CN113810916B - Multi-server mixed deployment architecture and method in 5G/6G edge computing scene - Google Patents
Multi-server mixed deployment architecture and method in 5G/6G edge computing scene Download PDFInfo
- Publication number
- CN113810916B CN113810916B CN202111073797.4A CN202111073797A CN113810916B CN 113810916 B CN113810916 B CN 113810916B CN 202111073797 A CN202111073797 A CN 202111073797A CN 113810916 B CN113810916 B CN 113810916B
- Authority
- CN
- China
- Prior art keywords
- server
- computing
- deployment
- edge
- dynamic
- 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.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
- H04W16/18—Network planning tools
-
- 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
- H04L67/104—Peer-to-peer [P2P] networks
- H04L67/1042—Peer-to-peer [P2P] networks using topology management mechanisms
-
- 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
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/70—Services for machine-to-machine communication [M2M] or machine type communication [MTC]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Business, Economics & Management (AREA)
- General Business, Economics & Management (AREA)
- Computer And Data Communications (AREA)
Abstract
The invention discloses a multi-server mixed deployment architecture and a method under a 5G/6G edge computing scene, wherein the method comprises the following steps: the first stage faces a plurality of dynamic edge servers, and outputs an equipment distribution scheme through operation genetic algorithm optimization; and in the second stage, the single dynamic edge server is oriented, through Gibbs sampling, the path is planned by the output of the first step, the task is scheduled by the second step, and the first step and the second step are circularly and iteratively replaced to find the optimal solution. Aiming at the problems of over-dense deployment and over-high expenditure of the edge server, the invention provides a deployment framework, a joint optimization model and a two-stage optimization algorithm of the moving edge, establishes parallel execution between the movement and the operation of the edge server, reduces the expenditure of the deployment server and improves the resource utilization rate.
Description
Technical Field
The invention belongs to the technical field of Internet of things and edge computing, and particularly relates to a multi-server hybrid deployment architecture and a multi-server hybrid deployment method in a 5G/6G edge computing scene.
Background
With the development of 5G/6G communication technology and Internet of Things (IoT), the number of "killer" applications represented by autopilot, augmented/virtual reality games has increased dramatically, and such applications have the characteristics of being time-delay sensitive and computationally intensive. The terminal equipment of the Internet of things does not have the computing capacity of locally running the application; although cloud computing has strong computing power, the end-to-end delay is too large due to too long transmission distance, and the requirement of extremely low delay of application cannot be met.
Edge computing is a computing paradigm for sinking computing resources to the edge of a network, overcomes the defect of too long transmission delay of cloud computing, and can meet the requirements of high computing amount and extremely low delay of application. The development of the 5G/6G communication technology greatly reduces the transmission delay of the last hop in the edge calculation, and further improves the performance of the edge network. Meanwhile, the new characteristics introduced by the 5G/6G communication technology bring many challenges to edge calculation.
The excessive deployment cost is one of the main challenges facing 5G/6G edge computing, i.e. a large number of edge servers are needed to complete the full coverage of the target network. Due to the fact that the 5G/6G communication distance is greatly shortened, more servers need to be deployed to cover the same range to achieve the same coverage effect of 4G. Specifically, the effective communication range of the base station is reduced from 1-25 kilometers in a 4G network to 100-300 meters in a 5G network and several tens of meters in a 6G network. The deployment cost of 5G/6G is about 100-10000 times of that of 4G edge network when the terminal nodes cover the same range. Considering the rapid increase of the delay-sensitive application in the edge calculation, a more dense server needs to be deployed to meet the extremely high requirement of the application on Quality-of-Service (QoS), and the actual deployment cost may be higher.
Disclosure of Invention
The invention provides a multi-server mixed deployment architecture under a 5G/6G edge computing scene, which maximizes the resource utilization rate of servers and reduces the deployment number of edge servers through a mixed deployment architecture consisting of dynamic and static edge servers, so as to solve the problem of intensive deployment of the servers in a 5G/6G edge network, reduce the deployment cost and improve the resource utilization rate.
The invention is realized by the following technical scheme:
a multi-server mixed deployment architecture under a 5G/6G edge computing scene comprises a plurality of dynamic servers and a plurality of static servers, and a mixed deployment architecture composed of the static and dynamic edge servers is used for servicing delay sensitive or computation intensive tasks in edge computing, so that service coverage of nodes of the Internet of things under the edge scene is completed.
Preferably, the dynamic edge server has "compute-move parallelism," which is the decoupling of the compute offload process into request uploads, task computations, and result returns, enabling the dynamic edge server to perform task processing during moves.
Preferably, the number of static and dynamic edge servers is based on a two-stage optimization algorithm to determine a minimum deployment number of servers for a given edge network device; the location of the static edge server is a deployment location of the multiple static servers determined based on a two-phase optimization algorithm.
Preferably, the multi-server hybrid deployment architecture under the 5G/6G edge computing scene further comprises a path planning module and a task scheduling module which jointly optimize the multi-dynamic server based on a two-stage optimization algorithm.
Preferably, the dynamic edge server is preferentially used for processing compute-intensive and delay-tolerant tasks, and the static edge server is preferentially used for processing delay-sensitive tasks.
On the other hand, the application also relates to a multi-server hybrid deployment method under the 5G/6G edge computing scene, which comprises the following steps:
s1, using a mixed deployment framework consisting of dynamic and static edge servers;
s2, providing an optimization problem model based on the hybrid deployment architecture;
s3, optimizing the deployment scheme of the dynamic and static edge servers;
and S4, completing full coverage of the nodes of the Internet of things in the 5G/6G edge network according to the optimization result.
Preferably, the step S3 includes a two-stage optimization algorithm for jointly optimizing path planning and task scheduling of the multi-dynamic server, and the algorithm includes the following steps:
in the first stage, equipment allocation based on the Elitism Genetic Algorithm (EGA) allocates a plurality of internet of things equipment to multiple dynamic edge servers, so that a multi-server planning problem is converted into a single-server planning problem.
And in the second stage, based on the single dynamic server scheduling of Gibbs sampling, path planning and task scheduling of a single server are optimized in a combined mode, so that the total completion time delay is minimized.
Preferably, the second stage further comprises: iterative updating based on Gibbs sampling solves two coupled sub-problems; path planning is carried out based on a 2-OPT algorithm, so that the path planning method is suitable for a 'movement-calculation parallelism' movement mechanism; based on the moving mechanism, a heuristic task scheduling algorithm is further provided, and the resource utilization rate is improved.
Preferably, the optimization object of the two-stage optimization algorithm is to determine the minimum deployment number of the servers and the deployment positions of the static servers in the given edge network.
The multi-server mixed deployment architecture under the 5G/6G edge computing scene has the following characteristics:
a mixed deployment framework consisting of dynamic and static edge servers is used;
the dynamic edge server can process tasks in the moving process by utilizing the 'moving-computing parallelism';
optimizing a deployment plan using an optimization problem model based on the hybrid deployment framework;
determining the minimum deployment number of the servers of the given edge network equipment by using a two-stage optimization algorithm;
jointly optimizing path planning and task scheduling of the multiple dynamic servers by using a two-stage optimization algorithm;
a two-phase optimization algorithm is used to determine the deployment of multiple static servers.
The invention uses static and dynamic edge servers to solve various computing requests in the edge network, so as to reduce the deployment cost of the servers. The dynamic edge server is preferentially used for processing the tasks which are intensive in calculation and tolerant to time delay, and the static edge server is preferentially used for processing the tasks which are sensitive to time delay.
Preferably, the method decouples the computation unloading process into request uploading, task computation and result returning, and enables the mobile server to perform task processing in the moving process by utilizing the 'moving-computation parallelism'.
The invention provides an optimization problem model based on the hybrid deployment architecture, and optimizes the deployment scheme of the dynamic and static edge servers, so that the full coverage of the nodes of the Internet of things in the 5G/6G edge network is completed with the minimum deployment cost.
Preferably, the specific optimization objectives of the optimization model are to determine device-server allocation, dynamic server movement paths and task scheduling.
The invention proposes a two-phase optimization algorithm for planning a mobile edge server. The first stage of the method is based on the Elite genetic algorithm to distribute equipment, and distributes numerous Internet of things equipment to the multi-mobile edge server, so that the multi-server planning problem is converted into the single-server planning problem. And in the second stage, single mobile server scheduling is carried out based on Gibbs sampling, and path planning and task scheduling of a single server are optimized in a combined mode so as to minimize total completion delay.
Preferably, the first stage of the algorithm performs the assignment of the equipment based on the elite genetic algorithm. The chromosome representation, fitness calculation, population initialization and corresponding genetic steps in the EGA are customized for the special movement mechanism of the edge architecture, so that the algorithm can quickly converge to a near-optimal solution in polynomial time.
Preferably, the planning of the single dynamic server is completed in the second stage of the algorithm based on Gibbs sampling, and the coupling sub-problem is optimized in a combined manner: path planning and task scheduling. Specifically, a server moving path is optimized based on a 2-OPT algorithm, and the TSP problem is completed in polynomial time; and task scheduling is performed based on a heuristic algorithm, so that the parallelism of server movement and task calculation is further improved, and the resource utilization rate is improved.
The invention proposes an implementation scheme for determining the minimum deployment number and the static server deployment position of the servers for a given network based on the two-stage optimization algorithm.
The invention has the following advantages and beneficial effects:
1. aiming at the problems of limited coverage area of a 5G/6G edge server and intensive server deployment, the invention provides a mixed deployment architecture, and by utilizing different application requirements in dynamic and static edge server service edge computing, the dynamic server innovatively utilizes 'movement-computing parallelism', so that the server can complete processing of computing requests in the moving process, the resource utilization rate is greatly improved, and the deployment cost is reduced.
2. The invention provides a two-stage optimization algorithm to plan the server deployment of the framework, and jointly optimizes two coupled sub-problems of path planning and task scheduling of the dynamic server. For the NP-hard problem, the algorithm can converge to a near optimal solution in polynomial time.
3. For a given network, the two-phase algorithm provided by the invention can also be used for determining the number of servers to be deployed and the deployment position of the static server in a given scene, thereby further perfecting the deployment scheme of the edge architecture.
Drawings
The accompanying drawings, which are included to provide a further understanding of the embodiments of the invention and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the invention and together with the description serve to explain the principles of the invention. In the drawings:
fig. 1 is a schematic flow chart of a two-stage optimization algorithm proposed by the present invention.
FIG. 2 is an algorithm flow diagram of the first phase of a two-phase optimization algorithm.
FIG. 3 is an example of genetic operations in the first phase of a two-phase optimization algorithm.
FIG. 4 is an algorithm flow diagram of the second phase of the two-phase optimization algorithm.
FIG. 5 is an example of task scheduling in the second phase of the two-phase optimization algorithm.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail below with reference to examples and accompanying drawings, and the exemplary embodiments and descriptions thereof are only used for explaining the present invention and are not meant to limit the present invention.
Example 1
The embodiment provides a multi-server hybrid deployment architecture in a 5G/6G edge computing scene. The framework of the embodiment provides an edge framework for mixed deployment of static and dynamic edge servers aiming at the problem of dense deployment caused by a small communication range of a 5G/6G server, so that the deployment cost of the server is reduced, and the resource utilization rate is improved.
The framework of the embodiment includes a deployment scheme of a static edge server and a dynamic edge server, wherein the static server is adopted to process a delay sensitive task, and the dynamic server is adopted to process a delay tolerant task. The 5G/6G edge architecture of the present embodiment is specifically as follows:
1. the static edge server deployment complies with the ETSI standard and may be deployed at an access point or a base station. The dynamic server is an edge server loaded on an unmanned vehicle or an unmanned aerial vehicle and has the moving capability.
2. Theoretically, task requests of edge calculation are classified into two categories, namely delay sensitivity and delay tolerance, according to whether the delay requirements are strict or not. For the time delay sensitive task, a static edge server is deployed to ensure the service quality; for the delay tolerant task, the dynamic edge server is deployed, so that the computing resource of the same server can be shared by nodes divided at multiple geographic positions, and the number of the servers to be deployed is effectively reduced.
3. In specific implementation, clear division standards are difficult to find for delay sensitive and delay tolerant tasks, and the two-stage optimization algorithm provided by the invention can skip the task type division link and automatically determine the deployment positions of the static and dynamic servers according to the configuration of a network.
Example 2
The present embodiment is a mobility mechanism using the dynamic server in the edge architecture in embodiment 1.
The dynamic edge server in the invention simultaneously allows two moving mechanisms of 'moving-computing parallel' and 'moving-computing serial', which are as follows:
1. the "move-compute parallel" mechanism. The invention innovatively utilizes the characteristic that the moving process of the server and the task calculating process in the dynamic service architecture can be parallel, and decouples the calculation unloading process into the request uploading, the task calculation and the result returning, so that the server does not stay in-situ service after receiving the request, and performs the task calculation in the moving process from the previous place to the next place. The mechanism utilizes the moving time of the server, and can greatly reduce the service completion time delay under the condition of extremely long task calculation time, so that fewer servers are deployed to meet the service quality requirement of the given Internet of things equipment.
2. The "move-compute serial" mechanism. The present invention does not implement a "move-compute parallelism" mechanism for all terminal nodes. For a task request with extremely small calculation amount, the round-trip movement cost is far larger than the calculation cost, the server stays in place according to the existing movement mechanism, and moves to the next node after the task calculation is finished.
3. The dynamic server selects the two mechanisms, and makes a decision after comprehensive consideration according to factors such as the state of a terminal node in the network, the computing resources of the server, the edge network environment and the like.
Example 3
The embodiment is a specific deployment method of the edge deployment architecture, namely a two-stage optimization algorithm, proposed in the foregoing embodiment 1 and embodiment 2, and the specific deployment method is as follows:
the two-stage optimization algorithm schedules the multi-mobile edge server and plans the mobile path and the task processing sequence of the multi-mobile edge server, so that the server can achieve the minimum total service time under the service quality requirement of the IoT equipment and the resource limit of the edge server, thereby reducing the deployment cost and improving the resource utilization rate. The algorithm execution process is shown in fig. 1, and specifically includes the following steps:
1. in the first stage, equipment allocation based on an elite genetic algorithm is adopted, namely a plurality of Internet of things equipment are allocated to a plurality of mobile edge servers, so that a multi-server planning problem is converted into a single-server planning problem, and the total time delay of a system is minimized. The algorithm implementation example is shown in fig. 2, and specifically includes:
1.1, randomly generating a primary population.
And generating an initial population of the genetic algorithm, wherein the process increases the constraint condition of the hamming distance to generate the initial population which is more uniformly distributed, thereby accelerating the convergence of the genetic algorithm. The expression of the chromosome and the calculated distance of the hamming distance are shown in the upper part of fig. 3, and the blue-labeled genes are terms for calculating the hamming distance.
And 1.2, iteratively updating based on an elite genetic algorithm.
1.2.1, calling a second-stage algorithm and calculating the adaptive value of each chromosome. The adaptive value can be used to evaluate the quality of equipment allocation to facilitate the selection of superior individuals in the population by the tournament polling algorithm, and is calculated as follows:
the adaptation value consists of two parts: optimizing the target total delay and the penalty item exceeding the service quality limit.
The first term of the fitness function is the optimal target total delay. C k Numbering chromosomes, namely an Internet of things equipment-edge server distribution scheme;
the delay for a single server is obtained by invoking the second stage of the algorithm described below (optimizing the movement path and task processing order for a single server), and the sum of the completion delays for each edge server is the total system delay allocated to the device indicated in this chromosome. Obviously, the total delay should be as small as possible.
The second term of the fitness value function is a penalty term for allocation schemes that exceed the quality of service limit,
time required for the server to complete the calculation of task i, ddl i The requirement of the task i on service delay is specified. It is clear that an allocation scheme that completes a task within a specified time is more advantageous, i.e. the value of the penalty term should be as large as possible.
The calculation of the fitness value provides an index for the tournament polling algorithm to select superior chromosomes, with chromosomes with larger values being more advantageous in competition.
1.2,2, chromosome selection stage. And determining the individuals with excellent fitness in the current generation population based on the championship polling algorithm as parent chromosomes generated by the next generation individuals.
1.2.3, cross-recombination and genetic mutation stages.
For a selected pair of chromosomes, cross points are randomly selected, gene segments of corresponding chromosomes are crossed, and a mapping relation is established. And modifying related genes according to the mapping relation, and eliminating the repeatedly distributed Internet of things equipment. And (3) a genetic mutation stage, wherein mutation points are randomly selected, and genes of the two mutation points are exchanged. An example of the above process is shown in fig. 3.
1.2.4, generating a new generation of population.
The new generation population consists of the individuals with the maximum adaptability in the current generation and the newly generated individuals through the steps 1.2.2 to 1.2.3.
1.3, continuously iterating the process until reaching an algorithm convergence condition.
2. And in the second stage, based on the dispatching of the single mobile server of the Gibbs sampling, the path planning and the task dispatching of the single server are optimized in a combined mode, so that the total completion time delay is minimized. The embodiment of the algorithm is shown in fig. 4, and specifically includes:
2.1, based on the output result of the first stage of the algorithm, converting the multi-server planning problem into a single-server planning problem.
2.2, executing Gibbs sampling to each edge server in parallel, planning a path based on 2-OPT, eliminating the path violating the storage constraint and the service quality constraint, and enabling the generated path to be suitable for a 'movement-calculation parallelism' movement mechanism. The specific contents are as follows:
firstly, randomly initializing a moving path and task scheduling of the server; randomly selecting two nodes in a path and turning over all path points in the two nodes; if the newly generated path does not accord with the storage limit, two nodes are reselected; otherwise, the total delay of the path is calculated. If the time delay of the new path is better than that of the original path, updating the path and updating the iteration times; otherwise, two nodes are reselected. And if the new path after iteration does not reach the 2-OPT iteration requirement, reselecting the two nodes. Otherwise, calculating the acceptance probability, and deciding whether a new path is reserved or not according to the probability.
2.4 task scheduling based on the current path. And calculating the acceptance probability rho of the current task scheduling, and deciding whether the task processing sequence is updated.
2.4.1, wherein t * And (3) calculating the time delay of each server after changing path planning or task scheduling, wherein t is the time delay of each server in an initial state, and omega is a smooth parameter influencing algorithm convergence.
2.5, the basic idea of the priority scheduling algorithm is as follows: for a task set to be processed by a single server, each task is given a priority, then when the tasks are scheduled, the task with the highest priority is selected from all the tasks in the ready state to run, and different time delays can be obtained by changing the priorities of the tasks. Specific examples are shown in fig. 5, and the execution steps are as follows:
2.5.1, the task sets prepared to be processed by the single server are prioritized according to a default sequence, and the completion time of each task in the server and the time of the server reaching the internet of things equipment corresponding to the task can be calculated based on path planning and task scheduling.
2.5.2, sequentially inspecting the completion time and the arrival time of each task, if the completion time is greater than the arrival time, adjusting the priority of the task one bit ahead, and continuing to circulate the step until the completion time is not greater than the arrival time; otherwise, the task priority is not changed.
2.5.3, if the condition that the completion time is not more than the arrival time can not be achieved after the task is transferred to the highest priority level through the transferring step, the priority level of the task is not changed, and the next task is continuously operated until all tasks are operated. The operation reduces the time of the server waiting for the completion of the task in the equipment of the Internet of things, further reduces the total time delay and optimizes the task scheduling scheme.
3. The process is updated iteratively until the algorithm based on Gibbs sampling reaches the convergence condition.
Example 4
This embodiment is a specific implementation of determining the minimum deployment number of servers and the deployment of static servers for a given edge network by using the method in embodiment 3. The method comprises the following specific steps:
1. in order to obtain a relatively accurate minimum deployment number of the servers, based on the second stage in the two-stage optimization algorithm, a method of predicting the maximum deployment number of a single server is adopted to calculate the minimum deployment number, and the specific method is as follows:
1.1, a single server is given a task set with a certain number of tasks, and the critical maximum bearable task number is calculated by adopting the dichotomy idea.
1.1.1, if the server can complete this number of tasks, the number of tasks is increased to 2 times.
1.1.2, if the server can not complete the number of tasks, halving the number of tasks, if the number of tasks can be completed, selecting the average value of the number of the tasks and the number of the previous tasks to be rounded upwards, and taking the value as the number of the tasks in the new task level of the server; otherwise, the number of tasks is halved until the maximum value of the tasks which can be completed by the server is reached.
1.2, dividing the number of tasks required to be completed by the maximum value of the tasks which can be completed by the server and rounding up, and obtaining the approximate minimum deployment number of the server.
1.2.1, at the moment, the task set and the server set are brought into a second algorithm of the two-stage optimization algorithm for calculation, and if the calculation can be finished, the deployment number of the servers is reduced; otherwise, increasing the deployment quantity until obtaining the accurate minimum deployment quantity of the servers.
2. Based on the foregoing description of the delay sensitive task, the static edge server can guarantee its quality of service. However, by the calculation of embodiment 3, it is more efficient to automatically determine the deployment locations of the static and dynamic servers according to the configuration of the edge computing network, specifically as follows:
2.1, through the calculation of the embodiment 3, each server is allocated with a path plan, and servers of which the paths only move in the communication range are separately distinguished according to the communication range of each server, and the servers are default to be deployed as static servers to provide more stable and efficient calculation services.
The above-mentioned embodiments are intended to illustrate the objects, technical solutions and advantages of the present invention in further detail, and it should be understood that the above-mentioned embodiments are merely exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made within the spirit and principle of the present invention should be included in the scope of the present invention.
Claims (8)
1. A multi-server mixed deployment system under a 5G/6G edge computing scene comprises a plurality of dynamic servers and a plurality of static servers, and is characterized in that a mixed deployment architecture composed of the static servers and the dynamic servers is used for servicing delay sensitive or computation intensive tasks in edge computing, so that service coverage of nodes of the Internet of things under the edge scene is completed;
the dynamic server allows two moving mechanisms of 'moving-computing parallel' and 'moving-computing serial' at the same time;
the mobile-computing parallel is to decouple the computing unloading process into request uploading, task computing and result returning, so that the dynamic server can process tasks in the moving process;
the mobile-computing serial is that the server stays in place according to the existing mobile mechanism and moves to the next node after the task computing is finished;
the dynamic server makes a decision on the selection of the two mobile mechanisms after comprehensively considering the state of the terminal node in the network, the computing resources of the server and the environment of the edge network.
2. The multi-server hybrid deployment system under 5G/6G edge computing scenario according to claim 1, wherein the number of static servers and dynamic servers is the minimum deployment number of servers of a given edge network device determined based on a two-phase optimization algorithm; the location of the static server is a deployment location of the multiple static servers determined based on a two-phase optimization algorithm.
3. The system for multi-server hybrid deployment in a 5G/6G edge computing scenario as claimed in claim 1, further comprising a path planning module and a task scheduling module for jointly optimizing the multi-dynamic server based on a two-stage optimization algorithm.
4. The system of claim 1, wherein the dynamic server is preferentially used for processing compute intensive and delay tolerant tasks, and the static server is preferentially used for processing delay sensitive tasks.
5. A multi-server hybrid deployment method under a 5G/6G edge computing scene is characterized by comprising the following steps:
s1, a mixed deployment architecture consisting of a dynamic server and a static server is used, wherein the dynamic server simultaneously allows two moving mechanisms of 'moving-computing parallel' and 'moving-computing serial'; the 'mobile-computing parallel' is to decouple a computing unloading process into request uploading, task computing and result returning, so that a dynamic server can process tasks in a mobile process; the mobile-computing serial is that the server stays in place according to the existing mobile mechanism and moves to the next node after the task computing is finished;
s2, providing an optimization problem model based on the hybrid deployment architecture;
s3, optimizing the deployment schemes of the dynamic server and the static server, wherein the dynamic server decides to select the two mobile mechanisms after comprehensively considering the terminal node state, the server computing resources and the edge network environment in the network;
and S4, completing full coverage of the nodes of the Internet of things in the 5G/6G edge network according to the optimization result.
6. The multi-server hybrid deployment method under 5G/6G edge computing scenario according to claim 5, wherein the step S3 includes a two-stage optimization algorithm for jointly optimizing path planning and task scheduling of the multi-dynamic server, the algorithm includes the following steps:
the method comprises the following steps that in the first stage, equipment is distributed based on an elite genetic algorithm, namely, a plurality of Internet of things equipment are distributed to a plurality of dynamic servers, so that a multi-server planning problem is converted into a single-server planning problem;
and in the second stage, based on the single dynamic server scheduling of Gibbs sampling, path planning and task scheduling of a single server are optimized in a combined mode, so that the total completion time delay is minimized.
7. The method for multi-server hybrid deployment in a 5G/6G edge computing scenario as claimed in claim 6, wherein the second stage further comprises: iterative updating based on Gibbs sampling solves two coupled sub-problems; path planning is carried out based on a 2-OPT algorithm, so that the path planning method is suitable for a 'movement-calculation parallelism' movement mechanism; based on the moving mechanism, a heuristic task scheduling algorithm is further provided.
8. The multi-server hybrid deployment method in a 5G/6G edge computing scenario according to claim 6,
the optimization object of the two-stage optimization algorithm is to determine the minimum deployment number and the static server deployment position of the server by the given edge network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111073797.4A CN113810916B (en) | 2021-09-14 | 2021-09-14 | Multi-server mixed deployment architecture and method in 5G/6G edge computing scene |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111073797.4A CN113810916B (en) | 2021-09-14 | 2021-09-14 | Multi-server mixed deployment architecture and method in 5G/6G edge computing scene |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113810916A CN113810916A (en) | 2021-12-17 |
| CN113810916B true CN113810916B (en) | 2022-11-18 |
Family
ID=78895220
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111073797.4A Active CN113810916B (en) | 2021-09-14 | 2021-09-14 | Multi-server mixed deployment architecture and method in 5G/6G edge computing scene |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113810916B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115225671B (en) * | 2022-07-15 | 2024-03-15 | 湖南大学 | AIoT intelligent medical multiple access edge unloading method, system, equipment and medium |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111343651A (en) * | 2020-02-18 | 2020-06-26 | 电子科技大学 | Service chain deployment method and system for serving crowd-sourcing computing environment |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103218265B (en) * | 2013-05-06 | 2016-04-27 | 杭州电子科技大学 | Based on the parallel type virtual role division methods of space density clustering |
| US10440058B2 (en) * | 2016-06-09 | 2019-10-08 | LGS Innovations LLC | Methods and systems for controlling traffic to VPN servers |
| CN107707583B (en) * | 2016-08-08 | 2020-11-17 | 环旭电子股份有限公司 | Cloud data transmission system and dynamic distribution method thereof |
| CN106371931B (en) * | 2016-09-30 | 2019-11-05 | 电子科技大学 | A kind of high-performance geoscience computing service system based on Web frame |
| CN112925652B (en) * | 2021-04-06 | 2023-11-21 | 北京百度网讯科技有限公司 | Application resource deployment method, device, electronic equipment and medium |
-
2021
- 2021-09-14 CN CN202111073797.4A patent/CN113810916B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111343651A (en) * | 2020-02-18 | 2020-06-26 | 电子科技大学 | Service chain deployment method and system for serving crowd-sourcing computing environment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113810916A (en) | 2021-12-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109413724B (en) | MEC-based task unloading and resource allocation scheme | |
| CN109818865B (en) | SDN enhanced path boxing device and method | |
| CN110602633B (en) | Explosive flow-oriented mobile edge computing unmanned aerial vehicle cluster auxiliary communication method | |
| Cong et al. | EdgeGO: A mobile resource-sharing framework for 6G edge computing in massive IoT systems | |
| CN114338504B (en) | Micro-service deployment and routing method based on network edge system | |
| CN110134493B (en) | Service function chain deployment algorithm based on resource fragment avoidance | |
| CN113950081B (en) | Dynamic service migration and request routing method for micro service | |
| CN108737569B (en) | Service selection method facing mobile edge computing environment | |
| CN110519776B (en) | Balanced clustering and joint resource allocation method in fog computing system | |
| CN111601327B (en) | Service quality optimization method and device, readable medium and electronic equipment | |
| CN111447005B (en) | Link planning method and device for software defined satellite network | |
| CN112784362A (en) | Hybrid optimization method and system for unmanned aerial vehicle-assisted edge calculation | |
| CN113709876A (en) | Satellite service resource allocation method and electronic equipment | |
| CN111885493B (en) | Micro-cloud deployment method based on improved cuckoo search algorithm | |
| Chen et al. | Learning-based computation offloading for IoRT through Ka/Q-band satellite–terrestrial integrated networks | |
| CN112073237B (en) | Large-scale target network construction method in cloud edge architecture | |
| CN113810916B (en) | Multi-server mixed deployment architecture and method in 5G/6G edge computing scene | |
| Wang et al. | Radio resource allocation for bidirectional offloading in space-air-ground integrated vehicular network | |
| CN114595049A (en) | Cloud-edge cooperative task scheduling method and device | |
| Dai et al. | A learning algorithm for real-time service in vehicular networks with mobile-edge computing | |
| CN115866788A (en) | 3C resource scheduling method of heaven and earth fusion network for active migration of MEC tasks | |
| CN113076177A (en) | Dynamic migration method of virtual machine in edge computing environment | |
| Parvaresh et al. | A continuous actor–critic deep Q-learning-enabled deployment of UAV base stations: Toward 6G small cells in the skies of smart cities | |
| Birabwa et al. | Service-aware user association and resource allocation in integrated terrestrial and non-terrestrial networks: A genetic algorithm approach | |
| Shinde et al. | Network selection and computation offloading in non-terrestrial network edge computing environments for vehicular applications |
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 |