CN113346938A - Edge computing resource fusion management method for air-space-ground integrated network - Google Patents
Edge computing resource fusion management method for air-space-ground integrated network Download PDFInfo
- Publication number
- CN113346938A CN113346938A CN202110549911.XA CN202110549911A CN113346938A CN 113346938 A CN113346938 A CN 113346938A CN 202110549911 A CN202110549911 A CN 202110549911A CN 113346938 A CN113346938 A CN 113346938A
- Authority
- CN
- China
- Prior art keywords
- computing
- edge
- user equipment
- service
- edge node
- 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
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/02—Arrangements for optimising operational condition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
- G06F9/45533—Hypervisors; Virtual machine monitors
- G06F9/45558—Hypervisor-specific management and integration aspects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
- G06N5/04—Inference or reasoning models
- G06N5/042—Backward inferencing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/14—Relay systems
- H04B7/15—Active relay systems
- H04B7/185—Space-based or airborne stations; Stations for satellite systems
- H04B7/18502—Airborne stations
- H04B7/18506—Communications with or from aircraft, i.e. aeronautical mobile service
- H04B7/18508—Communications with or from aircraft, i.e. aeronautical mobile service with satellite system used as relay, i.e. aeronautical mobile satellite service
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/14—Relay systems
- H04B7/15—Active relay systems
- H04B7/185—Space-based or airborne stations; Stations for satellite systems
- H04B7/1851—Systems using a satellite or space-based relay
- H04B7/18519—Operations control, administration or maintenance
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W84/00—Network topologies
- H04W84/02—Hierarchically pre-organised networks, e.g. paging networks, cellular networks, WLAN [Wireless Local Area Network] or WLL [Wireless Local Loop]
- H04W84/04—Large scale networks; Deep hierarchical networks
- H04W84/06—Airborne or Satellite Networks
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Software Systems (AREA)
- Astronomy & Astrophysics (AREA)
- General Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Artificial Intelligence (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
An edge computing resource fusion management method for an air-space-ground integrated network belongs to the technical field of network communication, analyzes service modes of space-ground integrated network space-base resources, provides two computing service fusion modes of combination and aggregation, models management of the space-base edge computing resources into a two-stage Steckelberg game model, and provides a hybrid iterative algorithm for solving Nash equilibrium. In the computing resource aggregation, the computing resources in any one terminal can independently exist, and general computing capability is provided for the user equipment.
Description
Technical Field
The invention belongs to the technical field of network communication, and particularly relates to an edge computing resource fusion management method for an air-space-ground integrated network.
Background
In recent years, information and communication technologies have been developed at a high speed, and a mobile network has made a major breakthrough. In the situation that the commercial use of the fifth generation mobile communication technology (5G for short) is about to expand gradually over the whole year, the International telecommunications Union (ITU for short) started the research of 6G in 2 months of 2020. With respect to 6G networks, many researchers at home and abroad have presented individual vision. In these vision, a world in which a ground wireless network, a satellite network and a near-earth unmanned aerial vehicle network are fully integrated to construct a wireless signal full-coverage and wireless network full-connection is mentioned. As a key infrastructure in the 6G network, the satellite and the unmanned aerial vehicle can provide network service for users in the environment inaccessible to the ground network; and under the condition that the ground network capacity is insufficient, the auxiliary ground network can provide expansion of the network capacity. The distance between the satellite and the unmanned aerial vehicle is relatively close to that between the user and the unmanned aerial vehicle, and along with the enhancement of the hardware capability of the satellite and the unmanned aerial vehicle, the perfection and the development of the transmission technology are realized, and the application of the satellite and unmanned aerial vehicle is more and more extensive. The satellite is used as an edge node to provide communication, cache and calculation services; applying the unmanned aerial vehicle to an edge calculation or edge cache scene; are the hot spots of current research. The integration of two space-based infrastructures, namely a satellite and an unmanned aerial vehicle, also becomes an important development direction of the current space-based network application. Nevertheless, the limited hardware capability of the satellite and the unmanned aerial vehicle, and the unstable operating environment, still exist for a long time as the main bottleneck limiting the application development.
Disclosure of Invention
1 air-space-ground integrated computing resource management architecture
1.1 air-space-ground integrated edge computing architecture
The air-ground integrated information network usually uses a high-orbit satellite as a space-based backbone network and uses a low-orbit satellite as a space-based access network or a space-based user. The general framework of the air-space-ground integrated information network is shown in fig. 1:
the 6G network provides ubiquitous coverage for various mobile terminals, and target users of the network are various mobile devices such as Internet of things nodes, Internet of vehicles nodes and personal communication terminals. In a 6G network scenario, a satellite can serve as a provider of a network to provide network services for a terminal. With the richness of terminal types, the perfection of functions and the diversified development of applications, the satellite serves as a cloud server or an edge server to provide various types of services such as calculation, communication, storage and the like for the terminal. The high-orbit satellite is deployed with computing equipment such as a large server and the like, cloud service can be provided for the mobile terminal, and the low-orbit satellite is used as an edge node to lease the computing capacity of the low-orbit satellite to the terminal. The space-ground-integrated edge computing architecture is shown in fig. 2.
Due to the limited hardware resources of the satellite, the transmission distance is long, and the wireless network is susceptible to weather. Therefore, the space-based infrastructure as an edge computing node necessarily provides network services which are difficult to provide by the ground-based infrastructure for the ground terminal. The scenarios of these services can be broadly divided into two types.
1) The space-based infrastructure serves as the assistance of the ground network, and provides general computing service for the inaccessible area of the ground network or the scene of insufficient ground network capacity.
2) The space-based infrastructure acts as a direct provider of services, providing specialized computing services around its unique remote or aerial images.
In both of the above exemplary scenarios, the ground terminal requires support of space-based edge computing resources. Because of the limited computing power of a single satellite, when computing resources of multiple edge servers are needed, a convergence mode of aggregating and combining the two computing resources is provided.
1.2 air-space-ground integration edge computing resource aggregation
The aggregation method of air-space-ground integrated edge computing resources is mainly oriented to general computing tasks and is suitable for the first service providing scenario mentioned in the above section. And (3) referring to object-oriented intra-object aggregation, and proposing an aggregation concept of space-ground integrated computing resources. In the computing resource aggregation, the computing resources in any one terminal can independently exist, and general computing capability is provided for the user equipment. The architecture is shown in fig. 3.
The aggregation of space-ground-integrated edge computing resources requires the following support:
1) hardware level: certain computing hardware is required to be carried in each infrastructure, and certain computing capacity is provided;
2) platform level: providing a general operating system platform and supporting the establishment of a virtual machine and the allocation of computing resources;
3) software layer: the method provides general virtual machine leasing service and provides functions of cross-node virtual machine distribution, management and the like.
The aggregation method of edge computing resources is essentially the scheduling and management of virtual machines. The user can rent the virtual machine to different edge servers according to the requirement of the computing resource of the user. Each virtual machine provides a general computing task, and coordination and function combination among the virtual machines are not needed. The user only needs to ensure that the total computing capacity of the virtual machine meets the application requirement.
The management of air-space-ground integrated edge computing resources corresponds to a ground edge computing scene, and the aggregation method is equivalent to the renting of a computing Platform provided at a Platform as a Service (PaaS) level, and is similar to the current public cloud renting scene.
1.3 air-space-ground integrated edge computing resource combination
The combination method of the air-space-ground integrated edge computing resources is mainly oriented to the special computing tasks of the remote sensing images and the aerial images, and is suitable for the second service providing scene mentioned in the above section. And (4) referring to the combination in the object-oriented way, and proposing a combination concept of air-space-ground integrated edge computing resources. In the computing resource combination, the computing resources in any one terminal do not exist independently, but cooperate with each other to jointly complete a complete task. The architecture is shown in fig. 4.
High-resolution remote sensing images taken by low earth orbit satellites are data specific to space-based infrastructure. Users want to use these data, which inevitably consumes computing and transmission resources of the space-based infrastructure. If the original data are directly sent to the terminal, on one hand, a large amount of transmission resources are consumed, and on the other hand, more computing resources are still needed for the terminal to process the data. Therefore, the space-based infrastructure can issue the result with smaller data size after processing to the user according to the requirement of the user.
Image processing typically requires a higher consumption of computing resources due to the limited computing power of a single infrastructure. If the space-based infrastructure needs to be cooperated, one task needs to be decomposed into a plurality of subtasks in advance, and the tasks are completed by cooperation of all edge nodes. The remote sensing image and the aerial image are irreplaceable, the computing power provided by the space-based infrastructure is irreplaceable, and the satellite needs to perform respective operations according to preset tasks.
The management of air-space-ground integrated edge computing resources corresponds to a ground edge computing scene, and the combination method is equivalent to the application of providing remote sensing images or aerial images for a user terminal at a Software as a Service (SaaS) level. When the application relates to a plurality of edge nodes, the cooperation mode and the cooperation process are preset by software, and a user only needs to rent the cloud software directly.
2-edge computing resource management model
2.1 edge computing resource management scenarios
In the air-ground integrated application scenario, the space-based infrastructure is managed by a space-based information network operator. The operator provides infrastructure rental services and allows for profit maximization to be achieved within the user's acceptable expense. And the user needs to rent space-based infrastructure and reaches the maximum benefit on the premise of meeting the calculation requirement.
The computing capability and transmission link of the space-based infrastructure are easily affected by the external environment, so that the service quality of the space-based infrastructure is unstable. In order to more effectively acquire the service quality of the space-based infrastructure, the service satisfaction of the user equipment is interacted, the service quality of the space-based infrastructure can be effectively acquired, and therefore, the recommendation evaluation is provided for the resource leasing decision. The scenario is shown in fig. 5.
The essence of edge computing resource management is the process of gaming between service providers and users of the services, both seeking to maximize benefits through certain strategies. The section models communication and calculation expenses involved in the edge calculation process, so that a general game model is obtained, and a method for solving the maximum benefit is provided.
2.2 computational overhead modeling
This section first models the overhead of computing resources. Consider a scenario where multiple edge nodes and multiple user terminals are randomly distributed in space. UE {1,2, …, M } represents a set of user equipments, the total number of user equipments being M. EN ═ {1,2, …, N } is used to denote a set of edge nodes, with the total number of edge nodes being N. Assuming that there is a dedicated computing service across nodes in these edge nodes, the set of edge nodes that provide the dedicated computing service is denoted as EG ═ EG1,eg2,…,egLThere are L dedicated services. For the l-th service, a set of nodes that collectively provide a dedicated service is denoted
It is assumed that the edge nodes are able to communicate with each other over a wireless link. For general purpose computing services, tasks in each user device may be offloaded to an edge node, and the device can only communicate with one edge node at a time. For specialized computing services that require cooperation of multiple nodes, a proxy communicates directly with the user device. The proxy node coordinates and controls the computation and transmission processes within the node set. According to two service providing forms of the air-space-ground integrated edge computing resource, two types of edge computing tasks are provided for user equipment m: general purpose computing tasks and specialized computing tasks. Two types of tasks are defined as follows:
1) general computational tasks, written as
Wherein QmRepresenting the amount of tasks transmitted to the edge node,
representative task JmThe maximum delay tolerance is calculated. According to different requirement combinations, tasks can be divided into delay sensitive tasks or energy consumption sensitive tasks, and the tasks can be completed by leasing computing resources of any edge node.
2) And the special computing task can be regarded as a computing task facing the special service node set. The mathematical expression is as follows:
wherein SQmRepresenting the amount of tasks to rent a dedicated service,
delegate task SJmThe maximum delay tolerance is calculated.
The section models the calculation of the general calculation task, and the special calculation task can be converted into a general calculation task model through the merging and mapping of the nodes. The calculation model in this section mainly takes into account two influencing factors: time delay and power consumption.
1) And (4) time delay. Based on the above task model, task JmTransport delay from user equipment m offload to edge node n
Expressed as a formula
Wherein r ismnRepresenting the data transmission rate. The time delay formula of the computing task in the edge node n is
CnMeter for representing edge nodeComputing power. Maximum queuing delay for tasks in edge nodes
Wherein F { xm0 waiting for a coefficient function representing the user equipment m to task JmThe waiting factor required to offload to the edge node. The total time delay formula of one unloading task is
2) And (4) energy consumption. Task JmThe formula of the calculated energy consumption at the edge node n is
Wherein etanThe energy consumption coefficient per CPU cycle for edge node n is represented. Transmission energy consumption formula unloaded from user equipment m to edge node n
Where ρ ismnRepresenting the transmission power between the edge node n and the user equipment m.
2.3 resource Allocation model
In an air-space-ground integrated edge computing scenario, different types of space-based infrastructures may belong to different operators. These infrastructures compete with each other, providing flexible access, and meeting the computing needs of the user equipment. The user equipment can arbitrarily select different space-based infrastructures according to the requirements of the user equipment to obtain satisfactory computing services. There is also a competitive relationship between user devices, which tend to achieve higher edge computing quality of service at lower prices. To attract more customers, the infrastructure needs to set an appropriate price per day to maximize their profit. The interaction process between the space based infrastructure and the user device is a typical stankberg gaming process. The space-based infrastructure serves as a leader, and the user equipment serves as a follower to participate in the game together.
In an actual scenario, around the two computing resource fusion modes proposed by the present invention, the edge computing service lease of the present invention can be divided into two scenarios:
1) a single service rental scenario. The scene is based on the provided air-space-ground integrated edge computing resource aggregation method, and all space-based edge nodes provide lease of a general computing task. All nodes have the same logical structure and provide the same computing service.
2) Mix service rental scenarios. The scene is oriented to an aerospace-ground integrated edge computing resource fusion method. At this point, the space-based edge node may provide for the lease of both general-purpose computing tasks and specialized computing tasks. The service agent provides lease service uniformly for the node set providing special calculation task, and the calculation and interaction process of multiple nodes is preset in advance by the service itself. After the user rents the service, the tasks of all the nodes are fixed and cannot be freely changed by the user. At this time, the node set providing the dedicated service may be regarded as one node, and the whole scenario may divide all the edge node sets into two groups, one group provides the general computing service, the other group provides the dedicated computing service, and decisions between the two groups do not affect each other.
In both scenarios, the interaction between the space based infrastructure and the user equipment can be modeled as a two-phase, sippansberg gaming framework. In the first phase, each service provider (i.e., space based infrastructure) sets an optimal price for computing resources. In the second phase, the user (i.e., user equipment) determines the optimal amount of computing resources needed based on the price of the edge node.
The Stancoberg game model consists of four parts, namely a service provider, a consumer, a utility function and an action strategy, and is specifically represented as follows:
wherein EN, UE represent provider and consumer of service respectively, G, F represent utility function of provider and consumer respectively,
representing the pricing policy of the service provider and the resource lease policy of the service consumer, respectively. All in one in the skyIn a chemolitho-edge computing scenario, the space-based edge node EN is the provider of the service, which announces the price at which computing resources are sold to the user equipment
The user equipments UE, called consumers, pay for edge computing resources and receive computing services Q from the edge nodes. Assume that the set of computing requirements of the user equipment is Q ═ Q1,Q2,…,QMIn which Qj,QmThe computing resource pricing set of the edge node is the computing requirement of the user equipment j, m
Wherein P isnPricing the computing resources of the edge node n.
The utility functions of the service provider and the consumer in this section model are as follows:
1) utility function of the user equipment. This section considers the interaction information between user terminals. For the user equipment, the evaluation of the edge nodes is modeled by the satisfaction of a logarithmic function. In addition to considering the satisfaction of the user equipment with each edge node, the user equipment also refers to the satisfaction of other equipment with each edge node. In this case, each edge node may be properly evaluated by different user equipment. The problem model includes self-satisfaction, other equipment evaluation, and costs computed at the edge nodes. For user equipment m, its utility function Fm(Pn,Qm,Q-m) Modeled as equation (1).
Wherein Q-mRepresenting the amount of tasks, U, transmitted to the edge node by users other than user equipment mj,UmAnd (3) representing the satisfaction degree of the user j, m, and calculating by adopting a logarithmic function, wherein the calculation is shown in formula (2).
Um＝aln(1+Qm) (2)
and a represents the satisfaction coefficient of the user equipment, and the value is set to be 10.
In the formula (1), the first and second groups,
representing service satisfaction information obtained from other user devices. Since the space-based infrastructure computing and transmission environment is unstable and is easily influenced by the outside, the user equipment is more likely to consider the service conditions of other equipment, so that the resource leasing strategy of the user equipment is formulated. Matrix array
Representing user equipment m1Slave user equipment m2The obtained influence coefficient is obtained by the statistical calculation of the previous interaction information, and the value range is [0,1 ]]In the meantime.
A higher value of (c) indicates a closer proximity of the two user equipments to the environment. This section considers that the influence of user equipment m and user equipment j on each other is the same, i.e.
DmRepresents the payment overhead of the user equipment m, and is formulated as Dm＝Pn×Qm。
It is the transmission energy consumption of the user equipment m, ζ is the coefficient factor of energy consumption, and is used to determine the appropriate weight of transmission energy consumption, and the value is 1.
2) Utility function G of edge noden. For edge node n, its utility function is equal to profit minus its own cost, and the formula is expressed as equation (3).
Wherein
Is the calculated energy consumption of the edge node n, and xi is the coefficient factor of the energy consumption, and is used for determining the proper weight of the calculated energy consumption, and the value is 1.
3-edge computing resource management policy
3.1 Nash equilibrium analysis based on Stencolberg Game
In the space-based information network edge computing scenario, when the user equipment only needs general computing services, each edge node independently announces its price without cooperation with other edge nodes. The pricing of the nodes is decided in a distributed manner. The pricing of the edge node n depends not only on the feedback of the user equipment but also on the actions of other edge nodes. The problem model formula can be modeled as formula (4).
Wherein
Representing the optimal price, G, of the entire edge node except for the edge node nnIs the utility function of the edge node.
A set of computing requirements is represented.
Optimal price based on the above edge nodes
For user equipment m, utility formula (5) is established.
Based on the above model formula, all edge nodes and user equipments want to be able to make maximum use of their utility function values. Next, this section will analyze the stanoeberg game with complete information between the edge node and the user device.
Definitions 1. assumptions
Is the optimal unit price of the edge node n,
is the optimal amount of tasks for the user equipment m to transmit to the edge node,
for optimal transmission requirements of user equipments other than user equipment m, then
The optimum nash equilibrium point satisfies the following condition:
theorem 1. consider the dynamic computational requirements of having a fixed number of user devices. For a user equipment whose utility function satisfies equation (2), then there is one and only nash equalization point for that user equipment.
For any user equipment m, its utility function is Fm(Pn,Qm,Q-m) The specific function is shown in formula (7).
And (5) solving the first derivative of the formula (7) to obtain a formula (8).
And solving the second derivative of the formula (7) to obtain a formula (9).
Due to a>0 and (1+ Q)m)2>0, and thus the second derivative of the utility function is negative. Therefore, the utility function of the user equipment is a strict convex function, and the existence of Nash equilibrium is proved. Theorem 1 proves that the process is finished.
The first derivative of the utility function of the user equipment, namely the numerical value of the formula (8), is made equal to zero, and the optimal demand of the computing resources can be obtained by solving.
According to equation (3), the utility function of the edge node n is shown in equation (10).
The first derivative of equation (10) is calculated to obtain equation (11).
The second derivative of equation (10) is calculated to obtain equation (12).
Since all parameters in equation (12) are greater than 0, the second derivative of the utility function is less than 0. In other words, the utility function of the edge node is a strict convex function, and nash equilibrium points exist. Theorem 2 proves that the process is finished.
Based on theorem 1 and theorem 2, nash equilibrium points exist in the pricing stage of the edge node and the computing resource leasing strategy stage of the user terminal, respectively. It can be concluded that the stegano-kelbergen-schwann equilibrium point exists and is unique.
3.2 resource optimization method under single service lease scene
For a single service rental scenario, the process is a typical Stancoberg game. In the air-space-ground integrated information network scenario, a central agent is assumed to exist in the computing resource leasing application. Each participant in the game model transmits own information to the central agent, and meanwhile, each participant can also obtain the information of other participants from the central agent, so that each participant can obtain the optimal strategy.
According to the proving process of theorem 1 in the above section, there is an optimal solution for the utility value of the user equipment. Therefore, when the provider of the service, i.e. the edge node, has given the pricing scheme, the user of the service, i.e. the user equipment, calculates the optimal calculation requirement of the user equipment according to equation (10).
The process of solving the optimal pricing scheme comprises the following steps: let the first derivative of the edge node utility function, i.e., the value of equation (12), equal to zero, yield the following equation.
User equipment r for task offloading due to selection of edge node nmn,ρmnThe numerical values are not necessarily the same, and the equation is difficult to directly solve, so that the pricing can gradually approach the optimal value in a dynamic iteration mode. To this end, this section proposes a hybrid dynamic iterative algorithm. The dynamic iterative algorithm is easy to realize and has good convergence performance. Algorithm 1 gives the pseudo code of the packet dynamic iterative algorithm. In the algorithm, the calculation requirement is directly calculated by formula (10), and the next generation price Pn(t +1) calculating according to the derivative direction to gradually enable the Steiner game modelApproximating nash equilibrium. The computational complexity of the algorithm is O (t)maxMN) where t ismaxRespectively the maximum number of iterations of the edge node. Further, the coefficient α is a single search step size of the edge node.
Algorithm 1. solving Nash equilibrium by a hybrid dynamic iterative algorithm.
Inputting: the method includes the steps that an edge node set EN is {1,2, …, N }, and a user equipment set UE is {1,2, …, M };
and (3) outputting: optimum price
Optimal computational demand
Initialization:
edge node price Pn,(n∈EN)；
User equipment requirement Qmn,(m∈UE,n∈EN)；
From the first iteration t to the maximum number of iterations tmaxAnd adding 1 to t each time, and calculating:
obtaining the optimal demand Q of the computing resource according to the formula (10) from i to 1 to i to M (i is added with 1 each time)in(t)；
Calculating the optimal price from i-1 to i-N (i adds 1 each time)
Wherein alpha is the single search step length of the edge node, M is the total number of the user equipment, a is the user satisfaction degree, and rmiFor the data transmission rate from the edge node to the user equipment m, ζ is the transmission power consumption coefficient factor, CiIs the computing power of the edge node, xi is the computing energy consumption coefficient factor, rhomiRepresenting the transmission power, η, between the edge node and the user equipment miThe energy consumption coefficient per CPU cycle representing the edge node.
3.3 resource optimization method in hybrid service rental scene
In a mixed service rental scene, for a special computing service, although a plurality of nodes provide services at the same time, the nodes can be merged into a virtual node, and then a Stancoberg game model is established. The method comprises the following specific steps: the edge nodes are first divided into two groups. EN ═ GEN, SEN }. Wherein GEN, SEN represent the set of edge nodes in the general purpose computing service and the special purpose computing service, respectively. Merging each set of nodes providing dedicated services, each set of nodes being merged into a virtual edge node:
the set of edge nodes may be mapped as:
the stancoberg game model. Wherein N is1The number of edge nodes for providing general computing services, and the number of virtual nodes for providing special computing services. The hybrid service rental can still be modeled as a Stancoberg gaming model.
Similar to algorithm 1, this subsection presents a packet-dynamic iterative algorithm to solve nash equilibrium, and pseudo-code is presented in algorithm 2. The computational complexity of the algorithm is O (t)maxM(N1+L))。
And 2, solving the Nash equilibrium based on a grouping dynamic iterative algorithm.
Inputting: the method includes the steps that an edge node set EN is {1,2, …, N }, and a user equipment set UE is {1,2, …, M };
and (3) outputting: optimum price
Optimal computational demand
Initialization:
edge node grouping, EN ═ { GEN, GEN };
SEn of the virtual node maps are obtained,
edge node price Pn,(n∈EN′)；
User equipment requirement Qmn,(m∈UE,n∈EN′)；
From the first iteration t to the maximum number of iterations tmaxAnd adding 1 to t each time, and calculating:
obtaining the optimal demand Q of the computing resource according to the formula (10) from i to 1 to i to M (i is added with 1 each time)in(t)；
From i-1 to i-N1(i add 1 each time) calculation
From i-1 to i-L (i adds 1 each time) calculation
Wherein alpha is the single search step length of the edge node, M is the total number of the user equipment, a is the user satisfaction degree, and rmiFor the data transmission rate from the edge node to the user equipment m, ζ is the transmission power consumption coefficient factor, CiIs the computing power of the edge node, xi is the computing energy consumption coefficient factor, rhomiRepresenting the transmission power, η, between the edge node and the user equipment miThe energy consumption coefficient per CPU cycle representing the edge node.
Drawings
FIG. 1 general architecture diagram of air-space-ground integrated information network
FIG. 2 air-space-ground integrated edge computing architecture
FIG. 3 is a schematic diagram of edge computing resource aggregation for air, space and ground integration
FIG. 4 is a schematic diagram of an edge computing resource combination for air-space-ground integration
FIG. 5 air-space-ground integrated edge computing resource management scenario
Detailed Description
The application process of the edge computing resource fusion management method of the air-space-ground integrated network provided by the invention is as follows:
step 1: ground end user sending network request to high orbit satellite
step 3, executing an algorithm, wherein the high-orbit satellite sends a resource distribution result to a low-orbit satellite constellation for executing the task;
and Step 4, the low-orbit satellite executes the tasks according to the resource allocation amount and the specific task requirements and returns the results to the ground user.
Claims (4)
1. An air-space-ground integrated information network is generally characterized in that a high-orbit satellite is used as a space-based backbone network, and a low-orbit satellite is used as a space-based access network or a space-based user;
deploying computing equipment in the high-orbit satellite can provide cloud service for the mobile terminal, and leasing the computing capacity of the low-orbit satellite serving as an edge node to the terminal;
the space-based infrastructure is used as an edge computing node and is bound to provide network services which are difficult to provide by the space-based infrastructure for the ground terminal;
the method is characterized in that: these services are divided into two types;
1) the space-based infrastructure serves as the assistance of a ground network, and provides general computing service for the inaccessible area of the ground network or the scene with insufficient ground network capacity;
2) the space-based infrastructure serves as a direct provider of services, and provides special computing services around unique remote sensing or aerial images of the space-based infrastructure;
in the two typical scenes, the ground terminal needs to support space-based edge computing resources; because a single satellite has limited computing capacity, when computing resources of a plurality of edge servers are needed, two computing resource fusion modes of aggregation and combination are provided;
1.2 air-space-ground integration edge computing resource aggregation
The aggregation method of the air-space-ground integrated edge computing resources is mainly oriented to general computing tasks and is suitable for the first service providing scene mentioned in the above section; referring to aggregation in an object-oriented way, a concept of aggregation of space-sky-ground integrated computing resources is proposed; in the computing resource aggregation, the computing resources in any terminal can independently exist, and general computing capability is provided for the user equipment;
1.3 air-space-ground integrated edge computing resource combination
The combination method of the air-space-ground integrated edge computing resources is mainly oriented to the special computing tasks of the remote sensing images and the aerial images, and is suitable for the second service providing scene mentioned in the above section; referring to the combination in the object-oriented, the combination concept of the air-space-ground integrated edge computing resources is put forward; in the computing resource combination, the computing resources in any one terminal do not exist independently, but cooperate with each other to jointly complete a complete task;
the high-resolution remote sensing image shot by the low-orbit satellite is data specific to space-based infrastructure; users want to use the data, and the computing and transmission resources of the space-based infrastructure are consumed necessarily; the space-based infrastructure issues a result with smaller data volume after processing to the user according to the requirement of the user;
image processing typically requires a higher consumption of computing resources due to the limited computing power of a single infrastructure; if the space-based infrastructure needs to be cooperated, one task needs to be decomposed into a plurality of subtasks in advance, and each edge node is cooperated to complete the task; the satellite needs to execute respective operations according to a preset task;
the management of the air-space-ground integrated edge computing resources corresponds to the ground edge computing scene, and the combination method is equivalent to providing the application facing the remote sensing image or the aerial image for the user terminal on the software-service level; when the application relates to a plurality of edge nodes, the cooperation mode and the cooperation process are preset by software, and a user only needs to rent the cloud software directly;
2-edge computing resource management model
2.1 edge computing resource management scenarios
In the air-space-ground integrated application scene, the space-based infrastructure is managed by a space-based information network operator; the operator provides the rental business of the infrastructure and considers that the profit is maximized within the expense range acceptable by the user; the user needs to rent space-based infrastructure and reaches the maximum benefit on the premise of meeting the calculation requirement;
in order to more effectively acquire the service quality of the space-based infrastructure, the service satisfaction of the user equipment is interacted, and the service quality of the space-based infrastructure can be effectively acquired, so that the recommendation evaluation is provided for the resource lease decision;
the essence of edge computing resource management is a game process between a service provider and a service user, and both parties seek maximization of benefits through a certain strategy; modeling communication and calculation expenses involved in the edge calculation process so as to obtain a general game model and further provide a method for solving the benefit maximization;
2.2 computational overhead modeling
Firstly, modeling the expenditure of computing resources; considering scenes of a plurality of randomly distributed edge nodes and a plurality of user terminals in a space; UE {1,2, …, M } represents a set of user equipments, and the total number of user equipments is M; using EN ═ {1,2, …, N } to represent a set of edge nodes, where the total number of edge nodes is N; assuming that there is a dedicated computing service across nodes in these edge nodes, the set of edge nodes that provide the dedicated computing service is denoted as EG ═ EG1，eg2，...，egLH, there are L dedicated services in total; for the l-th service, a set of nodes that collectively provide a dedicated service is denoted
It is assumed that edge nodes can communicate with each other over wireless links; for general computing services, tasks in each user device can be offloaded to an edge node, and a device can only communicate with one edge node at the same time; for a dedicated computing service requiring cooperation of a plurality of nodes, communicating directly with the user equipment by an agent; the agent node coordinates and controls the calculation and transmission process in the node set; according to two service providing forms of the air-space-ground integrated edge computing resource, two types of edge computing tasks are provided for user equipment m: general computing tasks and special computing tasks; two types of tasks are defined as follows:
1) general computational tasks, written as
Wherein QmRepresenting the amount of tasks transmitted to the edge node,
representative task JmCalculating the maximum delay tolerance;
2) the special computing task can be regarded as a computing task facing the special service node set; the mathematical expression is as follows:
wherein SQmRepresenting the amount of tasks to rent a dedicated service,
delegate task SJmCalculating the maximum delay tolerance;
modeling the calculation of the general calculation task, and converting the special calculation task into a general calculation task model through the merging and mapping of the nodes; two influencing factors are considered: time delay and energy consumption;
1) time delay; based on the above task model, task JmTransport delay from user equipment m offload to edge node n
Expressed as a formula
Wherein r ismnRepresenting a data transmission rate; the time delay formula of the computing task in the edge node n is
CnRepresenting the computational power of the edge node; maximum queuing delay for tasks in edge nodes
Wherein F { xm0 waiting for a coefficient function representing the user equipment m to task JmWaiting coefficients needed for unloading to the edge nodes; the total time delay formula of one unloading task is
2) Energy consumption; task JmThe formula of the calculated energy consumption at the edge node n is
Wherein etanRepresenting the energy consumption coefficient of each CPU cycle of the edge node n; transmission energy consumption formula unloaded from user equipment m to edge node n
Where ρ ismnRepresenting the transmission power between the edge node n and the user equipment m;
2.3 resource Allocation model
Edge computing service tenancy is divided into two scenarios:
1) a single service rental scenario; the scene is based on the proposed air-space-ground integrated edge computing resource aggregation method, and all space-based edge nodes provide lease of a general computing task; all nodes have the same logic structure and provide the same computing service;
2) a hybrid service rental scenario; the scene is oriented to an aerospace-ground integrated edge computing resource fusion method; at the moment, the space-based edge node can provide the lease of the general computing task and the lease of the special computing task; the service agent provides lease service uniformly for the node set providing special calculation task, and the calculation and interaction process of multiple nodes is preset in advance by the service itself; after the user rents the service, the tasks of all the nodes are fixed and cannot be freely changed by the user; at this time, the node set providing the special service can be regarded as one node, and the whole scene can divide all edge node sets into two groups, one group provides general computing service, the other group provides the special computing service, and the decisions between the two groups are not influenced;
in the above two scenarios, the interaction between the space-based infrastructure and the user equipment can be modeled as a two-stage sippansberg game framework; in the first phase, each service provider, i.e. the space-based infrastructure, sets the optimal price for the computing resources; in the second stage, the user, namely the user equipment determines the required optimal computing resource amount according to the price of the edge node;
the Stancoberg game model consists of four parts, namely a service provider, a consumer, a utility function and an action strategy, and is specifically represented as follows:
wherein EN, UE represent provider and consumer of service respectively, G, F represent utility function of provider and consumer respectively,
respectively representing a pricing strategy of a service provider and a resource leasing strategy of a service consumer; in the air-space-ground integrated edge computing scenario, the space-based edge node EN is the provider of the service, which announces the price at which computing resources are sold to user equipment
User equipment, UE, called consumers, paying for edge computing resources and receiving computing services from edge nodes
Assume that the user device's set of computational requirements is
Wherein Qj，QmThe computing resource pricing set of the edge node is the computing requirement of the user equipment j, m
Wherein P isnPricing computing resources of the edge node n; the utility functions of the service provider and the consumer are as follows:
1) a utility function of the user equipment; interactive information between user terminals is considered; for the user equipment, the evaluation of the edge node is modeled by the satisfaction degree of a logarithmic function; in addition to considering the satisfaction of the user equipment to each edge node, the user equipment also refers to the satisfaction of other equipment to each edge node; in this case, each edge node may be properly evaluated by different user equipment; the problem model comprises self satisfaction, other equipment evaluation and cost calculated at the edge node; for user equipment m, its utility function Fm(Pn，Qm，Q-m) Modeling as formula (1);
wherein Q-mRepresenting the amount of tasks, U, transmitted to the edge node by users other than user equipment mj，UmThe satisfaction degree of the user j, m is represented, a logarithmic function is adopted for calculation, and the calculation is shown in a formula (2);
Um＝aln(1+Qm) (2)
a represents the satisfaction coefficient of the user equipment, and the value is set to be 10;
in the formula (1), the first and second groups,
representing service satisfaction information obtained from other user equipment; because the computing and transmission environment of the space-based infrastructure is unstable and is easily influenced by the outside, the user equipment is more likely to consider the service conditions of other equipment, so that the resource leasing strategy of the user equipment is formulated; matrix array
Representing user equipment m1Slave user equipment m2The obtained influence coefficient is obtained by the statistical calculation of the previous interaction information, and the value range is [0,1 ]]To (c) to (d);
a higher value of (d) indicates a closer proximity of the two user devices to the environment; it is considered that the influence of the user equipment m and the user equipment j on each other is the same, i.e.
DmRepresents the payment overhead of the user equipment m, and is formulated as Dm＝Pn×Qm；
Is the transmission energy consumption of the user equipment m, and ζ is the coefficient factor of energy consumption, and the value is 1;
2) utility function G of edge noden(ii) a For the edge node n, the utility function is equal to the profit minus the cost of the edge node n, and the formula is expressed as formula (3);
wherein
Is the calculated energy consumption of the edge node n, and xi is the coefficient factor of the energy consumption, and is used for determining the proper weight of the calculated energy consumption, and the value is 1.
2. The method of claim 1, further comprising an edge computing resource management policy:
in the space-based information network edge computing scene, when user equipment only needs general computing service, each edge node independently declares the price without cooperation with other edge nodes; the pricing of the nodes makes a decision in a distributed mode; pricing of the edge node n depends not only on feedback of user equipment, but also on actions of other edge nodes; modeling a problem model formula into a formula (4);
wherein
Representing the optimal price, G, of the entire edge node except for the edge node nnA utility function which is an edge node;
representing a set of computing requirements;
optimal price based on the above edge nodes
Establishing a utility formula (5) aiming at the user equipment m;
based on the model formula, all the edge nodes and the user equipment hope to utilize the utility function values of the edge nodes and the user equipment to the maximum extent; next, the Stencoerberg game with complete information between the edge node and the user equipment is analyzed;
definitions 1. assumptions
Is the optimal unit price of the edge node n,
is the optimal amount of tasks for the user equipment m to transmit to the edge node,
for optimal transmission requirements of user equipments other than user equipment m, then
The optimum nash equilibrium point satisfies the following condition:
theorem 1. consider the dynamic computational requirements of a fixed number of user devices; for a certain user equipment, the utility function of the certain user equipment meets the formula (2), and then the user equipment has one and only Nash equilibrium point;
for any user equipment m, its utility function is Fm(Pn，Qm，Q-m) The concrete function is shown in formula (7);
solving the first derivative of the formula (7) to obtain a formula (8);
solving the second derivative of the formula (7) to obtain a formula (9);
since a > 0 and (1+ Q)m)2> 0, so the second derivative of the utility function is negative; therefore, the utility function of the user equipment is a strict convex function, and the existence of Nash equilibrium is proved; theorem 1 proves that the test is finished;
making a first derivative of a utility function of the user equipment, namely a numerical value of a formula (8) equal to zero, and solving to obtain the optimal demand of the computing resources;
theorem 2. consider that a fixed number of user equipments have a dynamic price; for each edge node, the utility function of the edge node meets the formula (3), and then the edge node has a unique Nash equilibrium point;
according to formula (3), the utility function of the edge node n is shown in formula (10);
solving the first derivative of the formula (10) to obtain a formula (11);
solving a second derivative of the formula (10) to obtain a formula (12);
since all parameters in equation (12) are greater than 0, the second derivative of the utility function is less than 0; in other words, the utility function of the edge node is a strict convex function, and a nash equilibrium point exists; theorem 2 proves that the test is finished;
based on theorem 1 and theorem 2, Nash equilibrium points respectively exist in a pricing stage of the edge node and a computing resource leasing strategy stage of the user terminal; the conclusion is that the steiner-bergamosspoint exists and is unique.
3. The method of claim 2, further comprising a resource optimization method in a single service rental scenario:
for a single service rental scenario, the process is a typical Stancoberg game; in the air-space-ground integrated information network scene, a central agent is supposed to exist in the computing resource leasing application; each participant in the game model transmits own information to the central agent, and simultaneously each participant also acquires information of other participants from the central agent, so that each participant can calculate an optimal strategy;
according to the proving process of theorem 1 in the upper section, the utility value of the user equipment has an optimal solution; therefore, after the provider of the service, that is, the edge node, has given the pricing scheme, the user of the service, that is, the user equipment, calculates the optimal calculation requirement of the user equipment according to the formula (10);
the process of solving the optimal pricing scheme comprises the following steps: making the first derivative of the utility function of the edge node, namely the numerical value of the formula (12), equal to zero to obtain the following equation;
algorithm 1 gives pseudo code of the packet dynamic iterative algorithm; in the algorithm, the calculation requirement is directly calculated by formula (10), and the next generation price Pn(t +1) is then based on the derivative directionCalculating to enable the Steiner game model to gradually approach Nash equilibrium; the computational complexity of the algorithm is O (t)maxMN) where t ismaxRespectively the maximum iteration times of the edge nodes; in addition, the coefficient α is a single search step size of the edge node;
algorithm 1. solving Nash equilibrium by a hybrid dynamic iterative algorithm.
Inputting: the method includes the steps that an edge node set EN is {1,2, …, N }, and a user equipment set UE is {1,2, …, M };
and (3) outputting: optimum price
Optimal computational demand
Initialization:
edge node price Pn，(n∈EN)；
User equipment requirement Qmn，(m∈UE，n∈EN)；
From the first iteration t to the maximum number of iterations tmaxAnd adding 1 to t each time, and calculating:
obtaining the optimal demand Q of the computing resource according to the formula (10) from i to 1 to i to M (i is added with 1 each time)in(t)；
Calculating the optimal price from i-1 to i-N (i adds 1 each time)
Wherein alpha is the single search step length of the edge node, M is the total number of the user equipment, alpha is the user satisfaction degree, and rmiFor the data transmission rate from the edge node to the user equipment m, ζ is the transmission power consumption coefficient factor, CiIs the computing power of the edge node, xi is the computing energy consumption coefficient factor, rhomiRepresenting the transmission power, η, between the edge node and the user equipment miRepresentative edgeEnergy consumption coefficient of each CPU cycle of the edge node;
3.3 resource optimization method in hybrid service rental scene
In a mixed service rental scene, for a special computing service, although a plurality of nodes provide services at the same time, the special computing service becomes a virtual node through merging of the nodes, and then a Stancoberg game model is established; the method comprises the following specific steps: firstly, dividing edge nodes into two groups; EN ═ GEN, SEN }; wherein GEN and SEN respectively represent edge node sets in general computing service and special computing service; merging each set of nodes providing dedicated services, each set of nodes being merged into a virtual edge node:
the set of edge nodes may be mapped as:
the Stancoberg gaming model; wherein N is1The number of edge nodes for providing general computing service, and L is the number of virtual nodes for providing special computing service; the hybrid service rental is modeled as a Stancoberg gaming model.
4. The method of claim 1, further comprising grouping dynamic iterative algorithms to solve for nash equalization and presenting pseudo-code in algorithm 2; the computational complexity of the algorithm is O (t)maxM(N1+L))；
And 2, solving the Nash equilibrium based on a grouping dynamic iterative algorithm.
Inputting: the method includes the steps that an edge node set EN is {1,2, …, N }, and a user equipment set UE is {1,2, …, M };
and (3) outputting: optimum price
Optimal computational demand
Initialization:
edge node grouping, EN ═ { GEN, GEN };
the SEN virtual node is mapped on the map,
edge node price Pn，(n∈EN′)；
User equipment requirement Qmn，(m∈UE，n∈EN′)；
From the first iteration t to the maximum number of iterations tmaxAnd adding 1 to t each time, and calculating:
obtaining the optimal demand Q of the computing resource according to the formula (10) from i to 1 to i to M (i is added with 1 each time)in(t)；
From i-1 to i-N1(i add 1 each time) calculation
From i-1 to i-L (i adds 1 each time) calculation
Wherein alpha is the single search step length of the edge node, M is the total number of the user equipment, a is the user satisfaction degree, and rmiFor the data transmission rate from the edge node to the user equipment m, ζ is the transmission power consumption coefficient factor, CiIs the computing power of the edge node, xi is the computing energy consumption coefficient factor, rhomiRepresenting the transmission power, η, between the edge node and the user equipment miThe energy consumption coefficient per CPU cycle representing the edge node.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110549911.XA CN113346938A (en) | 2021-05-20 | 2021-05-20 | Edge computing resource fusion management method for air-space-ground integrated network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110549911.XA CN113346938A (en) | 2021-05-20 | 2021-05-20 | Edge computing resource fusion management method for air-space-ground integrated network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113346938A true CN113346938A (en) | 2021-09-03 |
Family
ID=77469942
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110549911.XA Pending CN113346938A (en) | 2021-05-20 | 2021-05-20 | Edge computing resource fusion management method for air-space-ground integrated network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113346938A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113691305A (en) * | 2021-10-26 | 2021-11-23 | 武汉兴图新科电子股份有限公司 | Cloud-edge-end-based picture transmission method for Beidou third-number short message channel |
| CN114286354A (en) * | 2021-12-31 | 2022-04-05 | 北京航空航天大学 | 6G-oriented air-space-ground integrated network architecture and implementation method |
| CN114499624A (en) * | 2021-12-08 | 2022-05-13 | 上海交通大学 | Multi-source data fusion processing method and system in heaven-earth integrated information network |
| CN114567370A (en) * | 2022-02-22 | 2022-05-31 | 中国电子科技集团公司第五十四研究所 | High-orbit backbone network distributed routing communication method |
| CN115622608A (en) * | 2022-09-29 | 2023-01-17 | 广州爱浦路网络技术有限公司 | Method, system and medium for optimization of offloading strategies based on low-earth-orbit satellite edge calculation |
| CN117118495A (en) * | 2023-08-23 | 2023-11-24 | 中国科学院微小卫星创新研究院 | Space-based general calculation integrated network system and remote sensing data on-orbit processing method |
| CN117118495B (en) * | 2023-08-23 | 2024-05-28 | 中国科学院微小卫星创新研究院 | Space-based general calculation integrated network system and remote sensing data on-orbit processing method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002039619A1 (en) * | 2000-11-07 | 2002-05-16 | Space Systems/Loral Inc. | Dynamic resource management systems |
| CN111400001A (en) * | 2020-03-09 | 2020-07-10 | 清华大学 | Online computing task unloading scheduling method facing edge computing environment |
| CN111555795A (en) * | 2020-04-21 | 2020-08-18 | 清华大学 | Network resource allocation method and device based on user grouping scheduling |
-
2021
- 2021-05-20 CN CN202110549911.XA patent/CN113346938A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002039619A1 (en) * | 2000-11-07 | 2002-05-16 | Space Systems/Loral Inc. | Dynamic resource management systems |
| CN111400001A (en) * | 2020-03-09 | 2020-07-10 | 清华大学 | Online computing task unloading scheduling method facing edge computing environment |
| CN111555795A (en) * | 2020-04-21 | 2020-08-18 | 清华大学 | Network resource allocation method and device based on user grouping scheduling |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113691305A (en) * | 2021-10-26 | 2021-11-23 | 武汉兴图新科电子股份有限公司 | Cloud-edge-end-based picture transmission method for Beidou third-number short message channel |
| CN114499624A (en) * | 2021-12-08 | 2022-05-13 | 上海交通大学 | Multi-source data fusion processing method and system in heaven-earth integrated information network |
| CN114499624B (en) * | 2021-12-08 | 2022-12-13 | 上海交通大学 | Multi-source data fusion processing method and system in heaven-earth integrated information network |
| CN114286354A (en) * | 2021-12-31 | 2022-04-05 | 北京航空航天大学 | 6G-oriented air-space-ground integrated network architecture and implementation method |
| CN114286354B (en) * | 2021-12-31 | 2023-12-08 | 北京航空航天大学 | 6G-oriented space-earth integrated network architecture and implementation method |
| CN114567370A (en) * | 2022-02-22 | 2022-05-31 | 中国电子科技集团公司第五十四研究所 | High-orbit backbone network distributed routing communication method |
| CN114567370B (en) * | 2022-02-22 | 2023-12-29 | 中国电子科技集团公司第五十四研究所 | Distributed routing communication method for high-rail backbone network |
| CN115622608A (en) * | 2022-09-29 | 2023-01-17 | 广州爱浦路网络技术有限公司 | Method, system and medium for optimization of offloading strategies based on low-earth-orbit satellite edge calculation |
| CN117118495A (en) * | 2023-08-23 | 2023-11-24 | 中国科学院微小卫星创新研究院 | Space-based general calculation integrated network system and remote sensing data on-orbit processing method |
| CN117118495B (en) * | 2023-08-23 | 2024-05-28 | 中国科学院微小卫星创新研究院 | Space-based general calculation integrated network system and remote sensing data on-orbit processing method |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113346938A (en) | Edge computing resource fusion management method for air-space-ground integrated network | |
| Gao et al. | Virtual network function placement in satellite edge computing with a potential game approach | |
| CN111757354B (en) | Multi-user slicing resource allocation method based on competitive game | |
| CN111400001B (en) | Online computing task unloading scheduling method facing edge computing environment | |
| CN111641973B (en) | Load balancing method based on fog node cooperation in fog computing network | |
| Xue et al. | A heuristic scheduling algorithm based on PSO in the cloud computing environment | |
| Tham et al. | Stochastic programming methods for workload assignment in an ad hoc mobile cloud | |
| Zhu et al. | Load-balanced virtual network embedding based on deep reinforcement learning for 6G regional satellite networks | |
| Sharmin et al. | Toward sustainable micro-level fog-federated load sharing in internet of vehicles | |
| CN111949409A (en) | Method and system for unloading calculation tasks in electric wireless heterogeneous network | |
| Aloqaily et al. | Fairness-aware game theoretic approach for service management in vehicular clouds | |
| Wang et al. | A profit maximization strategy of mec resource provider in the satellite-terrestrial double edge computing system | |
| CN115022894A (en) | Task unloading and computing resource allocation method and system for low-earth-orbit satellite network | |
| Lakshmi et al. | An adaptive multi-cloud offloading using hierarchical game-theoretic approach | |
| Yang et al. | Towards 6g joint haps-mec-cloud 3c resource allocation for delay-aware computation offloading | |
| CN116489708B (en) | Meta universe oriented cloud edge end collaborative mobile edge computing task unloading method | |
| Li et al. | iDaaS: Inter-datacenter network as a service | |
| Lyu et al. | Multi-leader multi-follower Stackelberg game based resource allocation in multi-access edge computing | |
| Chen et al. | Distributed task offloading game in multiserver mobile edge computing networks | |
| Liu et al. | Scalable traffic management for mobile cloud services in 5G networks | |
| Mai et al. | Computing resource allocation in LEO satellites system: a Stackelberg game approach | |
| Grasso et al. | Slicing a FANET for heterogeneous delay-constrained applications | |
| Kumar et al. | Backhaul-aware storage allocation and pricing mechanism for RSU-based caching networks | |
| Busacca et al. | Drone-assisted edge computing: a game-theoretical approach | |
| CN113708982A (en) | Service function chain deployment method and system based on group learning |
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 |