WO2023050469A1 - 一种基于边缘计算的FiWi网络负载均衡方法及系统 - Google Patents

一种基于边缘计算的FiWi网络负载均衡方法及系统 Download PDF

Info

Publication number
WO2023050469A1
WO2023050469A1 PCT/CN2021/123240 CN2021123240W WO2023050469A1 WO 2023050469 A1 WO2023050469 A1 WO 2023050469A1 CN 2021123240 W CN2021123240 W CN 2021123240W WO 2023050469 A1 WO2023050469 A1 WO 2023050469A1
Authority
WO
WIPO (PCT)
Prior art keywords
protection
services
user
network
server
Prior art date
Application number
PCT/CN2021/123240
Other languages
English (en)
French (fr)
Inventor
沈纲祥
王新宇
Original Assignee
苏州大学
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 苏州大学 filed Critical 苏州大学
Priority to US17/911,390 priority Critical patent/US12160466B2/en
Publication of WO2023050469A1 publication Critical patent/WO2023050469A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W28/00Network traffic management; Network resource management
    • H04W28/02Traffic management, e.g. flow control or congestion control
    • H04W28/08Load balancing or load distribution
    • 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
    • H04L67/1004Server selection for load balancing
    • 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
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02DCLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00Reducing energy consumption in communication networks
    • Y02D30/50Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate

Definitions

  • the present invention relates to the field of communication technology, in particular to a method and system for balancing loads in a FiWi network based on edge computing.
  • the fiber optic wireless access network shows great advantages.
  • MEC Mobile Edge Computing
  • 5G Mobile Edge Computing
  • the edge of the network provides computing, storage, and communication resources to meet the requirements of emerging applications, improve user experience, and build a smarter mobile network.
  • all user data requests and computing services need to be completed in the remote large data center, which not only increases the transmission delay of the business, but also increases the traffic between the user and the large data center. transmission bandwidth.
  • edge data centers on access network nodes, all user data requests and computing resource requirements can be obtained from these edge data centers, which can greatly save service feedback delays.
  • the waste of bandwidth resources caused by round-trip transmission can be avoided, so while improving user experience, it can greatly improve the efficiency of network bandwidth resource usage.
  • MEC research focuses on data caching, computing migration, and architecture.
  • Research on the survivability of FiWi networks supporting MEC services is still very scarce, especially when the wireless base station fails, users cannot obtain corresponding MEC services, which affects user experience, and how to restore MEC services for users has become a major problem;
  • load balancing among MEC servers needs to be considered.
  • the purpose of the present invention is to provide a FiWi network load balancing method and system based on edge computing, which can effectively restore MEC services, reduce the maximum load of MEC servers and the consumption of network protection bandwidth resources, and realize server resource load and optical fiber link traffic load balance.
  • the present invention provides a FiWi network load balancing method based on edge computing, comprising the following steps:
  • Establish a path protection strategy For user services, find a candidate protection path that meets the communication delay requirements with its local MEC server, and select the path with the lowest delay among the candidate protection paths that meet the service bandwidth to provide services for user services;
  • Establish a server protection strategy For user services, sort all non-local MEC servers that meet the communication delay requirements in ascending order of load, delete MEC servers that cannot provide sufficient MEC resources, and select the MEC server with the smallest load to provide services for user services;
  • R max represents the maximum MEC server load of the entire network
  • B max represents the maximum used link bandwidth of the entire network
  • R total represents the total capacity of each MEC server
  • B total represents the total capacity of each fiber link
  • Establish a load balancing protection strategy when providing protection for user services, calculate and compare the network resource utilization rates of the path protection strategy and server protection strategy for the business, and select the strategy with the lower network resource utilization rate to provide services for the user business.
  • establishing a path protection strategy specifically includes the following steps:
  • establishing a server protection strategy specifically includes the following steps:
  • establishing a load balancing protection strategy specifically includes the following steps:
  • establishing a load balancing protection strategy also includes the following steps:
  • the value of ⁇ is updated after applying the path protection strategy or the server protection strategy, and finally the strategy with the smallest ⁇ is applied to provide protection for user services. .
  • a FiWi network load balancing system based on edge computing comprising:
  • the path protection strategy module for the user business, searches for a candidate protection path that meets the communication delay requirements between the local MEC server, and selects the path with the lowest delay among the candidate protection paths that meet the service bandwidth to provide services for the user business;
  • the server protection policy module sorts all non-local MEC servers that meet the communication delay requirements in ascending order of load, deletes MEC servers that cannot provide sufficient MEC resources, and selects the MEC server with the smallest load to provide services for user services;
  • the load balancing protection policy module when providing protection for user services, calculates and compares the network resource utilization rates of the service selection path protection policy module and the server protection policy module respectively, and selects the policy module with a smaller network resource utilization rate as the user service Provide services;.
  • R max indicates the maximum MEC server load of the entire network
  • B max indicates the maximum link bandwidth of the entire network
  • R total indicates the total capacity of each MEC server
  • B total indicates the total capacity of each fiber link.
  • the load balancing protection strategy module when it provides protection for each service of the user, it tries the path protection strategy module and the server protection strategy module at the same time to find the corresponding ⁇ value, and the ⁇ value is obtained after the application of the strategy module Update, and finally apply the policy with the smallest value of ⁇ to provide protection for users.
  • the load balancing protection strategy module also includes: judging whether all services in the network are protected: if they are all protected, then detect whether all services of all users in the network are protected, otherwise continue to pass the load
  • the balanced protection policy module protects unprotected services.
  • Beneficial effects of the present invention Considering the survivability of the FiWi network supporting MEC services, for a single wireless base station failure, three protection strategies are proposed in order to restore the user's MEC service, and when a single wireless base station failure occurs in the network, load balancing protection The strategy can not only effectively restore the MEC service of user business, but also minimize the maximum load of the MEC server and network protection bandwidth resources, balance the resource load of the MEC server and the traffic load on the fiber link, and achieve load balancing.
  • Fig. 1 is a schematic flow sheet of the method of the present invention
  • FIG. 2 is a schematic diagram of a FiWi network supporting MEC services
  • Fig. 3 is a schematic diagram of a protection strategy for wireless base station faults in the present invention.
  • FIG. 4 is a schematic diagram of a test network in Embodiment 2 of the present invention.
  • FIG. 5 is a schematic diagram of the maximum load simulation of the MEC server in Embodiment 2 of the present invention.
  • FIG. 6 is a schematic diagram of a simulation of protection bandwidth resources in Embodiment 2 of the present invention.
  • the present invention provides a FiWi network load balancing method based on edge computing, comprising the following steps:
  • Establish a path protection strategy For user services, find a candidate protection path that meets the communication delay requirements with its local MEC server, and select the path with the lowest delay among the candidate protection paths that meet the service bandwidth to provide services for user services;
  • Establish a server protection strategy For user services, sort all non-local MEC servers that meet the communication delay requirements in ascending order of load, delete MEC servers that cannot provide sufficient MEC resources, and select the MEC server with the smallest load to provide services for user services;
  • R max represents the maximum MEC server load of the entire network
  • B max represents the maximum used link bandwidth of the entire network
  • R total represents the total capacity of each MEC server
  • B total represents the total capacity of each fiber link
  • Establish a load balancing protection strategy When providing protection for user services, calculate and compare the network resource utilization rates of the service path protection strategy and server protection strategy, and select the strategy with the lower network resource utilization rate to provide services for the user business.
  • the present invention studies the survivability of a FiWi network that supports MEC services.
  • three protection strategies are proposed in order to restore the user's MEC service.
  • the load balancing protection strategy can both Effectively restore the MEC service of user business, and at the same time minimize the maximum load of the MEC server and network protection bandwidth resources, balance the resource load of the MEC server and the traffic load on the fiber link, and achieve load balancing.
  • FIG. 2 gives an example of a FiWi network supporting MEC services. It consists of an Optical Line Terminal (OLT) and multiple Optical Network Unit Base Stations (ONU-BS). Each ONU-BS provides communication services for user services in the user group, and is connected to a local MEC server to provide MEC services.
  • OLT Optical Line Terminal
  • ONU-BS Optical Network Unit Base Stations
  • LBP load balancing protection
  • B1 and B2 are wireless base stations accessible to user A
  • N1 is the MEC node of user A.
  • user A is served by the MEC server connected to N1 through B1, and its corresponding communication path is A-B1-N1.
  • B1 fails, the wireless link A-B1 is interrupted, and user A cannot obtain the MEC service through the original communication path at this time.
  • PP Path Protection Policy
  • SP Server Protection Policy
  • the path to restore service will become A-B2-N2-N1 (1 dashed line).
  • Specific implementation plan For each business of the user, use the shortest route algorithm to find k candidate protection paths that meet the communication delay requirements between the user and its local MEC server, and save them in the candidate path list; Next, judge Whether all these candidate paths have enough bandwidth; the path that will not provide enough bandwidth will be deleted from the candidate path list; finally, the path with the lowest delay in the candidate list is selected as the protection path.
  • SP Server Protection Policy
  • computing and storage resources are reserved on non-local MEC servers as backups.
  • the service recovery path will become A-B2-N2-N3 (2 dotted lines).
  • Specific implementation plan For each business of the user, sort all non-local MEC servers that meet the communication delay requirements in ascending order of load and save them in a list; next, delete servers that cannot provide enough MEC resources from the list ;Finally, select the first server in the list to serve user traffic.
  • the SP can save the bandwidth resource for accessing the MEC service, the traffic load on the optical fiber link is small, and the protection bandwidth resource consumption in the network will also be small.
  • additional backup MEC resources need to be reserved at this time, which will increase the load of the MEC server, and the maximum load of the MEC server will be high, which may easily cause server load imbalance.
  • LBP load balancing protection strategy
  • LBP Load Balance Protection Policy
  • the LBP policy provides different protection methods for different services according to the current network resource utilization and user service characteristics. Some services are protected by the PP policy, and some services are protected by the SP policy. The goal is to allow the computing-intensive business in user A to adopt the PP strategy when it needs more MEC resources, and the bandwidth-constrained business to use the SP strategy when it needs more bandwidth resources.
  • R max and B max are the maximum MEC server load and the maximum used link bandwidth of the entire network
  • R total and B total are the total capacity of each MEC server and each fiber link.
  • network topology G (N, L), user set C, and service set S c for user c ⁇ C;
  • Step 1 For a user c ⁇ C;
  • Step 2 Obtain one of the services s ⁇ S c ;
  • Step 3 Try the PP strategy for business s and get the corresponding ⁇ value as ⁇ PP,s ; try the SP strategy for s and get the corresponding ⁇ value as ⁇ SP,s : if ⁇ PP,s ⁇ SP, s , use PP strategy for business s and record ⁇ PP,s ; otherwise use SP strategy and record ⁇ SP,s ;
  • Step 4 Determine whether all services in S c are protected; if not, return to step 2 to protect the next service; otherwise, skip to step 5;
  • Step 5 Determine whether all services of all users in C are protected; if not, go back to the first step to protect the next user; otherwise, terminate the process and find the final parameter ⁇ .
  • the load balancing protection policy can not only effectively restore the MEC service of the user business, but also minimize the maximum load of the MEC server and network protection bandwidth resources, and balance the resource load of the MEC server and The traffic load on the fiber optic link realizes load balancing.
  • the performance of the proposed strategy is evaluated in the test network shown in FIG. 4 , and the network has 50 MEC servers and 146 users.
  • the specific simulation parameters are as follows:
  • each MEC server is 800 units, and the maximum bandwidth capacity of each fiber link is 100Gb/s;
  • the MEC resources required by each user are evenly distributed in [X-10, X+10] units, where X is the average value of the MEC resources required by each user, and the bandwidth resources required by each user are uniform Distributed in [1, 3] Gb/s;
  • Each user has 100 MEC services, and the MEC resources and bandwidth resources of each user are randomly assigned to 100 MEC services;
  • each MEC service is randomly set in the range of [18, 22] ms, and the switching delay of each switch is set to 8 ms. Since the distance range is small, the signal propagation in the FiWi network is ignored delay.
  • FIG. 5 shows the results of the MEC server maximum load, which is defined as the highest server resource load among all servers
  • Figure 6 shows the results of the network protection bandwidth resource, which is defined as the sum of all fiber link protection bandwidth resources.
  • the LBP strategy when performing protection, according to the current network resource utilization and user service characteristics, the LBP strategy will try the PP and SP strategies to provide the most suitable protection method for different services. After protecting each MEC service with two strategies, always record the maximum MEC server load and maximum link capacity utilization, and choose the better strategy according to the value of ⁇ . Therefore, compared with PP and SP, the LBP strategy can effectively minimize the maximum load of the MEC server and protect bandwidth resources at the same time to achieve load balancing.
  • the embodiment of the present invention provides a FiWi network load balancing system based on edge computing, including:
  • the path protection strategy module for the user business, searches for a candidate protection path that meets the communication delay requirements between the local MEC server, and selects the path with the lowest delay among the candidate protection paths that meet the service bandwidth to provide services for the user business;
  • the server protection policy module sorts all non-local MEC servers that meet the communication delay requirements in ascending order of load, deletes MEC servers that cannot provide sufficient MEC resources, and selects the MEC server with the smallest load to provide services for user services;
  • the load balancing protection policy module when providing protection for user services, calculates and compares the network resource utilization rates of the service selection path protection policy module and the server protection policy module respectively, and selects the policy module with a smaller network resource utilization rate as the user service Provide services;.
  • R max indicates the maximum MEC server load of the entire network
  • B max indicates the maximum link bandwidth of the entire network
  • R total indicates the total capacity of each MEC server
  • B total indicates the total capacity of each fiber link.
  • the load balancing protection strategy module provides protection for each business of the user, it tries the path protection strategy module and the server protection strategy module at the same time to find the corresponding ⁇ value.
  • the ⁇ value is updated after applying the strategy module, and finally applies the strategy with the smallest ⁇ value Provide protection for users.
  • the load balancing protection strategy module also includes: judging whether all services in the network are protected: if they are all protected, then detecting whether all services of all users in the network are protected, otherwise continue to protect the unprotected services through the load balancing protection strategy module protected business. Detect whether all services of all users in the network are protected: if they are all protected, terminate the process and find the final network resource utilization, otherwise continue to protect the next user through the load balancing protection strategy module.

Landscapes

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

Abstract

本发明公开了一种基于边缘计算的FiWi网络负载均衡方法及系统,包括以下步骤:建立路径保护策略;建立服务器保护策略;建立负载均衡保护策略:为用户业务提供保护时,分别计算并比较该业务进行路径保护策略和服务器保护策略的网络资源利用率,取其中网络资源利用率较小的策略为该用户业务提供服务。本发明能有效恢复MEC服务,同时减少MEC服务器的最大负载和网络保护带宽资源消耗,实现服务器资源负载和光纤链路流量负载的均衡。

Description

一种基于边缘计算的FiWi网络负载均衡方法及系统 技术领域
本发明涉及通信技术领域,具体涉及一种基于边缘计算的FiWi网络负载均衡方法及系统。
背景技术
由于结合了光纤和无线通信技术的优势,光纤无线接入网络(FiWi)显示出巨大的优势。同时,移动边缘计算(MEC),一个新兴5G关键技术,通过将IT服务环境与云计算能力在网络边缘相结合,将原本位于云数据中心的服务和网络功能推向移动网络的边缘,在移动网络的边缘提供计算、存储和通信资源,从而满足新兴应用的要求,提升用户体验,构建更加智能的移动网络。在不支持MEC服务的场景下,所有用户的数据请求和计算业务都需要在远端的大型数据中心中完成,这不仅增加了业务的传输时延,也增加了业务往返用户与大型数据中心间的传输带宽。而当我们在接入网节点上部署边缘数据中心后,所有用户的数据请求和计算资源需求可以从这些边缘数据中心获得,因此可以极大地节省业务反馈时延。同时由于不需要从遥远的大型数据中心获取数据或计算资源,从而可以避免由于往返传输而造成的带宽资源浪费,所以在改善用户体验的同时,可以极大地提升网络带宽资源的使用效率。
目前,MEC的研究集中在数据缓存、计算迁移和架构方面。支持MEC服务的FiWi网络的生存能力的研究还十分稀少,尤其是当无线基站发生故障时,用户便无法获得相应的MEC服务,从而影响用户体验,如何为用户恢复MEC服务成为一大问题;对于负载较大的服务器,当其本地业务的MEC资源需求量增加时,就无法提供足够的资源满足该部分需求,从导致业务的阻塞。因此, 需要考虑均衡MEC服务器间的负载。同时,还应该避免因访问远程服务器而消耗太多的额外带宽,需要考虑使网络保护带宽最小化。传统的保护方案中虽然能够恢复用户的MEC服务,但是由于将用户的所有业务采用单一的保护方案,容易出现MEC服务器负载过大和网络保护带宽资源消耗过大等问题,难以上述两个指标的平衡。
发明内容
本发明的目的是提供一种基于边缘计算的FiWi网络负载均衡方法及系统,能有效恢复MEC服务,同时减少MEC服务器的最大负载和网络保护带宽资源消耗,实现服务器资源负载和光纤链路流量负载的均衡。
为了解决上述技术问题,本发明提供了一种基于边缘计算的FiWi网络负载均衡方法,包括以下步骤:
建立路径保护策略:对于用户业务,寻找与其本地MEC服务器之间满足通信时延要求的候选保护路径,并在满足业务带宽的候选保护路径中选择延迟最低的路径为用户业务提供服务;
建立服务器保护策略:对于用户业务,将所有满足通信时延要求的非本地MEC服务器按照负载升序排序,删除无法提供足够MEC资源的MEC服务器后选择负载最小的MEC服务器为用户业务提供服务;
定义网络资源利用率ρ:
ρ=R max/R total+B max/B total
其中,R max表示整个网络的最大MEC服务器负载,B max表示整个网络的最大使用链路带宽,R total表示每个MEC服务器的总容量,B total表示每条光纤链路的总容量;
建立负载均衡保护策略:为用户业务提供保护时,分别计算并比较该业务 进行路径保护策略和服务器保护策略的网络资源利用率,取其中网络资源利用率较小的策略为该用户业务提供服务。
作为本发明的进一步改进,建立路径保护策略具体包括以下步骤:
对于用户业务,采用最短路由算法寻找k条与其本地MEC服务器之间满足通信时延要求的候选保护路径,并保存在候选路径列表中;
判断所有候选路径是否具有足够的带宽,在候选路径列表中删除无法提供足够带宽的路径;
选择候选路径列表中延迟最低的路径作为保护路径向用户业务提供服务。
作为本发明的进一步改进,建立服务器保护策略具体包括以下步骤:
对于用户业务,将所有满足通信时延要求的非本地MEC服务器按照负载升序排序,并保存在列表中;
从列表中删除无法提供足够MEC资源的MEC服务器;
选择列表中第一个服务器为用户业务提供服务。
作为本发明的进一步改进,建立负载均衡保护策略具体包括以下步骤:
a、对于网络拓扑G=(N,L),用户集合C,用户c∈C的业务集合S c
b、获得一个业务s∈S c
c、对业务s尝试路径保护策略并得到相应的网络资源利用率ρ值,记为ρ PP,s;对业务s尝试服务器保护策略并得到相应的网络资源利用率ρ值,记为ρ SP,s;若ρ PP,s<ρ SP,s,对业务s使用路径保护策略并记录ρ PP,s,否则使用服务器保护策略并记录ρ SP,s
作为本发明的进一步改进,建立负载均衡保护策略还包括以下步骤:
d、判断S c中的所有业务是否都被保护:若都被保护,则进行下一步骤,否则返回步骤b保护下一个业务;
e、判断C中的所有用户的全部业务是否都被保护;若都被保护,则终止进程,并找到最终网络资源利用率,否则返回步骤a保护下一个用户。
作为本发明的进一步改进,所述步骤c中ρ值在应用路径保护策略或服务器保护策略后更新,最终应用ρ最小的策略为用户业务提供保护。。
一种基于边缘计算的FiWi网络负载均衡系统,包括:
路径保护策略模块,对于用户业务,寻找与其本地MEC服务器之间满足通信时延要求的候选保护路径,并在满足业务带宽的候选保护路径中选择延迟最低的路径为用户业务提供服务;
服务器保护策略模块,对于用户业务,将所有满足通信时延要求的非本地MEC服务器按照负载升序排序,删除无法提供足够MEC资源的MEC服务器后选择负载最小的MEC服务器为用户业务提供服务;
负载均衡保护策略模块,为用户业务提供保护时,分别计算并比较该业务选择路径保护策略模块和服务器保护策略模块的网络资源利用率,取其中网络资源利用率较小的策略模块为该用户业务提供服务;。
其中,网络资源利用率ρ:
ρ=R max/R total+B max/B total
R max表示整个网络的最大MEC服务器负载,B max表示整个网络的最大使用链路带宽,R total表示每个MEC服务器的总容量,B total表示每条光纤链路的总容量。
作为本发明的进一步改进,所述负载均衡保护策略模块为用户的每一个业 务提供保护时,同时尝试路径保护策略模块和服务器保护策略模块,找到各自对应的ρ值,ρ值在应用策略模块后更新,最终应用ρ值最小的策略为用户提供保护。
作为本发明的进一步改进,所述负载均衡保护策略模块还包括:判断网络中的所有业务是否都被保护:若都被保护,则检测网络中所有用户的全部业务是否被保护,否则继续通过负载均衡保护策略模块保护未被保护的业务。
作为本发明的进一步改进,检测网络中所有用户的全部业务是否被保护:若都被保护,则终止进程,并找到最终网络资源利用率,否则继续通过负载均衡保护策略模块保护下一个用户。
本发明的有益效果:考虑支持MEC服务的FiWi网的生存性,针对单个无线基站故障,为了恢复用户的MEC服务提出了三种保护策略,且当网络中发生单个无线基站故障时,负载均衡保护策略既能有效恢复用户业务的MEC服务,又能同时最小化MEC服务器的最大负载和网络保护带宽资源,平衡MEC服务器的资源负载和光纤链路上的流量负载,实现负载均衡。
附图说明
图1是本发明方法流程示意图;
图2是支持MEC服务的FiWi网络示意图;
图3是本发明针对无线基站故障的保护策略示意图;
图4是本发明实施例二中测试网络示意图;
图5是本发明实施例二中MEC服务器的最大负载仿真示意图;
图6是本发明实施例二中保护带宽资源仿真示意图。
具体实施方式
下面结合附图和具体实施例对本发明作进一步说明,以使本领域的技术人员可以更好地理解本发明并能予以实施,但所举实施例不作为对本发明的限定。
参考图1,本发明提供了一种基于边缘计算的FiWi网络负载均衡方法,包括以下步骤:
建立路径保护策略:对于用户业务,寻找与其本地MEC服务器之间满足通信时延要求的候选保护路径,并在满足业务带宽的候选保护路径中选择延迟最低的路径为用户业务提供服务;
建立服务器保护策略:对于用户业务,将所有满足通信时延要求的非本地MEC服务器按照负载升序排序,删除无法提供足够MEC资源的MEC服务器后选择负载最小的MEC服务器为用户业务提供服务;
定义网络资源利用率ρ:
ρ=R max/R total+B max/B total
其中,R max表示整个网络的最大MEC服务器负载,B max表示整个网络的最大使用链路带宽,R total表示每个MEC服务器的总容量,B total表示每条光纤链路的总容量;
建立负载均衡保护策略:为用户业务提供保护时,分别计算并比较该业务进行路径保护策略和服务器保护策略的网络资源利用率,取其中网络资源利用率较小的策略为该用户业务提供服务。
本发明研究了支持MEC服务的FiWi网的生存性,针对单个无线基站故障,为了恢复用户的MEC服务提出了三种保护策略,且当网络中发生单个无线基站故障时,负载均衡保护策略既能有效恢复用户业务的MEC服务,又能同时最小化MEC服务器的最大负载和网络保护带宽资源,平衡MEC服务器的资源 负载和光纤链路上的流量负载,实现负载均衡。
实施例一
图2给出了一个支持MEC服务的FiWi网络的示例。它由一个光线路终端(OLT)和多个光网络单元基站(ONU-BS)组成。每个ONU-BS为用户群中的用户业务提供通信服务,并连接到本地MEC服务器以提供MEC服务。
但随着5G网络复杂性的增加,在发生故障或受到攻击时,通信网络的服务能力将会受到严重的影响,这不仅会影响用户的体验,还会给网络运行商造成巨大的经济损失。因此该网络的生存性是需要重视的重点。考虑无线基站(BS)的故障,因为它们暴露在外部环境中,容易遭受故障,并且当网络中的无线基站发生故障时,即使MEC服务器、交换机等节点和光纤链路都完好无损,用户也会因为失去与基站之间的通信连接而失去相应的MEC服务。为了解决故障带来的问题,基于两个基本保护策略提出了一个负载平衡保护(LBP)策略,既能有效恢复MEC服务,又能够同时减少MEC服务器的最大负载和网络保护带宽资源消耗,实现服务器资源负载和光纤链路流量负载的均衡,提高网络性能。
在正常情况下(无网络故障),用户通过信号质量最好的无线基站接入网络,如果无线基站出现故障,用户可以接入附近的备用基站恢复业务。例如在图3中,B1和B2是用户A可访问的无线基站,N1是用户A的MEC节点。在正常情况下,用户A由通过B1连接到N1的MEC服务器获得服务,其对应的通信路径是A-B1-N1。如果B1发生故障,则无线链路A-B1中断,此时用户A无法通过原始通信路径获得MEC服务。为了恢复用户的MEC服务,我们通常可以考虑两种基本的保护策略:路径保护策略(PP)和服务器保护策略(SP)。
(1)路径保护策略(PP):
即通过备用基站绕回到原来的MEC服务器。如图3中的1路径所示,恢 复服务的路径将变为A-B2-N2-N1(1虚线)。具体的实现方案:对于用户的每一个业务,采用最短路由算法来寻找k条与其本地MEC服务器之间满足通信时延要求的候选保护路径,并将它们保存在候选路径列表中;接下来,判断所有这些候选路径是否具有足够的带宽;将不能提供足够带宽的路径将从候选路径列表中删除;最后,选择候选列表中延迟最低的路径作为保护路径。可以看出,在PP策略中,由于重用了原始服务器中的资源而不需要保留备份MEC资源,MEC服务器的最大负载不会很高,不容易造成服务器负载失衡的现象。但是,由于用户“绕道”访问原始MEC服务器,经过的链路可能比较多,因此可能会消耗额外的通信资源,光纤链路上的流量负载较大,网络中的保护带宽资源消耗也会较大。
(2)服务器保护策略(SP):
即在非本地MEC服务器上预留计算和存储资源作为备份。如图3中的2路径所示,假设选取N3中的服务器作为备份服务器,恢复服务的路径将变为A-B2-N2-N3(2虚线)。具体的实现方案:对于用户的每一个业务,将所有满足通信时延要求的非本地MEC服务器按照负载升序排序,并保存在一个列表中;接下来,从列表中删除无法提供足够MEC资源的服务器;最后,选择列表中的第一个服务器来为用户业务提供服务。可以看出,在SP策略中,如果备用MEC服务器就在附近,SP可以节省访问MEC服务的带宽资源,光纤链路上的流量负载较小,网络中的保护带宽资源消耗也会较小。但是此时需要预留额外的备用MEC资源,会增加MEC服务器负载,MEC服务器的最大负载会很高,容易造成服务器负载失衡的现象。
针对这两种基本方案的优缺点,希望获得一种折中的保护策略,避免MEC服务器最大负载过高和网络保护带宽资源消耗过高的极端现象发生。因此结合PP和SP两种策略,提出了一种负载均衡保护策略(LBP)来实现MEC服务恢复,有效地平衡MEC服务器的资源负载和光纤链路上的流量负载,平衡MEC服务器最大负载和网络保护带宽资源两个性能指标。
(3)负载均衡保护策略(LBP):
如图3中的3路径所示,LBP策略根据当前网络资源利用率和用户业务的特点来为不同的业务提供不同的保护方式,部分业务采用PP策略保护,部分业务采用SP策略保护。目标是让用户A中的计算密集型业务在需要更多的MEC资源时采用PP策略,而带宽紧张型业务在需要更多的带宽资源时采用SP策略。
具体的实现方案:为了平衡MEC服务器负载和光纤链路容量利用率,首先定义一个参数ρ,表示网络资源利用率,其值为ρ=R max/R total+B max/B total,其中R max和B max是整个网络的最大MEC服务器负载和最大使用链路带宽,R total和B total是每个MEC服务器和每条光纤链路的总容量。ρ值越小,代表发生服务器负载失衡和链路流量超负荷的可能性越低,更容易实现负载均衡,网络的整体性能也就越好。在为用户的每一个业务提供保护时,同时尝试PP和SP策略,找到它们对应的ρ值。ρ值会在应用PP或SP策略后更新,最终应用ρ最小的策略为用户提供保护。算法具体如下:
输入:网络拓扑G=(N,L),用户集合C,对于用户c∈C的业务集合S c
第一步:对于一个用户c∈C;
第二步:获得其中的一个业务s∈S c
第三步:对业务s尝试PP策略并得到相应的ρ值记为ρ PP,s;对s尝试SP策略并得到相应的ρ值记为ρ SP,s:如果ρ PP,s<ρ SP,s,对业务s使用PP策略并记录ρ PP,s;否则使用SP策略并记录ρ SP,s
第四步:判断S c中的所有业务是否都被保护了;如果没有,返回第二步保护下一个业务;否则跳到第五步;
第五步:判断C中的所有用户的全部业务是否都被保护了;如果没有,回到第一步保护下一个用户;否则终止进程,并找到最终参数ρ。
当网络中发生单个无线基站故障时,负载均衡保护策略(LBP)既能有效 恢复用户业务的MEC服务,又能同时最小化MEC服务器的最大负载和网络保护带宽资源,平衡MEC服务器的资源负载和光纤链路上的流量负载,实现负载均衡。
实施例二
基于上述实施方式和实施例一,如图4所示的测试网络中评估所提出策略的性能,网络有50个MEC服务器和146个用户。具体仿真参数如下:
(1)每个MEC服务器的最大资源容量为800个单元,每条光纤链路的最大带宽容量为100Gb/s;
(2)每个用户所需的MEC资源均匀分布在[X-10,X+10]个单元,其中X是每个用户所需的MEC资源的平均值,每个用户所需的带宽资源均匀分布在[1,3]Gb/s;
(3)每个用户有100个MEC业务,每个用户的MEC资源和带宽资源随机分配给其100个MEC业务;
(4)每个MEC服务的时延约束在[18,22]ms的范围内随机设置,每个交换机的交换时延设置为8ms,由于距离范围很小,因此忽略了FiWi网络中的信号传播时延。
图5显示了MEC服务器最大负载的结果,它被定义为所有服务器中最高的服务器资源负载,图6显示了网络保护带宽资源的结果,它被定义为所有光纤链路保护带宽资源的总和。通过对三种策略的对比,看到PP策略实现了最低的MEC服务器最大负载,但需要最高的保护带宽资源,因为它需要绕回到原始MEC服务器。相比之下,SP策略的MEC服务器最大负载最高,但消耗最低的保护带宽资源,因为它需要在其他服务器上预留额外的MEC资源。最后,所提出的LBP策略最有效地实现了MEC服务器最大负载和保护带宽资源的平衡。具体来说,它接近了PP策略的最大MEC服务器负载和SP策略的保护带宽资源。这是因为在进行保护时,根据当前网络资源利用率和用户业务的 特点,LBP策略会尝试PP和SP策略,为不同的业务提供最合适的保护方式。使用两种策略保护每个MEC业务后,始终记录最大MEC服务器负载和最大链路容量利用率,并根据ρ值选择更好的策略。因此,与PP和SP相比,LBP策略可以有效地同时最小化MEC服务器最大负载和保护带宽资源,实现负载均衡。
实施例三
本发明实施例提供了一种基于边缘计算的FiWi网络负载均衡系统,包括:
路径保护策略模块,对于用户业务,寻找与其本地MEC服务器之间满足通信时延要求的候选保护路径,并在满足业务带宽的候选保护路径中选择延迟最低的路径为用户业务提供服务;
服务器保护策略模块,对于用户业务,将所有满足通信时延要求的非本地MEC服务器按照负载升序排序,删除无法提供足够MEC资源的MEC服务器后选择负载最小的MEC服务器为用户业务提供服务;
负载均衡保护策略模块,为用户业务提供保护时,分别计算并比较该业务选择路径保护策略模块和服务器保护策略模块的网络资源利用率,取其中网络资源利用率较小的策略模块为该用户业务提供服务;。
其中,网络资源利用率ρ:
ρ=R max/R total+B max/B total
R max表示整个网络的最大MEC服务器负载,B max表示整个网络的最大使用链路带宽,R total表示每个MEC服务器的总容量,B total表示每条光纤链路的总容量。
进一步的,ρ值越小,代表发生服务器负载失衡和链路流量超负荷的可能性越低,更容易实现负载均衡,网络的整体性能也就越好。负载均衡保护策略模块为用户的每一个业务提供保护时,同时尝试路径保护策略模块和服务器保护策略模块,找到各自对应的ρ值,ρ值在应用策略模块后更新,最终应用ρ值 最小的策略为用户提供保护。所述负载均衡保护策略模块还包括:判断网络中的所有业务是否都被保护:若都被保护,则检测网络中所有用户的全部业务是否被保护,否则继续通过负载均衡保护策略模块保护未被保护的业务。检测网络中所有用户的全部业务是否被保护:若都被保护,则终止进程,并找到最终网络资源利用率,否则继续通过负载均衡保护策略模块保护下一个用户。
以上所述实施例仅是为充分说明本发明而所举的较佳的实施例,本发明的保护范围不限于此。本技术领域的技术人员在本发明基础上所作的等同替代或变换,均在本发明的保护范围之内。本发明的保护范围以权利要求书为准。

Claims (10)

  1. 一种基于边缘计算的FiWi网络负载均衡方法,其特征在于:包括以下步骤:
    建立路径保护策略:对于用户业务,寻找与其本地MEC服务器之间满足通信时延要求的候选保护路径,并在满足业务带宽的候选保护路径中选择延迟最低的路径为用户业务提供服务;
    建立服务器保护策略:对于用户业务,将所有满足通信时延要求的非本地MEC服务器按照负载升序排序,删除无法提供足够MEC资源的MEC服务器后选择负载最小的MEC服务器为用户业务提供服务;
    定义网络资源利用率ρ:
    ρ=R max/R total+B max/B total
    其中,R max表示整个网络的最大MEC服务器负载,B max表示整个网络的最大使用链路带宽,R total表示每个MEC服务器的总容量,B total表示每条光纤链路的总容量;
    建立负载均衡保护策略:为用户业务提供保护时,分别计算并比较该业务进行路径保护策略和服务器保护策略的网络资源利用率,取其中网络资源利用率较小的策略为该用户业务提供服务。
  2. 如权利要求1所述的一种基于边缘计算的FiWi网络负载均衡方法,其特征在于:建立路径保护策略具体包括以下步骤:
    对于用户业务,采用最短路由算法寻找k条与其本地MEC服务器之间满足通信时延要求的候选保护路径,并保存在候选路径列表中;
    判断所有候选路径是否具有足够的带宽,在候选路径列表中删除无法提供 足够带宽的路径;
    选择候选路径列表中延迟最低的路径作为保护路径向用户业务提供服务。
  3. 如权利要求1所述的一种基于边缘计算的FiWi网络负载均衡方法,其特征在于:建立服务器保护策略具体包括以下步骤:
    对于用户业务,将所有满足通信时延要求的非本地MEC服务器按照负载升序排序,并保存在列表中;
    从列表中删除无法提供足够MEC资源的MEC服务器;
    选择列表中第一个服务器为用户业务提供服务。
  4. 如权利要求1所述的一种基于边缘计算的FiWi网络负载均衡方法,其特征在于:建立负载均衡保护策略具体包括以下步骤:
    a、对于网络拓扑G=(N,L),用户集合C,用户c∈C的业务集合S c
    b、获得一个业务s∈S c
    c、对业务s尝试路径保护策略并得到相应的网络资源利用率ρ值,记为ρ PP,s;对业务s尝试服务器保护策略并得到相应的网络资源利用率ρ值,记为ρ SP,s;若ρ PP,s<ρ SP,s,对业务s使用路径保护策略并记录ρ PP,s,否则使用服务器保护策略并记录ρ SP,s
  5. 如权利要求4所述的一种基于边缘计算的FiWi网络负载均衡方法,其特征在于:建立负载均衡保护策略还包括以下步骤:
    d、判断S c中的所有业务是否都被保护:若都被保护,则进行下一步骤,否则返回步骤b保护下一个业务;
    e、判断C中的所有用户的全部业务是否都被保护;若都被保护,则终止进 程,并找到最终网络资源利用率,否则返回步骤a保护下一个用户。
  6. 如权利要求4所述的一种基于边缘计算的FiWi网络负载均衡方法,其特征在于:所述步骤c中ρ值在应用路径保护策略或服务器保护策略后更新,最终应用ρ最小的策略为用户业务提供保护。。
  7. 一种基于边缘计算的FiWi网络负载均衡系统,其特征在于:包括:
    路径保护策略模块,对于用户业务,寻找与其本地MEC服务器之间满足通信时延要求的候选保护路径,并在满足业务带宽的候选保护路径中选择延迟最低的路径为用户业务提供服务;
    服务器保护策略模块,对于用户业务,将所有满足通信时延要求的非本地MEC服务器按照负载升序排序,删除无法提供足够MEC资源的MEC服务器后选择负载最小的MEC服务器为用户业务提供服务;
    负载均衡保护策略模块,为用户业务提供保护时,分别计算并比较该业务选择路径保护策略模块和服务器保护策略模块的网络资源利用率,取其中网络资源利用率较小的策略模块为该用户业务提供服务;。
    其中,网络资源利用率ρ:
    ρ=R max/R total+B max/B total
    R max表示整个网络的最大MEC服务器负载,B max表示整个网络的最大使用链路带宽,R total表示每个MEC服务器的总容量,B total表示每条光纤链路的总容量。
  8. 如权利要求7所述的一种基于边缘计算的FiWi网络负载均衡系统,其特征在于:所述负载均衡保护策略模块为用户的每一个业务提供保护时,同时尝试路径保护策略模块和服务器保护策略模块,找到各自对应的ρ值,ρ值在应用策略模块后更新,最终应用ρ值最小的策略为用户提供保护。
  9. 如权利要求8所述的一种基于边缘计算的FiWi网络负载均衡方法,其特征在于:所述负载均衡保护策略模块还包括:判断网络中的所有业务是否都被保护:若都被保护,则检测网络中所有用户的全部业务是否被保护,否则继续通过负载均衡保护策略模块保护未被保护的业务。
  10. 如权利要求9所述的一种基于边缘计算的FiWi网络负载均衡方法,其特征在于:检测网络中所有用户的全部业务是否被保护:若都被保护,则终止进程,并找到最终网络资源利用率,否则继续通过负载均衡保护策略模块保护下一个用户。
PCT/CN2021/123240 2021-09-28 2021-10-12 一种基于边缘计算的FiWi网络负载均衡方法及系统 WO2023050469A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US17/911,390 US12160466B2 (en) 2021-09-28 2021-10-12 Fiwi network load balancing method and system based on edge computing

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111143910.1A CN113938953B (zh) 2021-09-28 2021-09-28 一种基于边缘计算的FiWi网络负载均衡方法及系统
CN202111143910.1 2021-09-28

Publications (1)

Publication Number Publication Date
WO2023050469A1 true WO2023050469A1 (zh) 2023-04-06

Family

ID=79277269

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/123240 WO2023050469A1 (zh) 2021-09-28 2021-10-12 一种基于边缘计算的FiWi网络负载均衡方法及系统

Country Status (3)

Country Link
US (1) US12160466B2 (zh)
CN (1) CN113938953B (zh)
WO (1) WO2023050469A1 (zh)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114629997A (zh) * 2022-03-16 2022-06-14 上海小度技术有限公司 资源分配方法、装置、设备、介质及程序产品
CN119155233B (zh) * 2024-11-21 2025-02-14 长春工程学院 基于延迟与带宽中继比较算子的网络通信路径选择方法

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104283804A (zh) * 2014-10-27 2015-01-14 杭州华三通信技术有限公司 一种链路负载均衡方法和装置
US10104039B1 (en) * 2017-09-28 2018-10-16 Cloudflare, Inc. Establishing and using a tunnel from an origin server in a distributed edge compute and routing service
CN110300024A (zh) * 2019-06-28 2019-10-01 中天宽带技术有限公司 一种服务器任务处理方法、装置及其相关设备
CN111770477A (zh) * 2020-06-08 2020-10-13 中天通信技术有限公司 一种mec网络的保护资源的部署方法及相关装置
CN113179331A (zh) * 2021-06-11 2021-07-27 苏州大学 面向移动边缘计算的分布式专用保护业务调度方法

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2526300A (en) * 2014-05-20 2015-11-25 Ibm Partitioning of a network using multiple poles for each part thereof
CN107846445A (zh) * 2017-09-20 2018-03-27 南京邮电大学 一种基于sdn的服务器负载均衡方法

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104283804A (zh) * 2014-10-27 2015-01-14 杭州华三通信技术有限公司 一种链路负载均衡方法和装置
US10104039B1 (en) * 2017-09-28 2018-10-16 Cloudflare, Inc. Establishing and using a tunnel from an origin server in a distributed edge compute and routing service
CN110300024A (zh) * 2019-06-28 2019-10-01 中天宽带技术有限公司 一种服务器任务处理方法、装置及其相关设备
CN111770477A (zh) * 2020-06-08 2020-10-13 中天通信技术有限公司 一种mec网络的保护资源的部署方法及相关装置
CN113179331A (zh) * 2021-06-11 2021-07-27 苏州大学 面向移动边缘计算的分布式专用保护业务调度方法

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
WANG NING; SHAO WEIDONG; BOSE SANJAY K.; SHEN GANGXIANG: "MixCo: Optimal Cooperative Caching for Mobile Edge Computing in Fiber-Wireless Access Networks", 2018 OPTICAL FIBER COMMUNICATIONS CONFERENCE AND EXPOSITION (OFC), OSA, 11 March 2018 (2018-03-11), pages 1 - 3, XP033357749 *
WANG TONG: "Protection Strategies for Mobile Edge Computing Network", CHINESE MASTER’S THESES FULL-TEXT DATABASE, INFORMATION SCIENCE AND TECHNOLOGY, no. 2, 15 February 2021 (2021-02-15), XP093014492 *

Also Published As

Publication number Publication date
US20240223648A1 (en) 2024-07-04
CN113938953B (zh) 2022-07-12
US12160466B2 (en) 2024-12-03
CN113938953A (zh) 2022-01-14

Similar Documents

Publication Publication Date Title
CN111901424A (zh) 云边协同网络资源平滑迁移与重构方法及系统
CN107666412A (zh) 服务功能链的虚拟网络功能部署方法
US10404546B2 (en) Multi-tier fault tolerant network design with quality of service considerations
CN111526057B (zh) 一种基于业务类型的网络切片可靠性映射方法
US9820021B2 (en) Methods and apparatus for determining a path in a communications network
CN113938953B (zh) 一种基于边缘计算的FiWi网络负载均衡方法及系统
CN114095075A (zh) 一种基于业务定制需求感知的光网络故障恢复方法
WO2020158016A1 (ja) バックアップシステム及びその方法並びにプログラム
Yi et al. Energy‐aware disaster backup among cloud datacenters using multiobjective reinforcement learning in software defined network
CN113193996A (zh) 一种电力光传输网优化方法、装置、设备及存储介质
CN114615184A (zh) 一种电力业务的双路由规划方法、装置及系统
Jamali et al. Survivability evaluation for networks carrying complex traffic flows
CN101155063A (zh) 一种业务保护方法以及路由设备
Han et al. Survivable virtual network link shared protection method based on maximum spanning tree
CN102647424B (zh) 一种数据传输方法及其装置
CN114466059B (zh) 一种向移动边缘计算系统提供可靠服务功能链的方法
CN110062404B (zh) 核心网中具备单链路故障恢复的虚拟链路映射方法
Liu et al. Fault recovery of classified services based on hybrid protection and restoration in F5G
Tivig et al. Critical analysis of multi-controller placement problem in large SDN networks
CN115277430A (zh) 一种光网络中链路故障概率量化方法及sdn控制器部署方法
Ramesh et al. Energy—Efficient resource allocation in WDM networks
Markidis et al. Routing and wavelength assignment algorithms in survivable WDM networks under physical layer constraints
Li et al. Reconfigurable Bandwidth Service based on optical network state for inter-data center communication
CN109831380A (zh) 一种电力通信业务的路径保护方法
CN118337694B (zh) 一种基于服务功能链的路由优化方法和系统

Legal Events

Date Code Title Description
WWE Wipo information: entry into national phase

Ref document number: 17911390

Country of ref document: US

121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21959033

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21959033

Country of ref document: EP

Kind code of ref document: A1