data flow priority control method, It is characterized in that The following steps are involved: Step 1: Network topology modeling and time-sensitive flow edge occupancy matrix definition; Step 1.1: Define the switch node set S, host node set H, and node set V in the network topology, and V = (S∪H); model the network topology as a directed graph G = (V, E); Step 1.2: Define a time-sensitive flow f, where f = (s f ,d f ,p f ,t f ,dl f ), where s f is the place where f is sent, d f is the destination of f, p f is the sending period of f, t f is the transmission delay of f, dl f is the deadline of f; Step 1.3: Define the edge occupancy matrix occ[f][e]∈{0,1} of the time-sensitive flow f on edge e |F|×|E| , if the time-sensitive flow f occupies edge e, let occ[f][e] = 1, otherwise, let occ[f][e] = 0, where F is the set of time-sensitive flows, e∈E; define tr[f][e] to represent the transmission start time of the time-sensitive flow f on edge e; Step 2: Calculate the number of hops of the time-sensitive flow path and the maximum number of flows on the path, as well as their fusion index; Step 3: Calculate the low priority flow f low The fusion index of the acyclic path set is sorted to obtain the sorted acyclic path set SP low ; Step 4: Establish an integer linear scheduling model and add multiple constraints; Step 4.1: Define relevant variables and initialize them; Step 4.2: Add transmission start time constraints to the integer linear scheduling model; The transmission period set x of f is defined as follows: f : Where f∈F conf , p F [f] represents the transmission period matrix of f, m is p F The greatest common divisor of [f]; x is the independent variable, ¥ is a natural number; For any f∈F conf , u[f][e]=1, there exists x∈χ f , so that the following constraint on the transmission start time tr[f][e] of f on edge e holds true; x×m≤tr[f is ][e]≤(x+λ)×m-t F [f] Where λ is the time slot occupancy parameter, λ∈(0,1], t F [f] is the transmission duration of f; Use the following formula to constrain tr[f][e]; tr[f][e]=tr[f][e]mod p F [f],if tr[f][e]≥p F [f]and occ[f][e]=1 In the formula, mod represents the modular operation; Step 4.3: Add conflict-free constraints to the integer linear scheduling model; Define the first high priority flow f 1 ∈F conf and the second high priority flow f 2 ∈F conf ; Define p F [f 1 ] indicates f 1 The transmission period matrix, p F [f 2 ] indicates f 2 The transmission period matrix, h is p F [f 1 ] and p F [f 2 ]; Define f using the following formula 1 The transmission period set χ f1 : Define f using the following formula 2 The transmission period set γ f2 : In the formula, y represents another independent variable; For any f 1 ∈F conf and f 2 ∈F conf , The conflict-free constraint is defined as follows: tr[f 1 ][e]+x·p F [f 1 ]≥tr[f 2 ][e]+y·p F [f 2 ]+t F [f 2 ],if occ[f 1 ][e]+occ[f 2 ][e]≥2 Where, t F [f 2 ] is f 2 Duration of transmission; Step 4.4: Add adjacent edge constraints to the integer linear scheduling model; Define any qth edge and any nth edge And e q ≠e n ; If edge e q and e n are adjacent edges, then let A EE [e q ][e n ]=1; For any f∈F conf , A EE [e q ][e n ]=1, and use the following formula to construct the adjacent edge constraint: tr[f][e n ]=t[f][e q ]+h F [f],if occ[f][e q ]+u[f][e n ]≥2 In the formula, h F [f] is the single-hop delay time of f; Step 4.5: Add flow deadline constraints to the integer linear scheduling model; Step 4.6: Minimize the low priority flow f low exist The first edge e 1 The transmission start time is taken as the target, and the objective function SST of the dynamic priority scheduling model is constructed using the following formula: In the formula, tr[f low ][e 1 ] indicates f low On the first edge e 1 The transmission start time on ; Step 5: Use the solver to solve the model and obtain f low The start transmission time is to ensure that the F conf In the reserved time slot, the low priority flow f low Also available in e z Transmit on the edge.
2. A deterministic network data flow priority control method according to claim 1, It is characterized in that In step 2, the number of hops HOP(p) of a loop-free path p is defined using the following formula: HOP(p)=len(p)-1 Where len(p) represents the total number of nodes in p.
3. A deterministic network data flow priority control method according to claim 2, It is characterized in that In step 2, the number of occupied flows on edge e, OCC[e], is defined using the following formula: Where |F| represents the number of data flows in F.
4. A deterministic network data flow priority control method according to claim 3, It is characterized in that In step 2, the maximum number of flows of all edges of the acyclic path p is defined by the following formula: max (p): In the formula, p i represents nodes i, p on the acyclic path p i+1 represents the node i+1 on the acyclic path p; (p i ,p i+1 ) represents the edge between node i and node i+1 on the acyclic path p.
5. A deterministic network data flow priority control method according to claim 4, It is characterized in that In step 2, HOP(p) and OCC are defined using the following formula: max (p) fusion index HOC(p): Where |E| represents the number of edges in E.
6. A deterministic network data flow priority control method according to claim 1, It is characterized in that The step 3 includes the following sub-steps: Step 3.1: Use the depth-first search algorithm to find the low-priority flow f in G low Place of delivery To destination The set of all loop-free paths AP low ; Step 3.2: Calculate AP low The fusion parameter index of each loop-free path in the AP is calculated and the AP is assigned a fusion parameter index. low Sort all acyclic paths in ascending order and output the sorted acyclic path set SP low ; Define the configuration flow set F conf , initialize F conf =F high .
7. A deterministic network data flow priority control method according to claim 1, It is characterized in that The specific operation steps of step 4.1 are: define variables and initialize k=1; define SP low The kth acyclic path in is Will The edge set converted to the k-th acyclic path Traversal If f low Occupy edge e, then let occ[f low ][e]=1, otherwise, occ[f low ][e]=0, where 8. A deterministic network data flow priority control method according to claim 1, It is characterized in that The specific operation steps of step 4.5 are: add the flow cutoff time constraint using the following formula: h F [f]×HOP(p f )≤dl F [f] Where f∈F conf , p f is an acyclic path of f, HOP(p f ) represents an acyclic path p f The number of hops, dl F [f] is the cutoff time of f; To minimize the low priority flow f low exist The first edge e 1 The transmission start time is taken as the target, and the objective function SST of the dynamic priority scheduling model is constructed using the following formula: In the formula, tr[f low ][e 1 ] indicates f low On the first edge e 1 The time at which the transfer started.
9. A deterministic network data flow priority control method according to claim 1, It is characterized in that The step 5 comprises the following sub-steps: Step 5.1: Use the solver to solve the model. If SST has no integer solution, then k = k + 1; if k < len (SP low ), go to step 4.1; if SST is an integer, go to step 5.2; Step 5.2: For edge sets The zth edge e in z To traverse, And z∈¥; f low Merge to F conf In the collection; Use the following formula to get f low The start transmission time tr[f low ][e z ] to ensure that the F conf In the reserved time slot, the low priority flow f low Also available in e z Transmit on the edge to achieve priority transmission control; tr[f low ][e z ]=(SST+h F [f low ]*(HC(e z )-1))mod p flow In the formula, h F [f low ] indicates f low The single-hop delay time, p flow represents f low The transmission cycle matrix; HC(e z ) represents the zth edge e z The current hop count of Step 5.3: According to tr[f low ][e z ]In the edge set Medium transmission low , complete f low deployment.
10. A deterministic network data flow priority control method according to claim 5, It is characterized in that The definition formula of the fusion index HOC(p) is: HOC(p)=αgHOP(p)+βgOCC max (p), where α is the weight factor of HOP(p), and β is the OCC max The weighting factor of (p).
Citation Information
Patent Citations
Deterministic Network Routing and Queue Scheduling Method and Apparatus
CN115484205B
