CN117440461A - Routing system based on edge calculation - Google Patents
Routing system based on edge calculation Download PDFInfo
- Publication number
- CN117440461A CN117440461A CN202311464017.8A CN202311464017A CN117440461A CN 117440461 A CN117440461 A CN 117440461A CN 202311464017 A CN202311464017 A CN 202311464017A CN 117440461 A CN117440461 A CN 117440461A
- Authority
- CN
- China
- Prior art keywords
- edge computing
- edge
- gateway
- computing gateway
- computing
- 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.)
- Pending
Links
- 238000004364 calculation method Methods 0.000 title description 9
- 238000012545 processing Methods 0.000 claims abstract description 5
- 230000001934 delay Effects 0.000 claims abstract description 4
- 230000005540 biological transmission Effects 0.000 claims description 5
- 238000013139 quantization Methods 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 2
- 230000000295 complement effect Effects 0.000 claims 3
- 238000004891 communication Methods 0.000 abstract description 10
- 238000012544 monitoring process Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 4
- 239000013307 optical fiber Substances 0.000 description 3
- 239000012212 insulator Substances 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004587 chromatography analysis Methods 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000000034 method Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W40/00—Communication routing or communication path finding
- H04W40/02—Communication route or path selection, e.g. power-based or shortest path routing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/38—Services specially adapted for particular environments, situations or purposes for collecting sensor information
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W40/00—Communication routing or communication path finding
- H04W40/34—Modification of an existing route
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W88/00—Devices specially adapted for wireless communication networks, e.g. terminals, base stations or access point devices
- H04W88/16—Gateway arrangements
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention discloses a routing system based on edge computing, which belongs to the field of edge computing networks and comprises the steps of defining each edge computing gateway Gi, wherein the edge computing gateway Gi can perform real-time reasoning and processing on sigma (Gi) sensor accesses, and is provided with rho (Gi) sensor accesses to the edge computing gateway Gi, the bandwidths of the edge computing gateways Gi and Gj are w (Gi, gj), the delays of the edge computing gateways Gi and Gj are l (Gi, gj), a central server S serves as a leader of the distributed edge computing system, knows the spending b (Gi, gj) of each link and controls Gi communication, so that the dynamic optimal routing between the edge computing gateways and the routing of computing task transfer can be realized, the overall computing cost of the edge computing network can be reduced, the computing task can be matched with the computing capacity of the edge computing gateway, the robustness of the edge computing system can be enhanced, and the high SLA of the edge computing system is possible.
Description
Technical Field
The present invention relates to the field of edge computing networks, and more particularly, to an edge computing-based routing system.
Background
With the rapid development of the internet of things (Internet of Things, ioT), more and more devices and sensors are connected to the internet, forming a huge network ecosystem. However, traditional cloud computing modes present challenges for large-scale internet of things applications, such as high latency, limited network bandwidth, and the like. To solve these problems, edge computation (Edge Computing) has been developed. Edge computing utilizes computing resources close to the data source to move computing and storage functions to the network edge, providing faster, reliable, and secure services.
Taking power transmission and transformation equipment state monitoring data as an example, the current-stage online monitoring system (transformer oil chromatography, partial discharge, insulator leakage current, line icing, substation image monitoring and the like) comprises structured (online monitoring values, standing account information and the like) and unstructured data (images and the like). And a great deal of infrared, ultraviolet images and video information are generated by the line inspection of unmanned aerial vehicles which are currently developing. As the scale of the monitoring system increases, the amount of data will also multiply, which undoubtedly increases the demands on the computing power of the edge computing gateway.
In the process of deploying and using the edge computing, we can find that in some cases, an observation sensor needs to be added, or an inference model of the edge computing gateway is updated, and then the situation that the sensor accessed by the edge computing gateway generates data exceeding the processing capability of the edge computing gateway may occur. If the data are not transmitted to other edge nodes, schemes such as quantization, model precision reduction and the like are needed to accelerate the model recognition speed, but the accuracy of a model result is likely to be reduced, and the reliability of a calculation result of an edge calculation system is likely to be reduced. Or because the third party construction breaks the optical fiber or the operator base station is overloaded and dropped, the link between part of the edge computing gateway and the cloud server breaks down, so that the abnormal result cannot be recorded in the cloud server in time and can be known by a user.
Therefore, there is a need for an edge computing based routing system that allows computing tasks and signaling messages to flow in the computing power network of an edge computing gateway, matching the computing tasks to the computing power of the edge computing gateway, while enhancing the robustness of the edge computing system.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a routing system based on edge calculation, which aims to overcome the limitations of the traditional cloud calculation and edge calculation modes and enable the edge calculation to be more efficient. In a conventional edge computing system, an edge computing gateway is directly connected to a cloud server using a protocol such as an optical fiber or a radio access network (Radio Access Network), and does not communicate with other edge computing gateways.
In the system, the edge computing gateway and part of other edge computing gateways are mutually communicated by using protocols such as optical fiber or radio access network (Radio Access Network) or IEEE 802.11, and the like, and different communication delays, bandwidths and costs can exist.
In the edge computing system, when the computing capacity of each edge computing gateway is larger than the computing task cost, the edge computing gateway can dynamically select the optimal route by combining the communication delay and the cost of the edge computing network, so that the communication cost can be reduced, the damage of link faults can be resisted, and the robustness of the edge computing system is enhanced.
When the computing task overhead is larger than the computing capacity of some edge computing gateways, the edge computing gateways can combine the communication bandwidth and cost of the edge computing network and the computing capacity and workload of each edge computing gateway to determine the route of one or more computing tasks, and the overloaded computing tasks are unloaded to the adjacent idle edge computing gateways.
The beneficial effects are that:
according to the technical scheme provided by the invention, the dynamic optimal route between the edge computing gateways and the route for delivering the computing tasks can be realized by introducing the route scheme related to communication delay, bandwidth and overhead, so that the total computing cost of the edge computing network is reduced, the computing tasks are matched with the computing capacity of the edge computing gateways, the robustness of the edge computing system can be enhanced, and the high SLA of the edge computing system is possible.
Drawings
FIG. 1 is a schematic overall flow chart of the present invention;
FIG. 2 is a diagram showing the communication relationship between the wireless edge computing gateway and the receiving sensors and the central server S according to the present invention;
FIG. 3 is a diagram showing the communication relationship between the wireless edge computing gateway and the receiving sensors and the central server S in the case of the link disconnection of the present invention;
FIG. 4 is a diagram showing the communication relationship between the wireless edge computing gateway and the receiving sensors and the central server S in the case of the link disconnection of the present invention;
FIG. 5 is a schematic diagram of dynamically offloading computing task routing in the event of an edge computing gateway overload in accordance with the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention; it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments, and that all other embodiments obtained by persons of ordinary skill in the art without making creative efforts based on the embodiments in the present invention are within the protection scope of the present invention.
Examples:
as shown in fig. 1 to fig. 5, for the application scenario in operation, the inference models of all edge computing gateways are the same, and all edge computing gateways are under one tenant, so that privacy protection actions such as federal learning and the like do not need to be considered, and sensors for collecting data are the same, and data with approximate unit computing power processing time can be generated in the same time unit under normal conditions, and are called as equivalent access sensors, hereinafter referred to as sensors.
Defining each edge computing gateway Gi, under the current version of the model, gi can perform real-time reasoning and processing on σ (Gi) sensor accesses, with ρ (Gi) sensor access edge computing gateways Gi, gi and Gj having bandwidths w (Gi, gj), gi and Gj having delays l (Gi, gj), the central server S acting as a leader of the distributed edge computing system, knowing the overhead b (Gi, gj) of each link and controlling Gi communication.
(1) Gi measures w (Gi, gj) and l (Gi, gj) for Gj using tools such as Iperf3 and MTR, and if the link is broken, places w (Gi, gj) at 0,l (Gi, gj) at infinity, and reports w (Gi, gj) and l (Gi, gj).
(2) S distributes the inference model, each Gi reports σ (Gi) and ρ (Gi).
(3) If there is Gi satisfying σ (Gi) < ρ (Gi), and there is an loadable trajectory in the network flow, then Dijkstra's algorithm solution is performed for each Gi. The cost between the nodes Gi and Gj is set as b (Gi, gj), the lowest cost transmission path with free computing power Gj from Gi to σ (Gj) > ρ (Gj) is obtained, the computing task flow from Gi to Gj is increased until the bandwidth of a certain link in Gi to Gj reaches the capacity upper limit w or Gj has no redundant computing power. If there is still a computation task flow to be sent, an idle computation node of the secondary low-cost transmission path needs to be selected until computation meets distribution.
(4) When Gi satisfies σ (Gi) > = ρ (Gi), gi is solved by Dijkstra algorithm using the delay l (Gi, gj) and the overhead b (Gi, gj) of each node by using the concept of OSPF based on cost or Babel based on distance vector routing. The cost between Gi and Gj nodes is set to l (Gi, gk) +α×b (Gi, gj), and the minimum Σ (l (Gj, gk) +α×b (Gj, gk)) to Gj, gk e Gi, S is obtained, and the parameter α varies with the optimization target. The optimal path from Gi to S can be dynamically changed, and the robustness of the edge computing system is enhanced.
(5) If there is no scalable track, i.e. there is no feasible flow of the supply and demand network flow, the edge computing system will operate downgrade, and model quantization is required to reduce accuracy and speed up operation, and personnel are required to intervene to adjust network state or increase computing power of the edge computing gateway.
(6) Returning to (1) when the network state of the edge computing system changes, the inference model changes, the number of access sensors changes or a certain time is reached;
it should be noted that, as the sensor may move, for example, different edge computing gateways are dynamically accessed on the unmanned aerial vehicle; the sensor may also temporarily generate more calculation task demands, for example, the unmanned aerial vehicle beats an insulator with suspected breakage in flight to increase the detection density of the video input, so that the network topology needs to be updated regularly and the number of equivalent access sensors of each edge computing gateway needs to be estimated, and therefore, step (6) needs to be executed regularly.
Judging whether an loadable track exists in the network flow:
for any number of edge computation gateways selected to be Q, the rest of the edge computation gateway sets are complements T of Q, the section C (S, T) of the network flow between Q and T is the sum of the bandwidths from Q to T, the total computation gap in Q is sigma (rho (Gi) -sigma (Gi)) to all Gi in Q, and if C (S, T) is smaller than the computation gap, no loadable track exists.
The above description is only of the preferred embodiments of the present invention; the scope of the invention is not limited in this respect. Any person skilled in the art, within the technical scope of the present disclosure, may apply to the present invention, and the technical solution and the improvement thereof are all covered by the protection scope of the present invention.
Claims (7)
1. A routing system based on edge computation, characterized in that: each edge computing gateway Gi is defined, which can perform real-time reasoning and processing on the data accessed by sigma (Gi) sensors, and is provided with rho (Gi) sensors accessing the edge computing gateway Gi, the bandwidths of the edge computing gateway Gi and the edge computing gateway Gj are w (Gi, gj), the delays of the edge computing gateway Gi and the edge computing gateway Gj are denoted as l (Gi, gj), and the central server S acts as a leader of the distributed edge computing system, knows the overhead b (Gi, gj) of each link and controls the edge computing gateway Gi to communicate.
2. An edge computation based routing system according to claim 1, wherein: the edge computing gateway Gi measures the bandwidth w (Gi, gj) and the delay l (Gi, gj) for Gj by using tools such as Iperf3 and MTR, and if the link is disconnected, the bandwidth w (Gi, gj) is set to 0, the delay l (Gi, gj) is set to infinity, and the bandwidth w (Gi, gj) and the delay l (Gi, gj) are reported.
3. An edge computation based routing system according to claim 2, wherein: the central server S distributes the inference model, and each edge computing gateway Gi reports σ (Gi) and ρ (Gi).
4. A routing system based on edge computation according to claim 3, wherein: if the edge computing gateway Gi satisfies σ (Gi) < ρ (Gi), and there is an loadable track in this network flow, then a Dijkstra algorithm solution is performed on each edge computing gateway Gi, the overhead between the edge computing gateway Gi and the edge computing gateway Gj node is set to b (Gi, gj), a lowest cost transmission path with free computing force Gj from the edge computing gateway Gi to σ (Gj) > ρ (Gj) is obtained, a computing task flow from the edge computing gateway Gi to the edge computing gateway Gj is increased until the bandwidth of a link from the edge computing gateway Gi to the edge computing gateway Gj reaches its capacity upper limit bandwidth w (Gi, gj) or the edge computing gateway Gj has no redundant computing force, and if there is still a computing task flow to be transmitted, then a free computing force node of a low cost transmission path needs to be selected until the computing force satisfies the allocation.
5. An edge computation based routing system according to claim 4, wherein: if the edge computing gateway Gi satisfies σ (Gi) > = ρ (Gi), the edge computing gateway Gi is solved by Dijkstra algorithm using the delay l (Gi, gj) and the overhead b (Gi, gj) of each node by using the concept of OSPF based on cost or Babel based on distance vector routing, and the cost between the edge computing gateway Gi and the edge computing gateway Gj node is set to be the delay l (Gi, gj) +α×b (Gi, gj), the optimal path of Σ (l (Gj, gk) +α×b (Gj, gk)) to Gj, gk e Gi, S is obtained, and the parameter α varies with the optimization target, so that the obtained optimal path from the edge computing gateway Gi to the center server S can dynamically vary, and the robustness of the edge computing system is enhanced.
6. An edge computation based routing system according to claim 5, wherein: optionally selecting a plurality of edge computing gateways Gi to make the set of the edge computing gateways Gi be Q, leaving an edge computing gateway set as a complement T of Q, wherein a section C (S, T) of a network flow between the edge computing gateway set Q and the complement T is a sum of bandwidths from the edge computing gateway set Q to the complement T, a total computing power gap in the edge computing gateway set Q is Σ (ρ (Gi) - σ (Gi)) and all edge computing gateways Gi in the edge computing gateway set Q, and if the section C (S, T) of the network flow is smaller than the computing power gap, no load increasing track exists.
7. An edge computation based routing system according to claim 6, wherein: if there is no scalable track, i.e. there is no feasible flow of the supply and demand network flow, the edge computing system will operate downgrade, and model quantization is required to reduce accuracy and speed up operation, and personnel are required to intervene to adjust network state or increase computing power of the edge computing gateway.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311464017.8A CN117440461A (en) | 2023-11-06 | 2023-11-06 | Routing system based on edge calculation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311464017.8A CN117440461A (en) | 2023-11-06 | 2023-11-06 | Routing system based on edge calculation |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117440461A true CN117440461A (en) | 2024-01-23 |
Family
ID=89551306
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311464017.8A Pending CN117440461A (en) | 2023-11-06 | 2023-11-06 | Routing system based on edge calculation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117440461A (en) |
-
2023
- 2023-11-06 CN CN202311464017.8A patent/CN117440461A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP3213546B1 (en) | System and method for placing virtual serving gateways for mobility management | |
| Natarajan et al. | An IoT and machine learning‐based routing protocol for reconfigurable engineering application | |
| CN108076158B (en) | Minimum load route selection method and system based on naive Bayes classifier | |
| CN111787069A (en) | Method, device and equipment for processing service access request and computer storage medium | |
| CN109617810B (en) | Data transmission method and device | |
| WO2021244247A1 (en) | Data message forwarding method, network node, system, and storage medium | |
| Tham et al. | A load balancing scheme for sensing and analytics on a mobile edge computing network | |
| US20110206034A1 (en) | Route allocation apparatus and method | |
| Nguyen et al. | Flexible computation offloading in a fuzzy-based mobile edge orchestrator for IoT applications | |
| JP6660283B2 (en) | Traffic demand forecasting device, traffic demand forecasting method, and program | |
| CN113328953A (en) | Method, device and storage medium for network congestion adjustment | |
| Hussain et al. | Towards minimizing delay and energy consumption in vehicular fog computing (VFC) | |
| Zhang et al. | Multi-attribute-based QoS-aware virtual network function placement and service chaining algorithms in smart cities | |
| US11387926B2 (en) | Efficient transfer of sensor data on dynamic software defined network (SDN) controlled optical network | |
| Islam et al. | Software-defined network-based proactive routing strategy in smart power grids using graph neural network and reinforcement learning | |
| CN117440461A (en) | Routing system based on edge calculation | |
| Atat et al. | Stochastic geometry model for interdependent cyber-physical communication-power networks | |
| Zhao et al. | Message‐sensing classified transmission scheme based on mobile edge computing in the Internet of Vehicles | |
| CN116455817A (en) | Software-defined cloud network fusion architecture and route implementation method | |
| Liang et al. | Queue‐based congestion detection and multistage rate control in event‐driven wireless sensor networks | |
| CN114448838B (en) | System reliability evaluation method | |
| Lin et al. | ALPS: An adaptive link-state perception scheme for software-defined vehicular networks | |
| Hu et al. | A packet scheduling method based on dynamic adjustment of service priority for electric power wireless communication network | |
| Singh et al. | Analysis of energy optimization approaches in internet of everything: An SDN prospective | |
| Lourenço et al. | A deep neural network with a fuzzy multi-objective optimization model for fault analysis in an elastic optical network |
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 |