WO2022021176A1 - 云边协同网络资源平滑迁移与重构方法及系统 - Google Patents

云边协同网络资源平滑迁移与重构方法及系统 Download PDF

Info

Publication number
WO2022021176A1
WO2022021176A1 PCT/CN2020/105691 CN2020105691W WO2022021176A1 WO 2022021176 A1 WO2022021176 A1 WO 2022021176A1 CN 2020105691 W CN2020105691 W CN 2020105691W WO 2022021176 A1 WO2022021176 A1 WO 2022021176A1
Authority
WO
WIPO (PCT)
Prior art keywords
cloud
edge
delay
user
path
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.)
Ceased
Application number
PCT/CN2020/105691
Other languages
English (en)
French (fr)
Inventor
陈伯文
刘玲
沈纲祥
高明义
向练
陈虹
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Suzhou University
Original Assignee
Suzhou University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Suzhou University filed Critical Suzhou University
Publication of WO2022021176A1 publication Critical patent/WO2022021176A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/01Protocols
    • H04L67/10Protocols in which an application is distributed across nodes in the network
    • H04L67/1001Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/01Protocols
    • H04L67/10Protocols in which an application is distributed across nodes in the network
    • H04L67/1001Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004Server selection for load balancing
    • H04L67/101Server selection for load balancing based on network conditions
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/01Protocols
    • H04L67/10Protocols in which an application is distributed across nodes in the network
    • H04L67/1001Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004Server selection for load balancing
    • H04L67/1012Server selection for load balancing based on compliance of requirements or conditions with available server resources
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/01Protocols
    • H04L67/10Protocols in which an application is distributed across nodes in the network
    • H04L67/1001Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1029Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers using data related to the state of servers by a load balancer

Definitions

  • the invention relates to the technical field of cloud computing and edge computing networks, in particular to a method and system for smooth migration and reconstruction of cloud-edge collaborative network resources.
  • MEC Mobile Edge Computing
  • the MEC system allows devices to offload computing tasks to network edge nodes, such as base stations, wireless access points, etc., which not only satisfies the expansion requirements of terminal devices' computing capabilities, but also makes up for the long delay of cloud computing.
  • MEC technology helps to achieve key technical indicators such as ultra-low latency, ultra-high energy efficiency, and ultra-high reliability for 5G services.
  • both of these allocation methods are resource allocation methods designed for only one purpose.
  • the delay-first allocation method can effectively reduce the data transmission delay, but it cannot improve the computing capability of the node to process services, and the data processing delay still needs to be optimized; the resource-first allocation method can improve the speed of the node processing services, but the transmission Latency is likely to increase significantly.
  • the technical problem to be solved by the present invention is to overcome the problems of low data transmission delay and high node processing business computing capability in the prior art, so as to provide a method that can ensure low data transmission delay while ensuring node processing A method and system for smooth migration and reconstruction of cloud-edge collaborative network resources with high business computing capability.
  • a method for smooth migration and reconstruction of cloud-edge collaborative network resources of the present invention includes: reading cloud-edge collaborative network topology information, initializing the parameters of the cloud-edge collaborative network; A set of connection requests; process any connection request in the connection request set until the processing of all connection requests is completed; the method for processing any connection request is: first determine whether the local area requested by the user is There are enough computing resources. If so, the user request is sent to the edge computing server in the local area for processing. If not, continue to judge whether the server in the adjacent area has enough computing resources to process the user request. If so, migrate the user request.
  • network topology information in the cloud-edge collaboration network, network topology information, network connection status, number of user requests, number of edge computing servers, number of base stations and switches are configured.
  • the method for calculating the sum of the path transmission delay and data processing delay of multiple candidate paths from the user request to the edge computing server is: calculating the path transmission delay and data processing delay of the multiple paths respectively. Delay, and then calculate the sum of the path transmission delay and data processing delay.
  • the method for allocating spectrum resources to the working path is: assigning spectrum resources to the working path according to the constraints of spectrum consistency and spectrum continuity.
  • the computing resources of the edge computing server processing user requests are updated in real time, the number of successfully established connection requests is recorded, and the cloud-edge collaboration network status is updated.
  • the spectrum resources occupied by the working path are released; at the same time, the computing resources of the edge computing server processing the user request are released; finally, the connection request is established the working path to clear the information.
  • the method for calculating the average path transmission delay and the average data processing delay requested by the user is: according to the formula Calculate the average path transmission delay aveTRD requested by the user, where sucConReq represents the number of successful connection requests; according to the formula Calculate the average path transmission delay aveCOD requested by the user, where sucConReq indicates the number of successful connection establishment requests.
  • step S5 after the step S5 is completed, the state of the cloud-edge collaboration network is continued to be monitored.
  • the present invention also provides a system for smooth migration and reconstruction of cloud-edge collaborative network resources, comprising: a cloud-edge collaborative network initialization module for reading cloud-edge collaborative network topology information, and performing parameters of the cloud-edge collaborative network. initialization; a connection request generation module, used to generate a set of connection request sets according to user requests; a connection request processing module, used to process any connection request in the connection request set until the processing of all connection requests is completed;
  • the method for processing any connection request is as follows: first determine whether the local area requested by the user has sufficient computing resources, if so, the user request is sent to the edge computing server in the local area for processing, if not, continue to determine the relevant Whether the server in the adjacent area has enough computing resources to process the user request, if so, migrate the user request to other regional edge computing servers on the same switch as the local area; if not, send the user request to the cloud server through the switch Perform data processing on the server; calculate the sum of the path transmission delay and data processing delay of multiple candidate paths from the user request to the edge computing
  • the working path is set according to the sensitivity of the service to the delay, and the working path with low delay sensitivity is selected for transmission and processing, which is conducive to reducing the delay of the service.
  • Fig. 1 is the flow chart of the method for smooth migration and reconstruction of cloud-edge collaborative network resources according to the present invention
  • Fig. 2 is the network diagram of cloud-side collaborative network resource migration and reconstruction of the present invention
  • FIG. 3 is a schematic diagram of the cloud-edge collaborative network resource smooth migration and reconstruction system of the present invention.
  • this embodiment provides a method for smooth migration and reconstruction of cloud-edge collaborative network resources, including: step S1: reading cloud-edge collaborative network topology information, and initializing the parameters of the cloud-edge collaborative network; Step S2: generate a group of connection request sets according to user requests; Step S3: process any connection request in the connection request set until the processing of all connection requests is completed; wherein the method for processing any connection request It is: first judge whether the local area requested by the user has sufficient computing resources, if so, the user request is sent to the edge computing server in the local area for processing, if not, continue to judge whether the server in the adjacent area has sufficient computing resources for processing If there is a user request, migrate the user request to other regional edge computing servers on the same switch as the local area; if not, send the user request to the cloud server through the switch for data processing; calculate the data from the user request separately The sum of the path transmission delay and data processing delay of multiple candidate paths to the edge computing server, and the path with the lowest delay is used as the working path;
  • the computing is: first judge whether
  • step S1 the cloud-edge collaborative network topology information is read, and the parameters of the cloud-edge collaborative network are initialized, which is beneficial to the operation of services;
  • step S2 a set of connection requests is generated according to the user request, which is conducive to processing the connection request; in the step S3, for each connection request, first determine whether the local area requested by the user has sufficient computing resources, If yes, the user request is sent to the edge computing server in the local area for processing.
  • the network transmission delay can be ignored, and the end-to-end processing delay only needs to consider the calculation delay; if not, continue to judge the adjacent area Whether the internal server has enough computing resources to process user requests, and if so, migrate user requests to other regional edge computing servers on the same switch in the local area.
  • the end-to-end delay includes network transmission delay and computing Resource delay, if not, the user request is sent to the cloud server through the switch for data processing.
  • the end-to-end delay of the user request includes the network transmission delay and computing resource delay in the edge computing area and cloud computing area.
  • the path transmission delay and data processing delay of the service calculate the sum of the path transmission delay and data processing delay of multiple candidate paths from the user request to the edge computing server respectively, and calculate the delay The lowest path is used as the working path. Since the working path with low delay sensitivity is selected for transmission and processing, it is beneficial to reduce the path transmission delay and data processing delay of the service;
  • the requested edge computing server updates the computing resources in real time, which is conducive to recording the number of connection requests successfully established; in the step S4, the average path transmission delay and average data processing delay requested by the user are calculated, and the present invention is conducive to reasonable planning.
  • Resource scheduling, migration, and reconstruction in edge computing and cloud computing areas effectively balancing the relationship between network resources and service latency, and reducing latency as much as possible while reasonably allocating resources to minimize service latency.
  • the nodes have high computing capacity for processing business, so as to improve the service quality of the network.
  • the topology information of the network, the network connection state, the number of user requests, the number of edge computing servers, the number of base stations and switches are configured.
  • the cloud-edge collaboration network G U, B, J, S
  • the topology information of the network, the network connection status, the number of user requests, the number of edge computing servers, the number of base stations and switches are configured.
  • step S2 when a set of connection requests is generated according to the user request, information such as the number of connection requests, the number of spectrum gaps required by different connection requests, and computing resources are configured.
  • step S3 for any connection request, it is determined whether the local server requested by the user has the computing resources required by the user request, and if the computing resources of the local server are sufficient, the user request is directly processed locally. If the computing resources of the local server are insufficient, consider whether the edge computing servers in other regions outside the local region have the computing resources required by the user request. If the servers in other regions have sufficient computing resources, the user request will be migrated to other regions through the switch for processing. ; If the computing resources of the local and other regions do not meet the computing resources requested by the user, the user request will be migrated to the cloud server through the switch for processing.
  • the above-mentioned hierarchical deployment will help reduce the path transmission delay and data processing time of the service. extension.
  • the method for calculating the sum of the path transmission delay and data processing delay of multiple candidate paths from the user request to the edge computing server is as follows: calculate the path transmission delay and data processing delay of the multiple paths respectively, and then calculate the path transmission delay.
  • the sum of the data processing delay and the path transmission delay of the service is beneficial to reduce the service path transmission delay and data processing delay.
  • the K shortest path algorithm is used to calculate the K candidate paths from the user request to the server, so as to find the optimal path path as the working path.
  • the K shortest path algorithm is used to calculate the working path from the user request to the edge computing server.
  • the K shortest path algorithm calculates K candidate paths and arranges them in ascending order of distance, that is, the smaller the path distance, the higher the priority.
  • paths with lower priorities are selected in turn to allocate spectrum resources until the resources are allocated successfully or all paths are blocked.
  • the method for allocating spectrum resources to the working path is: assigning spectrum resources to the working path according to the constraints of spectrum consistency and spectrum continuity.
  • the selected working path is searched for the bandwidth resources required to satisfy the connection request. If both spectrum continuity and spectrum consistency are satisfied at the same time If the constraint conditions are met, the connection request is successfully established; if the dual constraints of spectrum continuity and spectrum consistency cannot be satisfied at the same time, the connection request establishment fails.
  • connection request CR(u, f, r) successfully establishes the working path
  • the spectrum allocation algorithm that hits the first time is used, according to all links on the path.
  • a spectrum resource table is generated for numbering the spectrum resource status, and the available spectrum gaps are searched from the end with the smaller number. If an available spectrum gap is found, spectrum resource allocation is performed and spectrum status update is performed; if no spectrum is found, spectrum allocation fails and services are blocked.
  • the spectrum resources occupied by the working path are released; at the same time, the computing resources of the edge computing server processing user requests are released; finally, the information of the working path established by the connection request is cleared.
  • the method for calculating the average path transmission delay and the average data processing delay requested by the user is: according to the formula Calculate the average path transmission delay aveTRD requested by the user, where sucConReq represents the number of successful connection requests; according to the formula Calculate the average path transmission delay aveCOD requested by the user, where sucConReq indicates the number of successful connection establishment requests.
  • Transmission delay After all connection requests are processed, record the cumulative computing resources of each connection request and the initial computing resources of the edge computing server, and use the average path transmission delay aveCOD to calculate the average data processing delay of a group of connection requests .
  • step S4 continue to monitor the state of the cloud-edge collaboration network. Specifically, it mainly completes the initialization of cloud-edge collaborative network, connection request generation, service priority selection, edge computing server selection, working path establishment, spectrum resource allocation, computing resource update, resource release, network transmission delay calculation and data processing time.
  • the state monitoring function of delayed computing is implemented to achieve the goal of reducing service delay as much as possible when computing resources are allocated.
  • the method also includes the steps of judging and warning. Specifically, the coordination function between each module is performed, and the judgment and early warning function of each module is successfully established, so as to achieve the goal of reducing service delay in mobile edge computing.
  • the business with higher delay sensitivity is preferentially processed according to the business delay sensitivity, and the edge computing server processing the business is selected according to the mobile edge computing hierarchical deployment strategy.
  • the K shortest path algorithm is used to calculate the working path between the service and the edge computing server.
  • the spectrum allocation algorithm with the first hit is used to allocate spectrum resources to the path, which needs to satisfy both spectrum consistency and spectrum continuity. Constraints; then update the status of network computing resources and spectrum resources in real time; after each connection request is successfully established, the path transmission delay is calculated according to the path length and path transmission rate, and the resource capacity and Data processing delay.
  • the present invention effectively balances the relationship between network resources and service delay through cloud-side coordination network resource migration and reconstruction sensitive to service delay, and reduces the delay as much as possible under the condition of rational resource allocation.
  • the present invention adopts the classification processing method of service delay sensitivity according to the difference of service transmission sensitivity to time delay, and preferentially transmits and processes services with high time delay sensitivity.
  • the present invention only needs to consider two different delays, including network transmission delay and computing resource delay.
  • the network transmission delay refers to the shortest path length between the user's service area and the edge computing server, which is calculated by the cumulative delay of the link delay; the computing resource delay is related to the computing resource requirements of each user and the edge computing server. computing power.
  • the following describes how to allocate computing resources through hierarchical deployment of mobile edge computing, how to process user requests according to service delay sensitivity, how to calculate path transmission delay through link accumulation delay, and how to use edge computing server resources and user request resources. Calculate the data processing delay to minimize the total business delay.
  • layer 1 has three local areas, each area has base stations and corresponding edge computing servers, and layer 2 is a cloud network composed of switches and corresponding edge computing servers. Assume that the computing resources of the edge computing server in the local area of layer 1 are 20, the computing resources of the edge computing server of the cloud network of layer 2 are 1000, and the number of service computing resources and the number of spectrum slots are randomly generated.
  • the connection request is represented by CR(u, f, r), where u represents the user request number, f represents the number of spectrum slots required to establish a working path, and r represents the computing resources required by the user request.
  • Three user request sets CR(1,3,15), CR(2,8,10) and CR(3,5,30) are generated in the base station area of node 3 and node 2 in Fig. 2 .
  • the dynamic computing resource of node 2 is 5 at this time, it cannot meet the computing resources required by the user request.
  • the computing resources are migrated to the edge computing server node 8 in other regions, the computing resources of the node 8 are updated to 10, and the transmission path is 2.
  • the computing resources on the local and adjacent edge computing area servers cannot meet the computing resources required by the user request, and the user request needs to pass through the switch node. 4 is transmitted to the cloud computing area server node 5 connected to the switch, and the data transmission route is 3.
  • the routing algorithm of K shortest paths needs to be used to calculate node 2 to node 1, node 1 K paths from 2 to node 8 and from node 3 to node 5.
  • the first hit spectrum allocation algorithm is used, and the user requests CR(1,3,15), CR(2,8,10), CR according to the constraints of spectrum consistency and spectrum continuity.
  • the working paths of (3, 5, 30) perform spectrum resource allocation. After the spectrum resource is allocated, the connection request is successfully established. At this time, the status of computing resources and spectrum resources is updated in real time and the number of successful connections is recorded.
  • this embodiment provides a cloud-edge collaborative network resource smooth migration and reconstruction system.
  • the principle of solving the problem is similar to the cloud-edge collaborative network resource smooth migration and reconstruction method. Repeat will not be repeated here.
  • This embodiment provides a cloud-edge collaborative network resource smooth migration and reconstruction system, including:
  • the cloud-edge collaborative network initialization module is used to read the cloud-edge collaborative network topology information, and initialize the parameters of the cloud-edge collaborative network;
  • connection request generation module is used to generate a set of connection request sets according to user requests
  • connection request processing module is used to process any connection request in the connection request set until the processing of all connection requests is completed; the method for processing any connection request is: first determine the local area requested by the user Whether there are enough computing resources, if so, the user request is sent to the edge computing server in the local area for processing; Migrate to other regional edge computing servers on the same switch in the local area, if not, send the user request to the cloud server through the switch for data processing; calculate multiple candidate paths from the user request to the edge computing server separately
  • the sum of the path transmission delay and the data processing delay is the sum of the path transmission delay and the data processing delay, and the path with the lowest delay is used as the working path; after the spectrum resources are allocated to the working path, the computing resources of the edge computing server processing user requests are updated in real time;
  • the calculation module is used to calculate the average path transmission delay and average data processing delay requested by the user.
  • the embodiments of the present application may be provided as a method, a system, or a computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
  • computer-usable storage media including, but not limited to, disk storage, CD-ROM, optical storage, etc.
  • These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory result in an article of manufacture comprising instruction means, the instructions
  • the apparatus implements the functions specified in the flow or flow of the flowcharts and/or the block or blocks of the block diagrams.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明涉及一种云边协同网络资源平滑迁移与重构方法及系统,包括:读取云边协同网络拓扑信息,对所述云边协同网络的参数进行初始化;根据用户请求生成一组连接请求集合;对所述连接请求集合中的任意一个连接请求进行处理,直至完成对所有连接请求的处理;计算用户请求的平均路径传输时延和平均数据处理时延。本发明可以最大程度地减少业务时延,提高网络的服务质量。

Description

云边协同网络资源平滑迁移与重构方法及系统 技术领域
本发明涉及云计算与边缘计算网络的技术领域,尤其是指一种云边协同网络资源平滑迁移与重构方法及系统。
背景技术
近年来,随着物联网(简称IoT)的高速发展和大量数据应用的广泛使用,用户对于网络计算资源的需求量急剧增长。此外,智能手机、笔记本电脑和平板电脑的技术进步,使得新的高要求服务和应用得以出现。尽管新的移动设备在中央处理器CPU方面越来越强大,但可能无法在短时间内处理业务量大的应用程序。云计算具有强大的计算能力,设备可以通过计算资源卸载,将计算任务传输到远端云服务器执行,从而能够有效缓解计算资源需求量较大的问题。然而,计算任务传输到云端服务器会造成不可接受的时延、增加额外传输能量消耗和带来数据泄露等问题。
为了解决云计算卸载过程中面临的高时延等问题,欧洲电信标准协会(ETSI)于2014年率先提出移动边缘计算(简称MEC)的概念:一种在靠近用户的无线接入网提供IT及云计算的平台,被认为是第五代移动通信的关键技术之一。MEC系统允许设备将计算任务卸载到网络边缘节点处,如基站、无线接入点等,既满足了终端设备计算能力的扩展需求,同时弥补了云计算时延较长的缺点。MEC技术有助于达到5G业务超低时延、超高能效、超高可靠性等关键技术指标。通过将云计算和云存储部署到网络边缘,提供一个具备高性能、低时延与高带宽的电信服务环境,加速网络中各项内容、服务以及应用的分发和下载,让消费者享有更高质量网络体验。
目前,在云计算与边缘计算的研究中,主要有以下两种计算资源分配的 方法:时延优先的计算资源分配方法和资源优先的计算资源分配方法。然而,这两种分配方法都是仅仅针对一个目的而设计的资源分配方法。时延优先的分配方法能够有效降低数据传输时延,但却不能提高节点处理业务的计算能力,数据处理时延仍然需要进行优化处理;资源优先的分配方法可以提高节点处理业务的速度,但是传输时延却有可能显著增加。
发明内容
为此,本发明所要解决的技术问题在于克服现有技术中不能同时满足数据传输时延低,节点处理业务计算能力高的问题,从而提供一种可以保证数据传输时延低,同时保证节点处理业务计算能力高的云边协同网络资源平滑迁移与重构方法及系统。
为解决上述技术问题,本发明的一种云边协同网络资源平滑迁移与重构方法,包括:读取云边协同网络拓扑信息,对所述云边协同网络的参数进行初始化;根据用户请求生成一组连接请求集合;对所述连接请求集合中的任意一个连接请求进行处理,直至完成对所有连接请求的处理;其中对任意一个连接请求进行处理的方法为:先判断用户请求的本地区域是否具有足够的计算资源,若有,则用户请求被发送到本地区域内边缘计算服务器进行处理,若否,继续判断相邻区域内服务器是否具有足够计算资源处理用户请求,若有,把用户请求迁移到与本地区域内的同一交换机上的其他区域边缘计算服务器上,若否,则将用户请求通过交换机发送到云端服务器上进行数据处理;分别计算从用户请求到边缘计算服务器的多条候选路径的路径传输时延和数据处理时延的总和,将时延最低的路径作为工作路径;对所述工作路径进行频谱资源分配后,对处理用户请求的边缘计算服务器的计算资源进行实时更新;计算用户请求的平均路径传输时延和平均数据处理时延。
在本发明的一个实施例中,在所述云边协同网络中,配置网络的拓扑信息、网络连接状态、用户请求数、边缘计算服务器的数目、基站和交换机的数目。
在本发明的一个实施例中,计算从用户请求到边缘计算服务器的多条候 选路径的路径传输时延和数据处理时延的总和方法为:分别计算多条路径的路径传输时延和数据处理时延,再计算路径传输时延和数据处理时延的总和。
在本发明的一个实施例中,计算多条路径的路径传输时延和数据处理时延的方法为:根据公式TRD=Weight k/c计算多条路径的路径传输时延,其中Weight k表示第k条工作路径的路径长度,c表示光纤传输速率;根据公式COD=r u/TR i计算多条路径的数据处理时延,其中r u表示用户请求u所需的计算资源,TR i表示节点i的计算资源总和。
在本发明的一个实施例中,对所述工作路径进行频谱资源分配的方法为:根据频谱一致性和频谱连续性的约束条件对所述工作路径进行频谱资源分配。
在本发明的一个实施例中,对处理用户请求的边缘计算服务器的计算资源进行实时更新后,记录成功建立连接请求数,更新云边协同网络状态。
在本发明的一个实施例中,更新云边协同网络状态时,对工作路径占用的频谱资源进行资源释放;同时,对处理用户请求的边缘计算服务器的计算资源进行释放;最后,将连接请求建立的工作路径进行信息清除。
在本发明的一个实施例中,计算用户请求的平均路径传输时延和平均数据处理时延的方法为:根据公式
Figure PCTCN2020105691-appb-000001
计算用户请求的平均路径传输时延aveTRD,其中sucConReq表示成功建立连接请求的数目;根据公式
Figure PCTCN2020105691-appb-000002
计算用户请求的平均路径传输时延aveCOD,其中sucConReq表示成功建立连接请求的数目。
在本发明的一个实施例中,所述步骤S5完成后,继续对所述云边协同网络的状态进行监控。
本发明的还提供了一种云边协同网络资源平滑迁移与重构系统,包括:云边协同网络初始化模块,用于读取云边协同网络拓扑信息,对所述云边协同网络的参数进行初始化;连接请求生成模块,用于根据用户请求生成一组连接请求集合;连接请求处理模块,用于对所述连接请求集合中的任意一个 连接请求进行处理,直至完成对所有连接请求的处理;其中对任意一个连接请求进行处理的方法为:先判断用户请求的本地区域是否具有足够的计算资源,若有,则用户请求被发送到本地区域内边缘计算服务器进行处理,若否,继续判断相邻区域内服务器是否具有足够计算资源处理用户请求,若有,把用户请求迁移到与本地区域内的同一交换机上的其他区域边缘计算服务器上,若否,则将用户请求通过交换机发送到云端服务器上进行数据处理;分别计算从用户请求到边缘计算服务器的多条候选路径的路径传输时延和数据处理时延的总和,将时延最低的路径作为工作路径;对所述工作路径进行频谱资源分配后,对处理用户请求的边缘计算服务器的计算资源进行实时更新;计算模块,用于计算用户请求的平均路径传输时延和平均数据处理时延。
本发明的上述技术方案相比现有技术具有以下优点:
本发明所述的云边协同网络资源平滑迁移与重构方法及系统,根据业务对时延敏感度的高低设置工作路径,选择时延敏感度低的工作路径传输与处理,有利于降低业务的路径传输时延和数据处理时延;另外,采用移动边缘计算分层部署的策略分配计算资源和带宽资源,首先考虑产生业务的本地区域内服务器是否具有处理业务所需网络资源,其次考虑通过交换机连接的其他区域内服务器是否具有足够的网络资源,若本地区域和其他区域都没有足够的网络资源,则考虑通过交换机将业务迁移到云端内服务器进行处理,从而通过分层部署进一步降低业务的路径传输时延和数据处理时延。
附图说明
为了使本发明的内容更容易被清楚的理解,下面根据本发明的具体实施例并结合附图,对本发明作进一步详细的说明,其中
图1是本发明云边协同网络资源平滑迁移与重构方法流程图;
图2是本发明云边协同网络资源迁移与重构的网络图;
图3是本发明云边协同网络资源平滑迁移与重构系统的示意图。
具体实施方式
实施例一
如图1所示,本实施例提供一种云边协同网络资源平滑迁移与重构方法,包括:步骤S1:读取云边协同网络拓扑信息,对所述云边协同网络的参数进行初始化;步骤S2:根据用户请求生成一组连接请求集合;步骤S3:对所述连接请求集合中的任意一个连接请求进行处理,直至完成对所有连接请求的处理;其中对任意一个连接请求进行处理的方法为:先判断用户请求的本地区域是否具有足够的计算资源,若有,则用户请求被发送到本地区域内边缘计算服务器进行处理,若否,继续判断相邻区域内服务器是否具有足够计算资源处理用户请求,若有,把用户请求迁移到与本地区域内的同一交换机上的其他区域边缘计算服务器上,若否,则将用户请求通过交换机发送到云端服务器上进行数据处理;分别计算从用户请求到边缘计算服务器的多条候选路径的路径传输时延和数据处理时延的总和,将时延最低的路径作为工作路径;对所述工作路径进行频谱资源分配后,对处理用户请求的边缘计算服务器的计算资源进行实时更新;步骤S4:计算用户请求的平均路径传输时延和平均数据处理时延。
本实施例所述云边协同网络资源平滑迁移与重构方法,所述步骤S1中,读取云边协同网络拓扑信息,对所述云边协同网络的参数进行初始化,有利于业务的运行;所述步骤S2中,根据用户请求生成一组连接请求集合,有利于对连接请求进行处理;所述步骤S3中,对于每一个连接请求,先判断用户请求的本地区域是否具有足够的计算资源,若有,则用户请求被发送到本地区域内边缘计算服务器进行处理,此时的网络传输时延可以忽略不计,端到端的处理时延只需考虑计算时延;若否,继续判断相邻区域内服务器是否具有足够计算资源处理用户请求,若有,把用户请求迁移到与本地区域内的同一交换机上的其他区域边缘计算服务器上,此时,端到端时延包括网络传输时延和计算资源时延,若否,则将用户请求通过交换机发送到云端服务器上进行数据处理,此时,用户请求端到端时延包括边缘计算区域和云计算区域的网络传输时延和计算资源时延,通过分层部署有利于降低业务的路径传输时延和数据处理时延;分别计算从用户请求到边缘计算服务器的多条候 选路径的路径传输时延和数据处理时延的总和,将时延最低的路径作为工作路径,由于选择时延敏感度低的工作路径传输与处理,有利于降低业务的路径传输时延和数据处理时延;对所述工作路径进行频谱资源分配后,对处理用户请求的边缘计算服务器进行计算资源的实时更新,有利于记录成功建立的连接请求数;所述步骤S4中,计算用户请求的平均路径传输时延和平均数据处理时延,本发明有利于合理规划边缘计算与云计算区域的资源调度、迁移、重构,有效平衡网络资源和业务时延之间的关系,在合理分配资源的情况下尽可能减少时延,以最大程度地减少业务时延,同时保证节点处理业务计算能力高,以提高网络的服务质量。
所述步骤S1中,在所述云边协同网络中,配置网络的拓扑信息、网络连接状态、用户请求数、边缘计算服务器的数目、基站和交换机的数目。具体地,在云边协同网络G(U,B,J,S)中,配置网络的拓扑信息、网络连接状态、用户请求数、边缘计算服务器的数目、基站和交换机的数目。
所述步骤S2中,根据用户请求生成一组连接请求集合时,配置连接请求数目、不同连接请求所需的频谱间隙个数和计算资源等信息。
所述步骤S3中,对于任意一个连接请求,判定用户请求的本地服务器是否具有用户请求所需计算资源,若本地服务器计算资源足够,则用户请求直接在本地进行处理。若本地服务器计算资源不够,考虑本地区域外的其他区域边缘计算服务器是否具有用户请求所需的计算资源,若其他区域的服务器具有足够的计算资源,则将用户请求通过交换机迁移到其他区域进行处理;而若本地和其他区域的计算资源都不满足用户请求的计算资源,则将用户请求通过交换机迁移到云服务器进行处理,上述通过分层部署有利于降低业务的路径传输时延和数据处理时延。
计算从用户请求到边缘计算服务器的多条候选路径的路径传输时延和数据处理时延的总和方法为:分别计算多条路径的路径传输时延和数据处理时延,再计算路径传输时延和数据处理时延总和,有利于降低业务的路径传输时延和数据处理时延。
具体地,根据连接请求CR(u,f,r)的用户请求和处理请求的边缘计算服务器,采用K条最短路径算法,计算出从用户请求到服务器的K条候选路径,以便查找出最优的路径作为工作路径。
其中采用K条最短路径路算法计算从用户请求到边缘计算服务器之间的工作路径。K条最短路径路算法计算了K条候选路径,并按距离由小到大升序排列,即路径距离越小,优先选择权就越高。当具有高优先级的路径在某一段链路上发生阻塞时,就依次选取较低优先级的路径进行频谱资源分配,直到分配资源成功或所有路径全都阻塞。
计算多条路径的路径传输时延和数据处理时延的方法为:根据公式TRD=Weight k/c计算多条路径的路径传输时延,其中Weight k表示第k条工作路径的路径长度,c表示光纤传输速率;根据公式COD=r u/TR i计算多条路径的数据处理时延,其中r u表示用户请求u所需的计算资源,TR i表示节点i的计算资源总和。
对所述工作路径进行频谱资源分配的方法为:根据频谱一致性和频谱连续性的约束条件对所述工作路径进行频谱资源分配。
具体地,根据连接请求CR(u,f,r)所需的频谱间隙数f,在所选择的工作路径中查找满足连接请求所需的带宽资源,若同时满足频谱连续性与频谱一致性双重约束条件,则成功建立连接请求;若不能同时满足频谱连续性与频谱一致性双重约束条件,则连接请求建立失败。
连接请求CR(u,f,r)成功建立工作路径后,根据频谱一致性和频谱连续性的约束条件对工作路径进行频谱资源分配时,采用首次命中的频谱分配算法,根据路径上所有链路的频谱资源状态生成一张频谱资源表进行编号,从标号小的一端开始查找可用的频谱间隙。如果找到可用的频谱间隙则进行频谱资源分配并进行频谱状态更新;如果没有找到则频谱分配失败,业务阻塞。
在频谱资源成功分配后,对处理用户请求的边缘计算服务器进行计算资源的实时更新,对处理用户请求的边缘计算服务器进行计算资源的实时更新 后,记录成功建立连接请求数,更新云边协同网络状态。
更新云边协同网络状态时,对工作路径占用的频谱资源进行资源释放;同时,对处理用户请求的边缘计算服务器的计算资源进行释放;最后,将连接请求建立的工作路径进行信息清除。
对于剩余的用户请求,均采用上述相同的方法直至完成对所有连接请求的处理。
所述步骤S4中,计算用户请求的平均路径传输时延和平均数据处理时延的方法为:根据公式
Figure PCTCN2020105691-appb-000003
计算用户请求的平均路径传输时延aveTRD,其中sucConReq表示成功建立连接请求的数目;根据公式
Figure PCTCN2020105691-appb-000004
计算用户请求的平均路径传输时延aveCOD,其中sucConReq表示成功建立连接请求的数目。
具体地,在每一个连接请求成功建立后,记录每个连接请求传输的最短路径长度,计算每个连接请求的路径传输时延,利用平均路径传输时延aveTRD计算得到一组连接请求的平均路径传输时延;在所有连接请求处理完成后,记录每个连接请求的用户请求累计计算资源和边缘计算服务器的初始计算资源,利用平均路径传输时延aveCOD计算一组连接请求的平均数据处理时延。
所述步骤S4完成后,继续对所述云边协同网络的状态进行监控。具体地,主要完成对云边协同网络初始化、连接请求生成、业务优先级选择、边缘计算服务器选择、工作路径建立、频谱资源分配、计算资源更新、资源释放、网络传输时延计算和数据处理时延计算的状态监控功能,以实现在计算资源分配时尽可能减少业务时延的目标。
另外,本方法还包括判决和预警的步骤。具体地,执行各个模块之间的协调功能,以及每个模块是否建立成功的判决与预警功能,完成移动边缘计算中减少业务时延的目标。
本发明中,对每一个连接请求,根据业务时延敏感度优先处理时延敏感度更高的业务,根据移动边缘计算分层部署的策略选择处理业务的边缘计算 服务器。采用K条最短路径算法计算业务到边缘计算服务器之间的工作路径,工作路径成功选择后,采用首次命中的频谱分配算法对路径进行频谱资源分配,需要同时满足频谱一致性和频谱连续性两个约束条件;然后对网络计算资源和频谱资源状态进行实时更新;每一个连接请求成功建立后,根据路径长度和路径传输速率计算路径传输时延,根据业务计算资源和边缘计算服务器共同计算资源容量和数据处理时延。本发明通过业务时延敏感的云边协同网络资源迁移与重构,有效平衡网络资源和业务时延之间的关系,在合理分配资源的情况下尽可能减少时延。
另外,本发明根据业务传输对时延敏感程度的不同,采用业务时延敏感程度分类处理方法,对时延敏感程度高的业务优先传输与处理。本发明只需要考虑两种不同的时延,包括网络传输时延和计算资源时延。网络传输时延是指用户的服务区域与边缘计算服务器之间的最短路径长度,以其链路时延的累计时延计算;计算资源时延与每个用户的计算资源需求和边缘计算服务器的计算能力有关。
下面具体阐述如何通过移动边缘计算分层部署分配计算资源,如何根据业务时延敏感度处理用户请求,如何通过链路累计时延计算路径传输时延,以及如何通过边缘计算服务器资源和用户请求资源计算数据处理时延,进而尽可能减少业务的总时延。
如图2所示,层1有三个本地区域,每个区域内都有基站和对应的边缘计算服务器,层2是由交换机和对应的边缘计算服务器组成的云端网络。假设层1本地区域内的边缘计算服务器计算资源为20,层2云端网络边缘计算服务器计算资源为1000,业务计算资源数和频谱隙数随机生成。
首先,对云边协同网络G(U,B,J,S)进行初始化,包括用户请求、基站、交换机、边缘计算服务器可以选取的位置,同时对边缘计算服务器的计算资源进行初始化。连接请求用CR(u,f,r)表示,u表示用户请求编号,f表示建立工作路径所需要的频谱隙数,r表示用户请求所需计算资源。在图2中的节点3和节点2基站区域内生成3个用户请求集合CR(1,3,15)、CR (2,8,10)、CR(3,5,30)。
第二,对于节点2产生的用户请求CR(1,3,15),首先判定用户请求是否能够在本地处理。由于用户请求的计算资源数为15,本地区域内边缘计算服务器节点2的计算资源为20,此时用户请求被发送到本地边缘计算服务器进行处理,数据传输路线为①。
第三,对于节点2产生的用户请求CR(2,8,10),由于节点2此时的动态计算资源为5,不能满足用户请求所需计算资源,因此需要通过交换机节点7把用户需求的计算资源迁移到其他区域边缘计算服务器节点8上,节点8的计算资源更新为10,传输路径为②。
第四,对于节点3产生的用户请求CR(3,5,30),此时本地和相邻边缘计算区域服务器上的计算资源都不能满足用户请求所需的计算资源,用户请求需要通过交换机节点4传输到与交换机相连的云计算区域服务器节点5,数据传输路线为③。
第五,对于用户请求CR(1,3,15)、CR(2,8,10)、CR(3,5,30)需要采用K条最短路径的路由算法分别计算节点2到节点1、节点2到节点8、节点3到节点5之间的K条路径。根据公式TRD=Weight k/c和公式COD=r u/TR i分别计算K条路径的路径传输时延和数据处理时延总和,选择K条路径中时延最低的一条路径作为工作路径。
第六,选定工作路径后,采用首次命中的频谱分配算法,根据频谱一致性和频谱连续性的约束条件对用户请求CR(1,3,15)、CR(2,8,10)、CR(3,5,30)的工作路径进行频谱资源分配。频谱资源分配后,连接请求成功建立,此时计算资源和频谱资源状态实时更新并记录连接成功数。
最后,每次连接请求成功建立后,记录工作路径的链路长度、连接请求的计算资源,用公式
Figure PCTCN2020105691-appb-000005
Figure PCTCN2020105691-appb-000006
分别计算平均路径传输时延和平均数据处理时延。
实施例二
如图3所述,基于同一发明构思,本实施例提供了云边协同网络资源平滑迁移与重构系统,其解决问题的原理与所述云边协同网络资源平滑迁移与重构方法类似,重复之处不再赘述。
本实施例提供一种云边协同网络资源平滑迁移与重构系统,包括:
云边协同网络初始化模块,用于读取云边协同网络拓扑信息,对所述云边协同网络的参数进行初始化;
连接请求生成模块,用于根据用户请求生成一组连接请求集合;
连接请求处理模块,用于对所述连接请求集合中的任意一个连接请求进行处理,直至完成对所有连接请求的处理;其中对任意一个连接请求进行处理的方法为:先判断用户请求的本地区域是否具有足够的计算资源,若有,则用户请求被发送到本地区域内边缘计算服务器进行处理,若否,继续判断相邻区域内服务器是否具有足够计算资源处理用户请求,若有,把用户请求迁移到与本地区域内的同一交换机上的其他区域边缘计算服务器上,若否,则将用户请求通过交换机发送到云端服务器上进行数据处理;分别计算从用户请求到边缘计算服务器的多条候选路径的路径传输时延和数据处理时延的总和,将时延最低的路径作为工作路径;对所述工作路径进行频谱资源分配后,对处理用户请求的边缘计算服务器的计算资源进行实时更新;
计算模块,用于计算用户请求的平均路径传输时延和平均数据处理时延。
本领域内的技术人员应明白,本申请的实施例可提供为方法、系统、或计算机程序产品。因此,本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且,本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流 程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器,使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
显然,上述实施例仅仅是为清楚地说明所作的举例,并非对实施方式的限定。对于所属领域的普通技术人员来说,在上述说明的基础上还可以做出其它不同形式变化或变动。这里无需也无法对所有的实施方式予以穷举。而由此所引伸出的显而易见的变化或变动仍处于本发明创造的保护范围之中。

Claims (10)

  1. 一种云边协同网络资源平滑迁移与重构方法,其特征在于,包括如下步骤:
    步骤S1:读取云边协同网络拓扑信息,对所述云边协同网络的参数进行初始化;
    步骤S2:根据用户请求生成一组连接请求集合;
    步骤S3:对所述连接请求集合中的任意一个连接请求进行处理,直至完成对所有连接请求的处理;其中对任意一个连接请求进行处理的方法为:先判断用户请求的本地区域是否具有足够的计算资源,若有,则用户请求被发送到本地区域内边缘计算服务器进行处理,若否,继续判断相邻区域内服务器是否具有足够计算资源处理用户请求,若有,把用户请求迁移到与本地区域内的同一交换机上的其他区域边缘计算服务器上,若否,则将用户请求通过交换机发送到云端服务器上进行数据处理;分别计算从用户请求到边缘计算服务器的多条候选路径的路径传输时延和数据处理时延的总和,将时延最低的路径作为工作路径;对所述工作路径进行频谱资源分配后,对处理用户请求的边缘计算服务器的计算资源进行实时更新;
    步骤S4:计算用户请求的平均路径传输时延和平均数据处理时延。
  2. 根据权利要求1所述的云边协同网络资源平滑迁移与重构方法,其特征在于:在所述云边协同网络中,配置网络的拓扑信息、网络连接状态、用户请求数、边缘计算服务器的数目、基站和交换机的数目。
  3. 根据权利要求1所述的云边协同网络资源平滑迁移与重构方法,其特征在于:计算从用户请求到边缘计算服务器的多条候选路径的路径传输时延和数据处理时延的总和方法为:分别计算多条路径的路径传输时延和数据处理时延,再计算路径传输时延和数据处理时延的总和。
  4. 根据权利要求3所述的云边协同网络资源平滑迁移与重构方法,其特征在于:计算多条路径的路径传输时延和数据处理时延的方法为:根据公式TRD=Weight k/c计算多条路径的路径传输时延,其中Weight k表示第k条工作路径的路径长度,c表示光纤传输速率;根据公式COD=r u/TR i计算多条路径的数据处理时延,其中r u表示用户请求u所需的计算资源,TR i表示节点i的计算资源总和。
  5. 根据权利要求1所述的云边协同网络资源平滑迁移与重构方法,其特征在于:对所述工作路径进行频谱资源分配的方法为:根据频谱一致性和频谱连续性的约束条件对所述工作路径进行频谱资源分配。
  6. 根据权利要求1所述的云边协同网络资源平滑迁移与重构方法,其特征在于:对处理用户请求的边缘计算服务器的计算资源进行实时更新后,记录成功建立连接请求数,更新云边协同网络状态。
  7. 根据权利要求6所述的云边协同网络资源平滑迁移与重构方法,其特征在于:更新云边协同网络状态时,对工作路径占用的频谱资源进行资源释放;同时,对处理用户请求的边缘计算服务器的计算资源进行释放;最后,将连接请求建立的工作路径进行信息清除。
  8. 根据权利要求1所述的云边协同网络资源平滑迁移与重构方法,其特征在于:计算用户请求的平均路径传输时延和平均数据处理时延的方法为:根据公式
    Figure PCTCN2020105691-appb-100001
    计算用户请求的平均路径传输时延aveTRD,其中sucConReq表示成功建立连接请求的数目;根据公式
    Figure PCTCN2020105691-appb-100002
    计算用户请求的平均路径传输时延aveCOD,其中sucConReq表示成功建立连接请求的数目。
  9. 根据权利要求1所述的云边协同网络资源平滑迁移与重构方法,其特征在于:所述步骤S5完成后,继续对所述云边协同网络的状态进行监控。
  10. 一种云边协同网络资源平滑迁移与重构系统,其特征在于,包括:
    云边协同网络初始化模块,用于读取云边协同网络拓扑信息,对所述云边协同网络的参数进行初始化;
    连接请求生成模块,用于根据用户请求生成一组连接请求集合;
    连接请求处理模块,用于对所述连接请求集合中的任意一个连接请求进行处理,直至完成对所有连接请求的处理;其中对任意一个连接请求进行处理的方法为:先判断用户请求的本地区域是否具有足够的计算资源,若有,则用户请求被发送到本地区域内边缘计算服务器进行处理,若否,继续判断相邻区域内服务器是否具有足够计算资源处理用户请求,若有,把用户请求迁移到与本地区域内的同一交换机上的其他区域边缘计算服务器上,若否,则将用户请求通过交换机发送到云端服务器上进行数据处理;分别计算从用户请求到边缘计算服务器的多条候选路径的路径传输时延和数据处理时延的总和,将时延最低的路径作为工作路径;对所述工作路径进行频谱资源分配后,对处理用户请求的边缘计算服务器的计算资源进行实时更新;
    计算模块,用于计算用户请求的平均路径传输时延和平均数据处理时延。
PCT/CN2020/105691 2020-07-28 2020-07-30 云边协同网络资源平滑迁移与重构方法及系统 Ceased WO2022021176A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010739986.X 2020-07-28
CN202010739986.XA CN111901424B (zh) 2020-07-28 2020-07-28 云边协同网络资源平滑迁移与重构方法及系统

Publications (1)

Publication Number Publication Date
WO2022021176A1 true WO2022021176A1 (zh) 2022-02-03

Family

ID=73183704

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/105691 Ceased WO2022021176A1 (zh) 2020-07-28 2020-07-30 云边协同网络资源平滑迁移与重构方法及系统

Country Status (2)

Country Link
CN (1) CN111901424B (zh)
WO (1) WO2022021176A1 (zh)

Cited By (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114698119A (zh) * 2022-03-31 2022-07-01 合肥工业大学 配网分布式保护系统的5g通信/云边计算资源协同分配方法
CN114785777A (zh) * 2022-03-04 2022-07-22 杭州未名信科科技有限公司 一种用于传输资源的端-边-云计算最优解耦方法
CN114827152A (zh) * 2022-07-01 2022-07-29 之江实验室 一种星地协同网络低延时云边端协同计算方法和装置
CN115225630A (zh) * 2022-07-19 2022-10-21 浪潮云信息技术股份公司 一种边缘计算场景下的云边消息通信方法
CN115277452A (zh) * 2022-07-01 2022-11-01 中铁第四勘察设计院集团有限公司 基于边端协同的ResNet自适应加速计算方法及应用
CN115604718A (zh) * 2022-10-09 2023-01-13 东南大学(Cn) 基于dqn的单边缘节点系统频谱与计算资源动态分配方法
CN115696403A (zh) * 2022-11-04 2023-02-03 东南大学 一种边缘计算节点协助的多层边缘计算任务卸载方法
CN115700609A (zh) * 2022-11-16 2023-02-07 浪潮云信息技术股份公司 一种基于边缘云实现协同推理的方法及系统
CN115801811A (zh) * 2023-01-09 2023-03-14 江苏云工场信息技术有限公司 一种云边协同方法及装置
CN115834386A (zh) * 2022-11-04 2023-03-21 北京沐融信息科技股份有限公司 面向边缘计算环境中的智能服务部署方法、系统及终端
CN115866040A (zh) * 2022-11-22 2023-03-28 浙江省公众信息产业有限公司 云边协同的自感知资源分配方法和装置
CN115883661A (zh) * 2022-11-25 2023-03-31 缀初网络技术(上海)有限公司 一种云边协同云游戏场景下的请求动态调度方法
CN116016522A (zh) * 2023-02-13 2023-04-25 广东电网有限责任公司中山供电局 一种云边端协同的新能源终端监控架构
CN116347524A (zh) * 2023-02-20 2023-06-27 中国科学院计算技术研究所 边缘网络中终端设备的低时延任务执行方法和系统
CN116755867A (zh) * 2023-08-18 2023-09-15 中国电子科技集团公司第十五研究所 一种面向卫星云的计算资源调度系统、方法及存储介质
CN116894469A (zh) * 2023-09-11 2023-10-17 西南林业大学 端边云计算环境中的dnn协同推理加速方法、设备及介质
CN117857423A (zh) * 2023-11-29 2024-04-09 慧之安信息技术股份有限公司 一种基于电力的低时延通信路由方法和系统
WO2024092991A1 (zh) * 2022-10-31 2024-05-10 苏州大学 数据中心弹性光网络距离自适应的流量分配方法和系统
CN118113463A (zh) * 2024-01-29 2024-05-31 重庆交通大学 车载边缘计算的服务协同迁移方法、装置、设备及介质
CN119516069A (zh) * 2024-10-20 2025-02-25 南京航空航天大学 一种基于语义通信的实时三维语义地图构建方法
CN119814777A (zh) * 2024-12-05 2025-04-11 天翼云科技有限公司 基于边缘计算系统的任务迁移方法、装置、设备及介质
CN119903552A (zh) * 2024-12-31 2025-04-29 华中科技大学 一种云边融合的安全存储系统

Families Citing this family (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113015217B (zh) * 2021-02-07 2022-05-20 重庆邮电大学 一种基于边云协同低成本的在线多功能业务的计算卸载方法
CN113344152A (zh) * 2021-04-30 2021-09-03 华中农业大学 一种奶品全链条生产信息智能检测上传系统及方法
CN113344728A (zh) * 2021-04-30 2021-09-03 华中农业大学 一种食品生产全链条信息智能监控系统及方法
CN113364850B (zh) * 2021-06-01 2023-02-14 苏州路之遥科技股份有限公司 软件定义云边协同网络能耗优化方法和系统
CN113489643B (zh) * 2021-07-13 2024-09-10 腾讯科技(深圳)有限公司 云边协同的数据传输方法、装置、服务器及存储介质
CN113784373B (zh) 2021-08-24 2022-11-25 苏州大学 云边协同网络中时延和频谱占用联合优化方法及系统
CN113742046A (zh) * 2021-09-17 2021-12-03 苏州大学 流量疏导的云边计算网络计算资源均衡调度方法及系统
CN114363984B (zh) * 2021-12-16 2022-11-25 苏州大学 一种云边协同光载网络频谱资源分配方法及系统
CN114726909B (zh) * 2022-03-15 2024-12-31 阿里云计算有限公司 云服务迁移信息处理方法、装置、设备、介质及产品
CN115396358B (zh) * 2022-08-23 2023-06-06 中国联合网络通信集团有限公司 算力感知网络的路由设置方法、装置及存储介质

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108121512A (zh) * 2017-12-22 2018-06-05 苏州大学 一种边缘计算服务缓存方法、系统、装置及可读存储介质
CN108541027A (zh) * 2018-04-24 2018-09-14 南京邮电大学 一种基于边缘云网络的通信计算资源置换方法
WO2019101056A1 (zh) * 2017-11-21 2019-05-31 华为技术有限公司 一种配置方法及装置
CN110891093A (zh) * 2019-12-09 2020-03-17 中国科学院计算机网络信息中心 一种时延敏感网络中边缘计算节点选择方法及系统

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105975330B (zh) * 2016-06-27 2019-06-18 华为技术有限公司 一种网络边缘计算的虚拟网络功能部署方法、装置和系统
CN106844051A (zh) * 2017-01-19 2017-06-13 河海大学 一种边缘计算环境中功耗优化的负载任务迁移算法
US10887198B2 (en) * 2017-09-29 2021-01-05 Nec Corporation System and method to support network slicing in an MEC system providing automatic conflict resolution arising from multiple tenancy in the MEC environment
US10206094B1 (en) * 2017-12-15 2019-02-12 Industrial Technology Research Institute Mobile edge platform servers and UE context migration management methods thereof
WO2020023115A1 (en) * 2018-07-27 2020-01-30 Futurewei Technologies, Inc. Task offloading and routing in mobile edge cloud networks
CN111010295B (zh) * 2019-11-28 2022-09-16 国网甘肃省电力公司电力科学研究院 一种基于sdn-mec的配用电通信网任务迁移方法
CN111131421B (zh) * 2019-12-13 2022-07-29 中国科学院计算机网络信息中心 一种工业互联网现场大数据与云端信息互联互通的方法
CN111240821B (zh) * 2020-01-14 2022-04-22 华南理工大学 一种基于车联网应用安全性分级的协同云计算迁移方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2019101056A1 (zh) * 2017-11-21 2019-05-31 华为技术有限公司 一种配置方法及装置
CN108121512A (zh) * 2017-12-22 2018-06-05 苏州大学 一种边缘计算服务缓存方法、系统、装置及可读存储介质
CN108541027A (zh) * 2018-04-24 2018-09-14 南京邮电大学 一种基于边缘云网络的通信计算资源置换方法
CN110891093A (zh) * 2019-12-09 2020-03-17 中国科学院计算机网络信息中心 一种时延敏感网络中边缘计算节点选择方法及系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
MAN JUNFENG; ZHAO LONGGAN; PENG CHENG; LI QIANQIAN: "Task Scheduling Method for Large-scale Factory Access in Cloud and Edge Collaborative Computing Architecture", COMPUTER INTEGRATED MANUFACTURING SYSTEMS, vol. 27, no. 8, 1 August 2020 (2020-08-01), CN , pages 2282 - 2294, XP009534058, ISSN: 1006-5911, DOI: 10.13196/j.cims.2021.08.011 *

Cited By (29)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114785777A (zh) * 2022-03-04 2022-07-22 杭州未名信科科技有限公司 一种用于传输资源的端-边-云计算最优解耦方法
CN114698119A (zh) * 2022-03-31 2022-07-01 合肥工业大学 配网分布式保护系统的5g通信/云边计算资源协同分配方法
CN115277452B (zh) * 2022-07-01 2023-11-28 中铁第四勘察设计院集团有限公司 基于边端协同的ResNet自适应加速计算方法及应用
CN114827152A (zh) * 2022-07-01 2022-07-29 之江实验室 一种星地协同网络低延时云边端协同计算方法和装置
CN114827152B (zh) * 2022-07-01 2022-09-27 之江实验室 一种星地协同网络低延时云边端协同计算方法和装置
CN115277452A (zh) * 2022-07-01 2022-11-01 中铁第四勘察设计院集团有限公司 基于边端协同的ResNet自适应加速计算方法及应用
CN115225630A (zh) * 2022-07-19 2022-10-21 浪潮云信息技术股份公司 一种边缘计算场景下的云边消息通信方法
CN115604718A (zh) * 2022-10-09 2023-01-13 东南大学(Cn) 基于dqn的单边缘节点系统频谱与计算资源动态分配方法
WO2024092991A1 (zh) * 2022-10-31 2024-05-10 苏州大学 数据中心弹性光网络距离自适应的流量分配方法和系统
CN115834386A (zh) * 2022-11-04 2023-03-21 北京沐融信息科技股份有限公司 面向边缘计算环境中的智能服务部署方法、系统及终端
CN115696403A (zh) * 2022-11-04 2023-02-03 东南大学 一种边缘计算节点协助的多层边缘计算任务卸载方法
CN115696403B (zh) * 2022-11-04 2023-05-16 东南大学 一种边缘计算节点协助的多层边缘计算任务卸载方法
CN115700609A (zh) * 2022-11-16 2023-02-07 浪潮云信息技术股份公司 一种基于边缘云实现协同推理的方法及系统
CN115866040A (zh) * 2022-11-22 2023-03-28 浙江省公众信息产业有限公司 云边协同的自感知资源分配方法和装置
CN115883661A (zh) * 2022-11-25 2023-03-31 缀初网络技术(上海)有限公司 一种云边协同云游戏场景下的请求动态调度方法
CN115801811B (zh) * 2023-01-09 2023-04-28 江苏云工场信息技术有限公司 一种云边协同方法及装置
CN115801811A (zh) * 2023-01-09 2023-03-14 江苏云工场信息技术有限公司 一种云边协同方法及装置
CN116016522B (zh) * 2023-02-13 2023-06-02 广东电网有限责任公司中山供电局 一种云边端协同的新能源终端监控系统
CN116016522A (zh) * 2023-02-13 2023-04-25 广东电网有限责任公司中山供电局 一种云边端协同的新能源终端监控架构
CN116347524A (zh) * 2023-02-20 2023-06-27 中国科学院计算技术研究所 边缘网络中终端设备的低时延任务执行方法和系统
CN116755867B (zh) * 2023-08-18 2024-03-08 中国电子科技集团公司第十五研究所 一种面向卫星云的计算资源调度系统、方法及存储介质
CN116755867A (zh) * 2023-08-18 2023-09-15 中国电子科技集团公司第十五研究所 一种面向卫星云的计算资源调度系统、方法及存储介质
CN116894469A (zh) * 2023-09-11 2023-10-17 西南林业大学 端边云计算环境中的dnn协同推理加速方法、设备及介质
CN116894469B (zh) * 2023-09-11 2023-12-15 西南林业大学 端边云计算环境中的dnn协同推理加速方法、设备及介质
CN117857423A (zh) * 2023-11-29 2024-04-09 慧之安信息技术股份有限公司 一种基于电力的低时延通信路由方法和系统
CN118113463A (zh) * 2024-01-29 2024-05-31 重庆交通大学 车载边缘计算的服务协同迁移方法、装置、设备及介质
CN119516069A (zh) * 2024-10-20 2025-02-25 南京航空航天大学 一种基于语义通信的实时三维语义地图构建方法
CN119814777A (zh) * 2024-12-05 2025-04-11 天翼云科技有限公司 基于边缘计算系统的任务迁移方法、装置、设备及介质
CN119903552A (zh) * 2024-12-31 2025-04-29 华中科技大学 一种云边融合的安全存储系统

Also Published As

Publication number Publication date
CN111901424A (zh) 2020-11-06
CN111901424B (zh) 2021-09-07

Similar Documents

Publication Publication Date Title
WO2022021176A1 (zh) 云边协同网络资源平滑迁移与重构方法及系统
CN113364850B (zh) 软件定义云边协同网络能耗优化方法和系统
CN113784373B (zh) 云边协同网络中时延和频谱占用联合优化方法及系统
Gu et al. AI-enhanced cloud-edge-terminal collaborative network: Survey, applications, and future directions
WO2023039965A1 (zh) 流量疏导的云边计算网络计算资源均衡调度方法及系统
Chen et al. Latency minimization for mobile edge computing networks
CN109862592B (zh) 一种基于多基站协作的移动边缘计算环境下资源管理和调度方法
CN109104464B (zh) 一种边缘计算环境下面向协同存储的分布式数据更新方法
CN114363984B (zh) 一种云边协同光载网络频谱资源分配方法及系统
CN112822050A (zh) 用于部署网络切片的方法和装置
Tziritas et al. Data replication and virtual machine migrations to mitigate network overhead in edge computing systems
WO2022142277A1 (zh) 一种通信架构的动态调节方法及系统
CN113076177B (zh) 一种边缘计算环境下的虚拟机动态迁移方法
CN113055487A (zh) 一种基于vmec服务网络选择的迁移方法
CN109617811B (zh) 一种云网络中移动应用的快速迁移方法
CN105391651B (zh) 一种虚拟光网络多层资源汇聚方法和系统
WO2023077791A1 (zh) 一种业务处理的方法、系统以及装置
Liu et al. Mobility-aware dynamic service placement for edge computing
Tyagi et al. Load balancing in sdn-enabled wsns toward 6g ioe: Partial cluster migration approach
CN117893294A (zh) 一种用于竞价的分布式低延迟延时任务
CN117579550A (zh) 一种基于可编程交换机和容器状态信息卸载的路由方法
CN113037805B (zh) 面向移动边缘计算的多实例微服务迁移方法
CN119907009B (zh) 一种基于vnf依赖组件的sfc部署和迁移方法
CN106933654B (zh) 一种基于缓存的虚拟机启动方法
Fan et al. Dynamic topology and resource allocation for distributed training in mobile edge computing

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 20947096

Country of ref document: EP

Kind code of ref document: A1