CN110933157A - Industrial Internet of things-oriented edge computing task unloading method - Google Patents
Industrial Internet of things-oriented edge computing task unloading method Download PDFInfo
- Publication number
- CN110933157A CN110933157A CN201911176742.9A CN201911176742A CN110933157A CN 110933157 A CN110933157 A CN 110933157A CN 201911176742 A CN201911176742 A CN 201911176742A CN 110933157 A CN110933157 A CN 110933157A
- Authority
- CN
- China
- Prior art keywords
- task
- edge computing
- node
- nodes
- edge
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 35
- 230000005540 biological transmission Effects 0.000 claims abstract description 62
- 238000004519 manufacturing process Methods 0.000 claims description 23
- 230000006870 function Effects 0.000 claims description 16
- 101001086191 Borrelia burgdorferi Outer surface protein A Proteins 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 12
- 238000004891 communication Methods 0.000 claims description 4
- 230000003213 activating effect Effects 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 239000000126 substance Substances 0.000 claims description 3
- 239000010410 layer Substances 0.000 description 27
- 235000008694 Humulus lupulus Nutrition 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 4
- 230000001934 delay Effects 0.000 description 3
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000001737 promoting effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000002356 single layer Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L45/00—Routing or path finding of packets in data switching networks
- H04L45/12—Shortest path evaluation
- H04L45/124—Shortest path evaluation using a combination of metrics
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/10—Protocols in which an application is distributed across nodes in the network
- H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
- H04L67/1004—Server selection for load balancing
- H04L67/1008—Server selection for load balancing based on parameters of servers, e.g. available memory or workload
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/02—Traffic management, e.g. flow control or congestion control
- H04W28/08—Load balancing or load distribution
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Hardware Design (AREA)
- General Engineering & Computer Science (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention relates to an industrial Internet of things-oriented edge computing task unloading method. The method comprises the steps of constructing a three-layer edge computing node network model; planning an optimal point-to-point path in the network by an SDN (software defined network) controller; defining a time delay model from the equipment to unload the tasks to the edge computing nodes of each layer, and selecting N better edge computing nodes as candidate nodes for task unloading through a primary selection rule from the edge computing nodes with the total time delay of the computing equipment tasks smaller than the maximum tolerant time delay of the tasks; and then taking the length and priority of the device task and the ratio of the transmission delay of task unloading and the maximum tolerance delay of the task as fuzzy logic input variables, and selecting the optimal node in the candidate nodes by adopting a fuzzy logic algorithm to unload the task. By the task unloading method, the difference requirements of the equipment on data real-time performance and computing resources are ensured, and the transmission and processing time of the equipment task is reduced.
Description
Technical Field
The invention belongs to the field of edge computing, and relates to an edge computing task unloading method for an industrial Internet of things.
Background
The edge computing is an open platform which is close to the network edge of an object or a data source and integrates core capabilities of a network, computing, storage and the like, the requirement of quick interaction response is met, and universal and flexible computing service is provided. Task unloading in edge computing is to unload the computing task of a device terminal to an edge computing environment at the edge of a network, so that the problems of the terminal device in the aspects of computing resources, storage space, energy efficiency and the like are solved.
In an industrial internet of things environment, a traditional cloud computing service adopts a centralized processing mode, has the advantages of low cost, high reliability, scalability and the like, and also has the problems that the data transmission distance of a device task is too long, and the real-time performance of data transmission is uncertain due to the variability of the network communication quality. With the continuous improvement of the real-time requirement of modern factory bottom equipment on data and the difference requirement on computing resources, cloud computing services are difficult to meet production environments with high real-time requirements on data, edge computing is achieved, defects of cloud computing are effectively overcome, a single-layer edge computing node unloading method is mostly adopted at the present stage, the problem of the real-time requirement on data is effectively solved, but the difference requirement of different bottom equipment tasks on the computing resources in a factory is ignored, and the computing resources cannot be fully utilized.
Disclosure of Invention
In view of this, the present invention provides an edge computing task offloading method for an industrial internet of things, which can meet the difference requirements of task requests of field production devices of different factories on computing resources. And selecting the optimal edge computing node for task unloading through an SDN technology and a fuzzy logic algorithm, thereby reducing the transmission time and the processing time of the computing task.
In order to achieve the purpose, the invention provides the following technical scheme:
an edge computing task unloading method facing an industrial Internet of things reasonably distributes computing resources to a factory production environment through layered edge computing nodes, an SDN controller completes point-to-point optimal transmission path planning in a network through an improved shortest path algorithm (OSPA), and unloading of equipment tasks to the optimal edge computing nodes is achieved through primary selection rules and fuzzy logic algorithms; the method comprises the following steps:
s1: constructing a network model of three layers of edge computing nodes, wherein the first layer of edge computing nodes are positioned in a factory workshop field, are close to industrial field production equipment and are responsible for accessing industrial wired and wireless equipment; the second layer of edge computing nodes are positioned at the general exit of data communication in the workshop and serve as a function of a workshop edge server; the third layer of edge computing nodes are positioned in a data center in a factory and are used for completing comprehensive computation and storage of data tasks among workshops; selecting the CPU frequency and the memory size performance parameters of the edge computing node according to the data volume scale of the field production equipment in the area where the edge computing node is located;
s2: the SDN controller plans an optimal point-to-point path in the industrial network through an improved shortest path algorithm OSPA;
s3: defining a time delay model for unloading tasks to each layer of edge computing nodes by equipment, and calculating the total time delay of the tasks of the equipmentLess than maximum tolerated delay for a taskSelecting N optimal edge computing NODEs as candidate NODEs for task unloading through an initial selection rule, wherein a candidate edge computing NODE set is represented as { NODE _1, NODE _ 2.,. NODE _ N };
s4: the length and priority of the device task and the ratio of the transmission delay of the task unloading to the maximum tolerance delay of the task are used as input variables of a fuzzy logic algorithm, membership functions of the three parameters are defined, a fuzzy rule and an inference method are selected, and the optimal node in candidate nodes is selected by the fuzzy logic algorithm to unload the task.
Optionally, the step S2 includes:
when planning an optimal point-to-point path in an industrial network, a network manager SDN controller adopts a shortest path algorithm to improve a traditional Dijkstra shortest path algorithm; since the SDN controller maintains the wholeIndustrial network topological graph and collection of link quality parameters, improved shortest path algorithm OSPA determines link weight W between two pointsijUsing the minimum average hop count k 'of the source to destination node path'hopMinimum average packet loss ratio loss'linkAnd minimum average transmission delay Tl′inkdelayA weighted average of the three parameters;
the performance requirements of tasks of different production devices in a factory on transmission paths are different, the SDN controller determines α delta coefficient values according to the requirements of task request types, and meanwhile, the SDN controller obtains the performance of current factory network link parameters and calculates k'hop、loss′link、Tl′inkdelayAnd according to formula Wij=αk′hop+βloss′link+δTl′inkdelayCalculate WijAnd the SDN controller plans an optimal point-to-point path in the industrial network by an improved shortest path algorithm OSPA.
Optionally, the step S3 includes:
defining a time delay model from the equipment to unload tasks to the edge computing nodes of each layer; computing nodes for first tier edgesIs responsible for the access of industrial wired and wireless devices on site, and has a certain difference in unloading time delay of the industrial wired and wireless devices due to the complex production environment of a factory, wherein,representing the mth edge computing node of the nth layer;
industrial wireless device deviOffloading tasks to first tier edge compute nodesHas an upload rate of riIn bps; reliability of wireless transmission determines stability of task offloading, industrial wireless device deviOffloading tasks to first tier edge compute nodesIs expressed as Pi;
When the grouping is collided, because the wireless transmission has a retransmission mechanism, each task request message can be ensured to be uploaded to the first layer edge computing node, and the dev is definediOffloading single-bit tasks toThe transmission time loss model of (1) is:
collecting data transmission rate B of each link of factory network through SDN controllerwIn bps, wherein Bw={B1,B2,B3…, industrial cable plant deviThe time delay for offloading the unit bit task to the edge compute node is:wherein q represents the industrial wireline equipment deviA number of links to a transmission path between edge computing nodes determined by the SDN controller;
the time delay of the transmission unit bit task between the computing nodes of each layer edge is expressed asWherein linkedgeRepresenting the number of links of a transmission path between two edge computing nodes determined by an SDN controller, so an industrial device deviWill task siIs unloaded toThe transmission delay is:
wherein x isi∈{0,1},For industrial wireless device xi0 for industrial wired device xi=1;ψiRepresenting a task siThe length of the message;
wherein λ isiψiRepresenting a task siCalculating the number of CPU cycles of the node by the needed edge; lambda [ alpha ]iCoefficient representation task siThe number of CPU cycles of the edge computing node required by the unit bit task depends on the task siThe computational complexity of (2);representing the CPU frequency of the mth edge computing node of the nth layer; in the method, the layering of the edge computing nodes is considered, and the edge computing nodes have the characteristics of multi-core and multi-task, so that the waiting time delay of the tasks in the edge computing nodes is ignored;
total latency at computing device tasksLess than maximum tolerated delay for a taskThe step of selecting candidate edge computing nodes through the primary selection rule in the edge computing nodes is as follows; firstly, in the first round, N edge calculation nodes with higher CPU frequency are selected from the edge calculation nodes with the CPU utilization rate smaller than Ut, wherein the Ut belongs to [0,1 ]]If the number of the CPU utilization rate of the edge computing nodes is less than N, the number of the edge computing nodes meeting the conditions in the round of selection is recorded as num1(ii) a The CPU utilization rate of the edge computing nodes in the second round is selected from N-num with higher CPU frequency in [ Ut + (R-2) M, Ut + (R-1) M ]1A candidateThe nodes, wherein M is the increment of the CPU utilization rate in each round of selection, R represents the R-th round, R is more than or equal to 2, and the like until N candidate nodes are selected; if the CPU utilization is equal to Ut until at the edge compute nodemaxAnd stopping selecting when the N candidate NODEs are not selected, taking the edge computing NODEs selected in the previous rounds as candidate NODEs, and recording the set of the candidate edge computing NODEs as { NODE _1, NODE _ 2.,. NODE _ N }.
Optionally, the step S4 includes:
taking the length and priority of the device task and the ratio of the transmission delay of task unloading and the maximum tolerance delay of the task as input variables of a fuzzy logic algorithm, wherein the fuzzy logic algorithm comprises the following steps:
(1) fuzzifying the length and priority of the device task and the ratio of the transmission delay of task unloading to the maximum tolerance delay of the task, and defining three variables mulen(u),μpri(u),μdelay(u) membership functions and corresponding fuzzy sets;
(2) selecting a corresponding fuzzy rule, obtaining the corresponding membership degree of each input variable through the membership function in the step (1), activating the fuzzy rule meeting the conditions, connecting the conditions of the fuzzy rule with logic and operation, and determining the fuzzy value meeting the conditions by adopting a minimum value method, namely:
wherein the content of the first and second substances,the number of fuzzy rules which are consistent with the fuzzy rule conclusion and are candidate edge computing NODEs NODE _ N is represented;
the final fuzzy value of the candidate edge calculation node is:
(3) defining fuzzy sets of edge compute nodes for optimal task offloadingAnd membership function A (u) thereof, using implication operator to obtain result { mu ] of precondition, i.e. length and priority of equipment task and ratio of transmission delay of task unloading to maximum tolerance delay of the taskNODE_1,μNODE_2,...,μNODE_NTruncating an edge calculation node fuzzy set unloaded by the optimal task, gathering the result of each rule after truncation, defuzzifying by using a centroid method to obtain a final result theta, wherein theta is more than 0 and less than or equal to 1, and when theta belongs to [0, y ∈ [ ]1]Selecting an edge computing NODE NODE _1 to unload the equipment task; when theta is equal to (y)1,y2]Selecting an edge computing NODE NODE _2 to unload the equipment task; when theta is equal to (y)N-1,yN]Then, selecting an edge computing NODE NODE _ N for unloading the device task, wherein y is more than 01<…<yN≤1。
The invention has the beneficial effects that: the invention provides a network model of three layers of edge computing nodes combined with the SDN technology, which distributes computing resources to a factory production environment through reasonable edge computing node layering, realizes the optimal transmission path planning of task unloading by adopting the SDN technology, and realizes the optimal unloading of factory equipment tasks by combining a fuzzy logic algorithm. By the method, the requirements of different bottom layer equipment in a factory on the difference of data instantaneity and computing resources are met, the cloud service computing pressure can be relieved, the data transmission quantity of a factory core network is reduced, network congestion of the core network is avoided, and the uncertainty of data transmission delay is reduced, so that economical, efficient and safe production of the factory is realized, and the production and operation costs are reduced.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objectives and other advantages of the invention may be realized and attained by the means of the instrumentalities and combinations particularly pointed out hereinafter.
Drawings
For the purposes of promoting a better understanding of the objects, aspects and advantages of the invention, reference will now be made to the following detailed description taken in conjunction with the accompanying drawings in which:
fig. 1 is a network topology structure diagram of an edge computing task offloading method for an industrial internet of things according to an embodiment of the present invention.
Fig. 2 is a flow chart of an edge computing task offloading method for an industrial internet of things according to an embodiment of the present invention.
Fig. 3 is a flowchart for calculating link weight W according to an embodiment of the present inventionijSchematic network topology diagram of (a).
Fig. 4 is a flowchart of an SDN controller planning an OSPA optimal path from point to point in an industrial network by using an improved shortest path algorithm according to an embodiment of the present invention.
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention in a schematic way, and the features in the following embodiments and examples may be combined with each other without conflict.
Wherein the showings are for the purpose of illustrating the invention only and not for the purpose of limiting the same, and in which there is shown by way of illustration only and not in the drawings in which there is no intention to limit the invention thereto; to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of an actual product; it will be understood by those skilled in the art that certain well-known structures in the drawings and descriptions thereof may be omitted.
The same or similar reference numerals in the drawings of the embodiments of the present invention correspond to the same or similar components; in the description of the present invention, it should be understood that if there is an orientation or positional relationship indicated by terms such as "upper", "lower", "left", "right", "front", "rear", etc., based on the orientation or positional relationship shown in the drawings, it is only for convenience of description and simplification of description, but it is not an indication or suggestion that the referred device or element must have a specific orientation, be constructed in a specific orientation, and be operated, and therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes, and are not to be construed as limiting the present invention, and the specific meaning of the terms may be understood by those skilled in the art according to specific situations.
An edge computing task unloading method facing an industrial Internet of things is designed to meet the difference requirements of task requests of different bottom layer production devices in a factory on computing resources and real-time performance, and a network structure diagram of the three-layer edge computing task unloading method combining an SDN technology is provided, wherein a network model of three-layer edge computing nodes is constructed as shown in figure 1, the first layer of edge computing nodes are located on the site of a factory workshop and close to industrial field production devices and are responsible for accessing industrial field wired and wireless production devices into a factory network to realize the conversion of a transmission protocol and the collection of data; the second layer edge computing node is positioned at a general exit of data communication in the workshop and has the function of serving as a workshop edge server to finish the comprehensive processing of data of each production line in the workshop and provide data service for an upper layer; the third layer of edge computing nodes are positioned in a data center in a factory and are used for completing comprehensive computation and storage of data tasks among workshops; computing resources are distributed to a factory production environment through three-layer edge computing division, an SDN controller realizes point-to-point optimal transmission path planning in a network through an improved shortest path algorithm (OSPA), and unloading of equipment tasks to optimal edge computing nodes is realized through a primary selection rule and a fuzzy logic algorithm, so that the requirements of modern factory equipment tasks on instantaneity and difference of the computing resources are met.
Fig. 2 is a flowchart of an unloading method for an edge computing task for an industrial internet of things, where the method includes:
s1: a network model of three-layer edge computing nodes is constructed, and a network topology diagram of the network model is shown in figure 1. And selecting performance parameters such as CPU frequency, memory size and the like of the edge computing node according to the data volume scale of the field production equipment in the area where the edge computing node is located.
S2: and the SDN controller plans an optimal point-to-point path in the industrial network by an improved shortest path algorithm OSPA.
The characteristic that a control plane and a forwarding plane of an SDN are separated and a counter of a flow table are used for acquiring performance parameters of a real-time industrial network link, so that the delay T of the link is completedi cLink packet loss ratioSum path hop number khopAnd (6) measuring.
And when planning the optimal point-to-point path in the industrial network, the SDN controller adopts a Dijkstra shortest path algorithm. In a traditional Dijkstra shortest path algorithm, link weights between two points are often obtained through a single variable or weighted average of performance parameters of the links, only performance indexes of a current link are considered when determining the weights of the links, and link quality of a subsequent path is not considered. In the invention, the improved shortest path algorithm OSPA determines the link weight W between two points because the SDN controller maintains the whole industrial network topological graph and collects the link quality parametersijUsing the minimum average hop count k 'of the source to destination node path'hopMinimum average packet loss ratio loss'linkAnd minimum average transmission delay T'linkdelayThe weighted average of the three parameters, as shown in fig. 3.
The source node of path L (① → ② → ④ → ⑥) is ① and the destination node is ⑥, where W is12K of'hop、loss′link、T′linkdelayThe calculation method is as follows:
wherein k ismaxRepresents the maximum number of hops for the path from node ① to ⑥;the sum of the packet loss rates, hop, of the segments of the path with the minimum packet loss rates from ① to ⑥lossIndicating the hop count of the path with the minimum packet loss rate from ① to ⑥;sum of transmission delays, hop, of segments of the link representing paths with minimum transmission delays from ① to ⑥TRepresenting the number of hops for the path with the minimum transmission delay from ① to ⑥, then W12The calculation method is as follows:
W12=αk′hop+βloss′link+δT′linkdelay
wherein the value of α + β + δ is 1.α, δ depending on the device task pair k'hop、loss′link、T′linkdelayThree parameters are required, and the SDN controller flexibly adjusts α, β and delta sizes of each task request23K of'hop、loss′link、T′linkdelayThe maximum number of hops of the path ② → ④ → ⑥, the sum of the packet loss ratios of the links of the path with the minimum packet loss ratio, the number of hops of the path with the minimum packet loss ratio, the sum of the transmission delays of the links of the path with the minimum transmission delay, the number of hops of the path with the minimum transmission delay, and other parameters are calculatedijThe calculation method is the same as above. The weight W calculated in this wayijNot only the current link quality parameter but also the subsequent destination of the path are consideredLink quality of the node. By the method, the influence on the selection of the subsequent link when the quality of a certain link in a certain path is in an extreme condition can be avoided.
The performance requirements of tasks of different production devices in a factory on transmission paths are different, the SDN controller determines α delta coefficient values according to the requirements of task request types, and meanwhile, the SDN controller obtains the performance of current factory network link parameters and calculates k'hop、loss′link、T′linkdelayAnd calculating W according to the formulaijThe SDN controller plans an optimal path from point to point in the industrial network by using an improved shortest path algorithm OSPA, and fig. 4 is a flowchart for planning an optimal path from point to point in the industrial network.
S3: defining a time delay model for unloading tasks to each layer of edge computing nodes by equipment, and calculating the total time delay of the tasks of the equipmentLess than maximum tolerated delay for a taskIn the edge computing NODEs, the optimal N edge computing NODEs are selected as candidate NODEs for task unloading through an initial selection rule, and a candidate edge computing NODE set is expressed as { NODE _1, NODE _ 2.
And defining a time delay model for unloading tasks to the edge computing nodes of each layer by the equipment. Computing nodes for first tier edgesIs responsible for the access of industrial wired and wireless devices on site, and has a certain difference in unloading time delay of the industrial wired and wireless devices due to the complex production environment of a factory, wherein,representing the mth edge compute node of the nth layer.
Industrial wireless device deviOffloading tasks to first tier edge compute nodesHas an upload rate of riReliability of wireless transmission in bps determines stability of task offloading, dev of industrial wireless deviceiOffloading tasks to first tier edge compute nodesIs expressed as Pi。
When the grouping is collided, because the wireless transmission has a retransmission mechanism, each task request message can be ensured to be uploaded to the first layer edge computing node, and the dev is definediOffloading single-bit tasks toThe transmission time loss model of (1) is:
collecting data transmission rate B of each link of factory network through SDN controllerwIn bps, wherein Bw={B1,B2,B3…, industrial cable plant deviThe time delay for offloading the unit bit task to the edge compute node is:wherein q represents the industrial wireline equipment deviA number of links to a transmission path between edge computing nodes determined by the SDN controller.
The time delay of the transmission unit bit task between the computing nodes of each layer edge is expressed asWherein linkedgeRepresenting the number of links of a transmission path between two edge computing nodes determined by an SDN controller, so an industrial device deviWill task siIs unloaded toThe transmission delay is:
wherein x isiE {0,1}, for industrial wireless device xi0 for industrial wired device xi=1;ψiRepresenting a task siThe length of the message.
wherein λ isiψiRepresenting a task siCalculating the number of CPU cycles of the node by the needed edge; lambda [ alpha ]iCoefficient representation task siThe number of CPU cycles of the edge computing node required by the unit bit task depends on the task siThe computational complexity of (2);representing the mth edge compute node CPU frequency of the nth layer. In the method, the edge computing node layering is considered, and the edge computing node has the characteristics of multi-core and multi-task, so that the waiting time delay of the task in the edge computing node is ignored.
Total latency at computing device tasksLess than maximum tolerated delay for a taskThe step of selecting candidate edge computing nodes according to the preliminary selection rule in the edge computing nodes of (1) is as follows. Firstly, selecting a CPU frequency from edge computing nodes with the CPU utilization rate less than Ut in the first roundLarge N, where Ut ∈ [0,1 ]]If the number of the CPU utilization rate of the edge computing nodes is less than N, the number of the edge computing nodes meeting the conditions in the round of selection is recorded as num1. The CPU utilization rate of the edge computing nodes in the second round is selected from N-num with higher CPU frequency in [ Ut + (R-2) M, Ut + (R-1) M ]1And (4) candidate nodes, wherein M is the increment of the CPU utilization rate in each round of selection, R represents the R-th round, R is more than or equal to 2, and the like until the N candidate nodes are selected. If the CPU utilization is equal to Ut until at the edge compute nodemaxAnd stopping selecting when the N candidate NODEs are not selected, taking the edge computing NODEs selected in the previous rounds as candidate NODEs, and recording the set of the candidate edge computing NODEs as { NODE _1, NODE _ 2.,. NODE _ N }.
S4: the length and priority of the device task and the ratio of the transmission delay of the task unloading to the maximum tolerance delay of the task are used as input variables of a fuzzy logic algorithm, membership functions of the three parameters are defined, a fuzzy rule and an inference method are selected, and the fuzzy logic algorithm is adopted to select the optimal node in candidate nodes for task unloading.
Taking the length and priority of the device task and the ratio of the transmission delay of task unloading and the maximum tolerance delay of the task as input variables of a fuzzy logic algorithm, wherein the fuzzy logic algorithm comprises the following steps:
(1) and (3) performing variable fuzzification on the length and priority of the equipment task and the ratio of the transmission delay of task unloading to the maximum tolerance delay of the task, and defining membership functions of three variables and corresponding fuzzy sets.
The task priority of the equipment is divided into four levels, from high to low, which are respectively:the fuzzy set is as follows:the length grades of the equipment tasks are divided into four grades, and the four grades are respectively from high to low:the fuzzy set is as follows:the ratio of the transmission delay of the equipment task unloading to the maximum tolerance delay of the task is divided into four levels from high to low:its corresponding fuzzy set is:and defining membership functions corresponding to the length of the device task, the priority of the device task and the ratio of the transmission delay of the device task unloading to the maximum tolerance delay of the task respectively.
(2) Selecting a corresponding fuzzy rule, obtaining the corresponding membership degree of each input variable through the membership function in the step (1), activating the fuzzy rule meeting the conditions, connecting the conditions of the fuzzy rule with logic and operation, and determining the fuzzy value meeting the conditions by adopting a minimum value method, namely:
wherein the content of the first and second substances,and the number of fuzzy rules which are consistent with the fuzzy rule conclusion and are used as candidate edge computing NODEs NODE _ N is represented.
The final fuzzy value of the candidate edge calculation node is:
(3) defining fuzzy sets of edge compute nodes for optimal task offloadingAnd its membership function A (u), using implication operator to unload preconditions, i.e. length, priority and task of equipment taskThe ratio of the transmission delay of the carrier to the maximum tolerated delay of the task, the result is obtained as muNODE_1,μNODE_2,...,μNODE_NTruncating an edge computing node fuzzy set unloaded by the optimal task, gathering the truncated result of each rule, defuzzifying by using a centroid method to obtain a final result theta
Wherein 0 < theta < 1, when theta belongs to [0, y ∈ [ ]1]Selecting an edge computing NODE NODE _1 to unload the equipment task; when theta is equal to (y)1,y2]Selecting an edge computing NODE NODE _2 to unload the equipment task; when theta is equal to (y)N-1,yN]Then, selecting an edge computing NODE NODE _ N for unloading the device task, wherein y is more than 01<…<yN≤1。
Finally, the above embodiments are only intended to illustrate the technical solutions of the present invention and not to limit the present invention, and although the present invention has been described in detail with reference to the preferred embodiments, it will be understood by those skilled in the art that modifications or equivalent substitutions may be made on the technical solutions of the present invention without departing from the spirit and scope of the technical solutions, and all of them should be covered by the claims of the present invention.
Claims (4)
1. An industrial Internet of things-oriented edge computing task unloading method is characterized by comprising the following steps: reasonably distributing computing resources to a factory production environment through layered edge computing nodes, finishing the optimal transmission path planning from point to point in a network by an SDN controller through an improved shortest path algorithm (OSPA), and unloading equipment tasks to the optimal edge computing nodes through a primary selection rule and a fuzzy logic algorithm; the method comprises the following steps:
s1: constructing a network model of three layers of edge computing nodes, wherein the first layer of edge computing nodes are positioned in a factory workshop field, are close to industrial field production equipment and are responsible for accessing industrial wired and wireless equipment; the second layer of edge computing nodes are positioned at the general exit of data communication in the workshop and serve as a function of a workshop edge server; the third layer of edge computing nodes are positioned in a data center in a factory and are used for completing comprehensive computation and storage of data tasks among workshops; selecting the CPU frequency and the memory size performance parameters of the edge computing node according to the data volume scale of the field production equipment in the area where the edge computing node is located;
s2: the SDN controller plans an optimal point-to-point path in the industrial network through an improved shortest path algorithm OSPA;
s3: defining a time delay model for unloading tasks to each layer of edge computing nodes by equipment, and calculating the total time delay of the tasks of the equipmentLess than maximum tolerated delay for a taskSelecting N optimal edge computing NODEs as candidate NODEs for task unloading through an initial selection rule, wherein a candidate edge computing NODE set is represented as { NODE _1, NODE _ 2.,. NODE _ N };
s4: the length and priority of the device task and the ratio of the transmission delay of the task unloading to the maximum tolerance delay of the task are used as input variables of a fuzzy logic algorithm, membership functions of the three parameters are defined, a fuzzy rule and an inference method are selected, and the optimal node in candidate nodes is selected by the fuzzy logic algorithm to unload the task.
2. The industrial internet of things-oriented edge computing task offloading method of claim 1, wherein: the step S2 includes:
when planning an optimal point-to-point path in an industrial network, a network manager SDN controller adopts a shortest path algorithm to improve a traditional Dijkstra shortest path algorithm; since the SDN controller maintains the entire industrial network topology and collects the link quality parameters, the improved shortest Path Algorithm OSPA determines the link weights W between two pointsijWhen it is in use, makeMinimum average hop number k 'from source node to destination node path'hopMinimum average packet loss ratio loss'linkAnd minimum average transmission delay T'linkdelayA weighted average of the three parameters;
the performance requirements of tasks of different production devices in a factory on transmission paths are different, the SDN controller determines α delta coefficient values according to the requirements of task request types, and meanwhile, the SDN controller obtains the performance of current factory network link parameters and calculates k'hop、loss′link、T′linkdelayAnd according to formula Wij=αk′hop+βloss′link+δT′linkdelayCalculate WijAnd the SDN controller plans an optimal point-to-point path in the industrial network by an improved shortest path algorithm OSPA.
3. The industrial internet of things-oriented edge computing task offloading method of claim 1, wherein: the step S3 includes:
defining a time delay model from the equipment to unload tasks to the edge computing nodes of each layer; computing nodes for first tier edgesIs responsible for the access of industrial wired and wireless devices on site, and has a certain difference in unloading time delay of the industrial wired and wireless devices due to the complex production environment of a factory, wherein,representing the mth edge computing node of the nth layer;
industrial wireless device deviOffloading tasks to first tier edge compute nodesHas an upload rate of riIn bps; reliability of wireless transmission determines stability of task offloading, industrial wireless device deviOffloading tasks to first tier edge compute nodesIs expressed as Pi;
When the grouping is collided, because the wireless transmission has a retransmission mechanism, each task request message can be ensured to be uploaded to the first layer edge computing node, and the dev is definediOffloading single-bit tasks toThe transmission time loss model of (1) is:
collecting data transmission rate B of each link of factory network through SDN controllerwIn bps, wherein Bw={B1,B2,B3…, industrial cable plant deviThe time delay for offloading the unit bit task to the edge compute node is:wherein q represents the industrial wireline equipment deviA number of links to a transmission path between edge computing nodes determined by the SDN controller;
the time delay of the transmission unit bit task between the computing nodes of each layer edge is expressed asWherein linkedgeRepresenting the number of links of a transmission path between two edge computing nodes determined by an SDN controller, so an industrial device deviWill task siIs unloaded toThe transmission delay is:
wherein x isiE {0,1}, for industrial wireless device xi0 for industrial wired device xi=1;ψiRepresenting a task siThe length of the message;
wherein λ isiψiRepresenting a task siCalculating the number of CPU cycles of the node by the needed edge; lambda [ alpha ]iCoefficient representation task siThe number of CPU cycles of the edge computing node required by the unit bit task depends on the task siThe computational complexity of (2);representing the CPU frequency of the mth edge computing node of the nth layer; in the method, the layering of the edge computing nodes is considered, and the edge computing nodes have the characteristics of multi-core and multi-task, so that the waiting time delay of the tasks in the edge computing nodes is ignored;
total latency at computing device tasksLess than maximum tolerated delay for a taskThe step of selecting candidate edge computing nodes through the primary selection rule in the edge computing nodes is as follows; firstly, in the first round, N edge calculation nodes with higher CPU frequency are selected from the edge calculation nodes with the CPU utilization rate smaller than Ut, wherein the Ut belongs to [0,1 ]]If the number of the CPU utilization rate of the edge computing nodes is less than N, the number of the edge computing nodes meeting the conditions in the round selection is countedThe number is num1(ii) a The CPU utilization rate of the edge computing nodes in the second round is selected from N-num with higher CPU frequency in [ Ut + (R-2) M, Ut + (R-1) M ]1The candidate nodes, wherein M is the increment of the CPU utilization rate in each round of selection, R represents the R-th round, R is more than or equal to 2, and the rest is done until N candidate nodes are selected; if the CPU utilization is equal to Ut until at the edge compute nodemaxAnd stopping selecting when the N candidate NODEs are not selected, taking the edge computing NODEs selected in the previous rounds as candidate NODEs, and recording the set of the candidate edge computing NODEs as { NODE _1, NODE _ 2.,. NODE _ N }.
4. The industrial internet of things-oriented edge computing task offloading method of claim 1, wherein: the step S4 includes:
taking the length and priority of the device task and the ratio of the transmission delay of task unloading and the maximum tolerance delay of the task as input variables of a fuzzy logic algorithm, wherein the fuzzy logic algorithm comprises the following steps:
(1) fuzzifying the length and priority of the device task and the ratio of the transmission delay of task unloading to the maximum tolerance delay of the task, and defining three variables mulen(u),μpri(u),μdelay(u) membership functions and corresponding fuzzy sets;
(2) selecting a corresponding fuzzy rule, obtaining the corresponding membership degree of each input variable through the membership function in the step (1), activating the fuzzy rule meeting the conditions, connecting the conditions of the fuzzy rule with logic and operation, and determining the fuzzy value meeting the conditions by adopting a minimum value method, namely:
wherein the content of the first and second substances,the number of fuzzy rules which are consistent with the fuzzy rule conclusion and are candidate edge computing NODEs NODE _ N is represented;
the final fuzzy value of the candidate edge calculation node is:
(3) defining fuzzy sets of edge compute nodes for optimal task offloadingAnd membership function A (u) thereof, using implication operator to obtain result { mu ] of precondition, i.e. length and priority of equipment task and ratio of transmission delay of task unloading to maximum tolerance delay of the taskNODE_1,μNODE_2,…,μNODE_NTruncating an edge calculation node fuzzy set unloaded by the optimal task, gathering the result of each rule after truncation, defuzzifying by using a centroid method to obtain a final result theta, wherein theta is more than 0 and less than or equal to 1, and when theta belongs to [0, y ∈ [ ]1]Selecting an edge computing NODE NODE _1 to unload the equipment task; when theta is equal to (y)1,y2]Selecting an edge computing NODE NODE _2 to unload the equipment task; when theta is equal to (y)N-1,yN]Then, selecting an edge computing NODE NODE _ N for unloading the device task, wherein y is more than 01<…<yN≤1。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911176742.9A CN110933157B (en) | 2019-11-26 | 2019-11-26 | Industrial Internet of things-oriented edge computing task unloading method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201911176742.9A CN110933157B (en) | 2019-11-26 | 2019-11-26 | Industrial Internet of things-oriented edge computing task unloading method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110933157A true CN110933157A (en) | 2020-03-27 |
CN110933157B CN110933157B (en) | 2022-03-11 |
Family
ID=69851210
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201911176742.9A Active CN110933157B (en) | 2019-11-26 | 2019-11-26 | Industrial Internet of things-oriented edge computing task unloading method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110933157B (en) |
Cited By (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111538571A (en) * | 2020-03-20 | 2020-08-14 | 重庆特斯联智慧科技股份有限公司 | Method and system for scheduling task of edge computing node of artificial intelligence Internet of things |
CN111918245A (en) * | 2020-07-07 | 2020-11-10 | 西安交通大学 | Multi-agent-based vehicle speed perception calculation task unloading and resource allocation method |
CN112486690A (en) * | 2020-12-11 | 2021-03-12 | 重庆邮电大学 | Edge computing resource allocation method suitable for industrial Internet of things |
CN113268341A (en) * | 2021-04-30 | 2021-08-17 | 国网河北省电力有限公司信息通信分公司 | Distribution method, device, equipment and storage medium of power grid edge calculation task |
CN113886094A (en) * | 2021-12-07 | 2022-01-04 | 浙江大云物联科技有限公司 | Resource scheduling method and device based on edge calculation |
CN114928607A (en) * | 2022-03-18 | 2022-08-19 | 南京邮电大学 | Collaborative task unloading method for multilateral access edge calculation |
CN115002108A (en) * | 2022-05-16 | 2022-09-02 | 电子科技大学 | Networking and task unloading method for serving smart phone as computing service node |
CN115665160A (en) * | 2022-10-14 | 2023-01-31 | 四川轻化工大学 | Multi-access edge computing system and method for electric power safety tool |
US20230217307A1 (en) * | 2020-10-20 | 2023-07-06 | L3Vel, Llc | Edge computing platform based on wireless mesh architecture |
CN116582540A (en) * | 2023-07-10 | 2023-08-11 | 北京智芯微电子科技有限公司 | Communication collaboration method and device for edge calculation, computer equipment and storage medium |
CN117349031A (en) * | 2023-12-05 | 2024-01-05 | 成都超算中心运营管理有限公司 | Distributed super computing resource scheduling analysis method, system, terminal and medium |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108777852A (en) * | 2018-05-16 | 2018-11-09 | 国网吉林省电力有限公司信息通信公司 | A kind of car networking content edge discharging method, mobile resources distribution system |
CN109005572A (en) * | 2018-08-20 | 2018-12-14 | 重庆邮电大学 | The access discharging method of mobile cloud service based on game theory |
CN109302709A (en) * | 2018-09-14 | 2019-02-01 | 重庆邮电大学 | The unloading of car networking task and resource allocation policy towards mobile edge calculations |
CN109684075A (en) * | 2018-11-28 | 2019-04-26 | 深圳供电局有限公司 | A method of calculating task unloading is carried out based on edge calculations and cloud computing collaboration |
CN109788069A (en) * | 2019-02-27 | 2019-05-21 | 电子科技大学 | Calculating discharging method based on mobile edge calculations in Internet of Things |
US20190174276A1 (en) * | 2017-12-01 | 2019-06-06 | Veniam, Inc. | Systems and methods for the data-driven and distributed interoperability between nodes to increase context and location awareness in a network of moving things, for example in a network of autonomous vehicles |
-
2019
- 2019-11-26 CN CN201911176742.9A patent/CN110933157B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20190174276A1 (en) * | 2017-12-01 | 2019-06-06 | Veniam, Inc. | Systems and methods for the data-driven and distributed interoperability between nodes to increase context and location awareness in a network of moving things, for example in a network of autonomous vehicles |
CN108777852A (en) * | 2018-05-16 | 2018-11-09 | 国网吉林省电力有限公司信息通信公司 | A kind of car networking content edge discharging method, mobile resources distribution system |
CN109005572A (en) * | 2018-08-20 | 2018-12-14 | 重庆邮电大学 | The access discharging method of mobile cloud service based on game theory |
CN109302709A (en) * | 2018-09-14 | 2019-02-01 | 重庆邮电大学 | The unloading of car networking task and resource allocation policy towards mobile edge calculations |
CN109684075A (en) * | 2018-11-28 | 2019-04-26 | 深圳供电局有限公司 | A method of calculating task unloading is carried out based on edge calculations and cloud computing collaboration |
CN109788069A (en) * | 2019-02-27 | 2019-05-21 | 电子科技大学 | Calculating discharging method based on mobile edge calculations in Internet of Things |
Non-Patent Citations (2)
Title |
---|
KE ZHANG: "Optimal Delay Constrained Offloading for Vehicular Edge Computing Networks", 《IEEE ICC 2017 AD-HOC AND SENSOR NETWORKING SYMPOSIUM》 * |
谢人超: "移动边缘计算卸载技术综述", 《通信学报》 * |
Cited By (17)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111538571B (en) * | 2020-03-20 | 2021-06-29 | 重庆特斯联智慧科技股份有限公司 | Method and system for scheduling task of edge computing node of artificial intelligence Internet of things |
CN111538571A (en) * | 2020-03-20 | 2020-08-14 | 重庆特斯联智慧科技股份有限公司 | Method and system for scheduling task of edge computing node of artificial intelligence Internet of things |
CN111918245A (en) * | 2020-07-07 | 2020-11-10 | 西安交通大学 | Multi-agent-based vehicle speed perception calculation task unloading and resource allocation method |
US20230217307A1 (en) * | 2020-10-20 | 2023-07-06 | L3Vel, Llc | Edge computing platform based on wireless mesh architecture |
CN112486690B (en) * | 2020-12-11 | 2024-01-30 | 重庆邮电大学 | Edge computing resource allocation method suitable for industrial Internet of things |
CN112486690A (en) * | 2020-12-11 | 2021-03-12 | 重庆邮电大学 | Edge computing resource allocation method suitable for industrial Internet of things |
CN113268341A (en) * | 2021-04-30 | 2021-08-17 | 国网河北省电力有限公司信息通信分公司 | Distribution method, device, equipment and storage medium of power grid edge calculation task |
CN113886094A (en) * | 2021-12-07 | 2022-01-04 | 浙江大云物联科技有限公司 | Resource scheduling method and device based on edge calculation |
CN114928607A (en) * | 2022-03-18 | 2022-08-19 | 南京邮电大学 | Collaborative task unloading method for multilateral access edge calculation |
CN114928607B (en) * | 2022-03-18 | 2023-08-04 | 南京邮电大学 | Collaborative task unloading method for polygonal access edge calculation |
CN115002108A (en) * | 2022-05-16 | 2022-09-02 | 电子科技大学 | Networking and task unloading method for serving smart phone as computing service node |
CN115665160A (en) * | 2022-10-14 | 2023-01-31 | 四川轻化工大学 | Multi-access edge computing system and method for electric power safety tool |
CN115665160B (en) * | 2022-10-14 | 2024-02-20 | 四川轻化工大学 | Multi-access edge computing system and method for electric power safety tools |
CN116582540B (en) * | 2023-07-10 | 2024-01-16 | 北京智芯微电子科技有限公司 | Communication collaboration method and device for edge calculation, computer equipment and storage medium |
CN116582540A (en) * | 2023-07-10 | 2023-08-11 | 北京智芯微电子科技有限公司 | Communication collaboration method and device for edge calculation, computer equipment and storage medium |
CN117349031A (en) * | 2023-12-05 | 2024-01-05 | 成都超算中心运营管理有限公司 | Distributed super computing resource scheduling analysis method, system, terminal and medium |
CN117349031B (en) * | 2023-12-05 | 2024-02-13 | 成都超算中心运营管理有限公司 | Distributed super computing resource scheduling analysis method, system, terminal and medium |
Also Published As
Publication number | Publication date |
---|---|
CN110933157B (en) | 2022-03-11 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110933157B (en) | Industrial Internet of things-oriented edge computing task unloading method | |
Zhang et al. | A hierarchical game framework for resource management in fog computing | |
CN112882815B (en) | Multi-user edge calculation optimization scheduling method based on deep reinforcement learning | |
Sun et al. | Autonomous resource slicing for virtualized vehicular networks with D2D communications based on deep reinforcement learning | |
CN111093203B (en) | Service function chain low-cost intelligent deployment method based on environment perception | |
CN108566659B (en) | 5G network slice online mapping method based on reliability | |
CN112039965B (en) | Multitask unloading method and system in time-sensitive network | |
US7826365B2 (en) | Method and apparatus for resource allocation for stream data processing | |
CN112118312B (en) | Network burst load evacuation method facing edge server | |
CN114422349B (en) | Cloud-edge-end-collaboration-based deep learning model training and reasoning architecture deployment method | |
CN111538571B (en) | Method and system for scheduling task of edge computing node of artificial intelligence Internet of things | |
CN110365748A (en) | Treating method and apparatus, storage medium and the electronic device of business datum | |
WO2023040022A1 (en) | Computing and network collaboration-based distributed computation offloading method in random network | |
CN112486690A (en) | Edge computing resource allocation method suitable for industrial Internet of things | |
CN112650581A (en) | Cloud-side cooperative task scheduling method for intelligent building | |
CN114745317A (en) | Computing task scheduling method facing computing power network and related equipment | |
CN104901989A (en) | Field service providing system and method | |
CN114595049A (en) | Cloud-edge cooperative task scheduling method and device | |
CN112162789A (en) | Edge calculation random unloading decision method and system based on software definition | |
Cao et al. | A deep reinforcement learning approach to multi-component job scheduling in edge computing | |
CN113315669B (en) | Cloud edge cooperation-based throughput optimization machine learning inference task deployment method | |
CN106211344A (en) | A kind of Ad Hoc network bandwidth management method based on context aware | |
CN116166444B (en) | Collaborative reasoning method oriented to deep learning hierarchical model | |
CN105335376B (en) | A kind of method for stream processing, apparatus and system | |
CN117156492A (en) | Deep reinforcement learning-based dual-time-scale resource allocation method for joint service caching, communication and calculation |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
TR01 | Transfer of patent right | ||
TR01 | Transfer of patent right |
Effective date of registration: 20240430 Address after: 400020 12-1 to 12-12, building 1, No. 8, West Ring Road, Jiangbei District, Chongqing Patentee after: Chongqing ruanjiang Turing Artificial Intelligence Technology Co.,Ltd. Country or region after: China Address before: 400065 Chongqing Nan'an District huangjuezhen pass Chongwen Road No. 2 Patentee before: CHONGQING University OF POSTS AND TELECOMMUNICATIONS Country or region before: China |