CN107332913B - Optimized deployment method of service function chain in 5G mobile network - Google Patents

Optimized deployment method of service function chain in 5G mobile network Download PDF

Info

Publication number
CN107332913B
CN107332913B CN201710536013.4A CN201710536013A CN107332913B CN 107332913 B CN107332913 B CN 107332913B CN 201710536013 A CN201710536013 A CN 201710536013A CN 107332913 B CN107332913 B CN 107332913B
Authority
CN
China
Prior art keywords
vnf
mapping
server
cost
link
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201710536013.4A
Other languages
Chinese (zh)
Other versions
CN107332913A (en
Inventor
孙罡
赵东成
廖丹
孙健
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.)
University of Electronic Science and Technology of China
Original Assignee
University of Electronic Science and Technology of China
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 University of Electronic Science and Technology of China filed Critical University of Electronic Science and Technology of China
Priority to CN201710536013.4A priority Critical patent/CN107332913B/en
Publication of CN107332913A publication Critical patent/CN107332913A/en
Application granted granted Critical
Publication of CN107332913B publication Critical patent/CN107332913B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/01—Protocols
    • H04L67/10—Protocols in which an application is distributed across nodes in the network
    • H04L67/1001—Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00—Network arrangements or protocols for supporting network services or applications
    • H04L67/50—Network services
    • H04L67/51—Discovery or management thereof, e.g. service location protocol [SLP] or web services
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W28/00—Network traffic management; Network resource management
    • H04W28/02—Traffic management, e.g. flow control or congestion control
    • H04W28/0252—Traffic management, e.g. flow control or congestion control per individual bearer or channel
    • H04W28/0263—Traffic management, e.g. flow control or congestion control per individual bearer or channel involving mapping traffic to individual bearers or channels, e.g. traffic flow template [TFT]
    • 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
    • G06F2009/45595—Network integration; Enabling network access in virtual machine instances

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Telephonic Communication Services (AREA)

Abstract

The invention discloses an optimized deployment method of a service function chain in a 5G mobile network, belonging to the field of mobile communication. When server resources and bandwidth resources of an underlying network are allocated to each dynamically-arriving service function chain request, three different mapping schemes are provided by considering a deployment mode of a VNF merging strategy, a virtual reuse strategy and a temporary link strategy during deployment: minimizing computational resource cost: limiting the maximum mapping cost which can be accepted by a service provider for each virtual link during deployment; path length of the shortest service function chain: limiting the maximum mapping cost that each virtual network function can be accepted by a service provider during deployment; and computing resource cost and link resource cost by using the sex warfare game theory model for fair optimization. The three implementation schemes of the invention can improve the mapping success rate of the service function chain request and the resource utilization rate of the underlying network, and simultaneously can minimize the total mapping cost as much as possible.

Description

Optimized deployment method of service function chain in 5G mobile network
Technical Field
The invention belongs to the field of mobile communication, and particularly relates to optimized deployment of a service function chain in a 5G mobile network.
Background
With the great increase of wireless traffic and services, the fourth generation (4G) will not meet the future network requirements, so that the research of the fifth generation (5G) mobile wireless network is promoted to cope with the huge challenge of diversified differentiated services in the future.
With the explosive growth of mobile wireless traffic, mobile operators are considering expanding their services using cloud computing and dealing with the tremendous growth of mobile data traffic. In conventional telecommunications networks, network functions or middleboxes (such as packet data network gateways, service gateways, firewalls, content filters, proxy servers, wide area network optimizers, intrusion detection systems and intrusion prevention systems) are implemented by specialized physical devices and facilities. However, as the demand for more diverse and new services increases, service providers must correspondingly purchase, store, and operate new physical devices to meet the demands of users. However, the procurement of new physical equipment will result in higher capital and operational expenditures. And these physical devices require specially trained personnel for deployment and maintenance.
To address the above difficulties, researchers have proposed Network Function Virtualization (NFV) aimed at mapping the processing of packets from a hardware middlebox onto a software middlebox running on commercial hardware. The network function running on the software middlebox is called a Virtual Network Function (VNF). In network function virtualization, multiple virtual network functions are typically connected in a particular order to form a service function chain to provide different network services. For example, in a mobile network, communication between a mobile network user and a service terminal needs to be performed through a service function chain: user → service gateway → packet data network gateway → firewall → intrusion detection system → agent → terminal, this service function chain typically deploys security policies that perform traffic filtering between the user and the service terminal. Generally, the type and order of the individual virtual network functions in the service function chain are determined according to traffic classification, service level protocol, provisioning policy of the operator, and the like.
In fact, core network virtualization and network function virtualization represent two key vision for future 5G architectures. As a key technology of 5G, network function virtualization has become an important direction for the evolution of wireless network architecture. As an emerging technology, network function virtualization has received a great deal of attention from the industry, academia, and standardization bodies. For service providers, efficient deployment/mapping of service function chains to 5G mobile networks is crucial. Currently, the placement of virtual network functions has also become a research focus, and there have been some researches on the problem of virtual network function placement.
Currently, a placement/deployment scheme related to virtual network function or Service Function Chain (SFC) requests has been proposed, but most of the proposed deployment approaches are not suitable for 5G mobile networks because the processed object is a virtual network or a federated cloud. For example, the said constrained NFV Location algorithm, the main idea is to reduce the overall network cost as much as possible when placing network functions, and at the same time, satisfy the size constraint of network nodes. Although this approach enables placement of virtual network functions, it is proposed for virtual networks or federated clouds, without considering the characteristics of 5G networks and related constraints, and therefore this approach is not suitable for 5G mobile networks. For those virtual network function or service function chain placement/deployment schemes suitable for the 5G mobile network, only the deployment problem of the virtual network functions (such as packet data network gateway and service gateway) of the radio access network is considered, and the deployment problem of the virtual network functions (such as firewall, content filter, proxy server, wide area network optimizer, intrusion detection system and intrusion prevention system) of the core network and the data center network is not considered. For example, by trading off the number of service gateways to be relocated and the large path length through a bidding nash theory, although the placement of virtual network functions in a 5G mobile network can be realized, only the problem of the deployment of the virtual network functions of a wireless access network is considered, but the problem of the deployment of the virtual network functions of a core network and a data center network is not considered by the service gateways.
Disclosure of Invention
The technical problem to be solved by the present invention is to provide a deployment method for placing service function chain requests, which aims at consuming the least server resources, bandwidth resources and reducing the blocking rate of the service function chain requests, under the precondition that the 5G mobile network (underlying network) and the on-line service function chain requests, the positions of mobile network users and the positions of service terminals are known, and under the consideration of each virtual network function and link connection condition in each service function chain request and the satisfaction of related constraint conditions. The deployment scheme of the invention comprehensively considers the specificity of the service function chain request, not only makes optimized configuration for common bandwidth resource requirements and server resource requirements, but also provides a corresponding solution strategy aiming at the strict requirement of the service function chain request in the aspect of communication delay.
In the mapping process of the service function chain request, three effective strategies are introduced to effectively map/deploy the service function chain request to the 5G mobile network so as to reduce the cost of computing resources and link resources, thereby improving the receiving rate of the service function chain request.
(1) Virtual machine reuse policies.
In the invention, in order to improve the resource utilization rate of the server, a virtual machine reuse strategy is introduced, namely, when mapping one Virtual Network Function (VNF), the existing/existing virtual machine which realizes the same VNF can be reused. The virtual machine reuse strategy can not only improve the resource utilization rate of the server, but also improve the receiving rate of the service function chain request when the server resource is limited. In order to improve the resource utilization rate of a server and reduce the cost of computing resources, when a VNF is mapped to the server, if the virtual machine reuse is considered, firstly, whether a reusable virtual machine exists in the server where the current VNF is placed is judged, and if the reusable virtual machine exists, the current VNF is hosted to the reusable virtual machine on the server; otherwise, the current VNF is hosted on the new virtual machine on this server. The reusable virtual machines are as follows: having hosted at least one and a VNF with a reuse identifier (indicating that the current virtual machine is allowed to be reused)iVirtual machines of the same type of VNF, and hosted VNF and VNFiNot belonging to the same service function chain request.
When the VNF is hosted on a reusable virtual machine, the present invention does not consider the cost of computing resources during the original service time of the virtual machine, i.e., the cost of computing resources is zero, but when the service time of the VNF exceeds the original service time of the virtual machine, the cost of computing resources for additional time needs to be computed. When a new virtual machine is used by the VNF, the present invention requires computing the cost of computing resources for the entire service time of the VNF. Thus, the computation of the computational resource cost of any VNF requested for one service function chain is as follows:
Cost(VNFi→nk)＝p(nk)×ε(VNFi)×Ti p
Figure BDA0001340546180000031
among them, Cost (VNF)i→nk) Representing the ith VNF (VNF)i) Is mapped to server nkThe cost of computing resources of p (n)k) Presentation Server nkThe unit cost of resources in (1), i.e. the unit cost of objects in parentheses, ε (VNF) is denoted by the symbol p (-)i) Representing VNFiResource constraints on the server, such as CPU, memory and storage capacity. N is a radical ofV＝{VNF1,VNF2,…,VNFnDenotes the set of virtual network functions in the service function chain request, the index n denotes the number of VNFs included in one service function chain request, Ti pRepresenting VNFiTime of need to pay, TiRepresenting VNFiService time of, ToRepresenting the original service time, pi, of the reused virtual machineiRepresenting the number of VNFs, π, hosted on a virtual machinei1 means that the current virtual machine is not reused, pii> 1 means reused by multiple VNFs. Reusing existing virtual opportunities affects the performance of other VNFs, therefore δ is defined as the maximum number of VNFs hosted on the existing virtual machine, i.e.:
Figure BDA0001340546180000032
(2) VNF merge policies.
In the present invention, when mapping VNFsiIf the VNF is hostedi-1Has sufficient available resources, the mapping VNF may be considerediTo hosting VNFi-1The present invention refers to this policy as VNF merge policy. In the present invention, rather than merging some of the VNFs before mapping this service function chain request, some VNFs are mergedIt is during the mapping process that the ith VNF is mapped to the server hosting the (i-1) th VNF, if this server has sufficient available resources. In the VNF merge policy, the ith VNF can only be mapped to the server hosting the i-1 st VNF, but not to the servers of the other VNFs hosting this service function chain request, thereby avoiding ping-pong routing problems. For example, a VNF when servicing function chain requests1Is mapped to physical node B, VNF2After being mapped to physical node F, VNF is mapped3It can be mapped to physical node F, if there are sufficient resources available, i.e. allowing VNF to be mapped to physical node F2And VNF3Are merged together such that VNF2And VNF3The communication does not need to consume bandwidth resources because of the VNF2And VNF3Communication takes place within the physical node F. However, mapping VNFs is not allowed3To the physical node B, which is undesirable because it can result in ping-pong routing.
(3) A temporary link mapping policy.
In the present invention, when mapping the ith VNF, the ith link e connecting the ith VNF and the (i-1) th VNF needs to be mapped at the same timei(virtual link) to ensure that a near optimal mapping scheme is obtained at the ith VNF. The traditional approach is to work on finding a locally optimal mapping solution for the ith VNF, but this does not guarantee a near optimal path for the entire service function chain. In order to ensure an approximately optimal path and improve the acceptance rate of the whole service function chain request, the virtual link e is calculated by the methodiWhen the cost of the link resources is high, a temporary link te connecting the ith VNF and the service terminal is generatedi(virtual link) the bandwidth requirement of this temporary link is equal to the bandwidth requirement of the (i + 1) th link, and this temporary link te is mappediAnd obtaining the mapping path by the bottom layer network. When mapping the ith VNF and the virtual link eiThen, by using the temporary link mapping policy, a near-optimal mapping scheme of the ith VNF is found. In the temporary link mapping policy, the temporary link does not need to consume actual link resources, it is only used to constrain the ith VNF not to deviate too far from the serving terminal,so as to ensure that the adjacent link of the physical server hosting the ith VNF has enough link resources to map the next virtual link, thereby improving the service function chain acceptance rate. The link resource cost of the ith virtual link is calculated as follows:
Figure BDA0001340546180000041
wherein,
Figure BDA0001340546180000042
representing a connection VNFiAnd VNFi-1The ith virtual link (e)i) Link resource cost of peiRepresenting a virtual link eiMapping paths in the underlying network (underlying paths), pteiIndicates a temporary link teiMapping path in underlying network (underlying path), esPhysical links, x, representing underlying networkiRepresenting a virtual link eiResource constraints such as bandwidth, etc., and xi+1Then it means that the second VNF is connectedi+1And VNFiResource constraints of the (i + 1) th virtual link. In the invention, the default virtual link connecting the first VNF and the user is a virtual link e1. The mapping of service function link requests can be divided into two parts. The first part is the VNF that places and allocates resources to service function chain requests. The second part is the virtual link that maps and allocates bandwidth resources to service function chain requests. The mapping process for the service function chain is described as follows.
(1) VNF mapping:
the VNF mapping process may be represented as:
Figure BDA0001340546180000043
Figure BDA0001340546180000044
Figure BDA0001340546180000051
Figure BDA0001340546180000052
Figure BDA0001340546180000053
Figure BDA0001340546180000054
Figure BDA0001340546180000055
wherein N isS1Set of servers and routers representing the underlying network assigned to the current service function chain request, CN1Representing the server resource allocated to the current service function chain request, MN＝{M(VNF1),M(VNF2),...,M(VNFn) Denotes a mapping record for each VNF requested by the current service function chain. M (VNF)i) Representing hosted VNFsiServer of (1), R (M (VNF)i) Represents the server M (VNF)i) The available resources of (1). CN＝{ε(VNF1),ε(VNF2),...,ε(VNFn) Denotes the resource constraint set, VM, of all virtual network functionsiRepresenting hosted VNFsiOf (existing) epsilon (VM)i) Represents the computational resources of this existing virtual machine, Y ∈ {0,1, 2.. and Y } represents the number of the network region, L (M (VNF)i) Represents the server M (VNF)i) The number of network areas where, and a server can only belong to one network area,
Figure BDA0001340546180000056
representing VNFiCan be mapped to this network area and,
Figure BDA0001340546180000057
representing VNFiCannot be mapped to this network area,
Figure BDA0001340546180000058
representation server M (VNF)i) Satisfy VNFiA position constraint of (2); if it is
Figure BDA0001340546180000059
It is not satisfied. In a 5G mobile network, a serving gateway and a packet data gateway belong to functions of a radio access network, and they are generally deployed only in the radio access network, whereas a firewall, a content filter, a proxy server, a wide area network optimizer, an intrusion detection system, and an intrusion prevention system are functions of a data center network or a core network, and they are generally deployed only in the core network and the data center network.
(2) Link mapping of service function chains:
the link mapping of the service function chain is described as follows:
Figure BDA00013405461800000510
Figure BDA00013405461800000511
Figure BDA00013405461800000512
Figure BDA00013405461800000513
wherein M isE＝{M(e1),M(e2),...,M(e|Ev|) Denotes a mapping record for each virtual link requested by the current service function chain,
Figure BDA0001340546180000061
set of virtual links, | E, representing current service function chain requestsvI represents the set EVI.e. the number of virtual links requested by the current service function chain.
Figure BDA00013405461800000618
A set of resource constraints representing all virtual links requested by the current serving function chain. P1Represents the end-to-end set of underlying paths to which the current service function chain request is mapped, and P1Each underlying path of
Figure BDA0001340546180000062
Set of physical links E being an underlying networkSA subset of (2). CE1Indicating the link resources allocated to this service function chain request.
Figure BDA0001340546180000063
Representing underlying paths
Figure BDA0001340546180000064
Available bandwidth resources of b (e)s) Representing a physical link esThe available bandwidth resources of (a) may be,
Figure BDA0001340546180000065
representing underlying paths
Figure BDA0001340546180000066
Path delay of d (e)s) Representing a physical link esTime delay of (2).
Therefore, each service function chain requires deployment issues in 5G mobile networks, I) link resource cost minimization; II) cost minimization of computational resources, which can be described according to the following linear program (1):
Figure BDA0001340546180000067
Figure BDA0001340546180000068
s.t.
Figure BDA0001340546180000069
Figure BDA00013405461800000610
Figure BDA00013405461800000611
Figure BDA00013405461800000612
Figure BDA00013405461800000613
Figure BDA00013405461800000614
Figure BDA00013405461800000615
Figure BDA00013405461800000616
Figure BDA00013405461800000617
the first objective is to reduce the cost of computing resources as much as possible. This will increase the probability that the service function chain has a longer path. The second objective is to minimize the cost of link resources, i.e., shorten the path of the entire service function chain as much as possible. At the same time, the constraints in the linear programming (1) are used to ensure the following constraints:
constraint 1 is used to ensure that the number of VNFs hosted on a virtual machine does not exceed the number δ given by the service provider.
Constraint 2 gives the time the VNF needs to pay.
Constraints 3 and 4 ensure that the servers being used meet the computational resource requirements of the VNF.
Constraints 5 and 6 ensure that the physical links used satisfy the constraints of the virtual links.
Constraints 7, 8 and 9 ensure that the servers used meet the location constraints of the VNF.
Because the linear programming (1) is a multi-objective problem and cannot be directly solved, the invention provides three solutions to solve the multi-objective problem (1). The first proposed solution is: the cost of computing resources is reduced to the maximum extent; the second solution is: the path of the whole service function chain is shortened; the third solution is: a fair solution is found for computing resource costs and link resource costs by allocating resources and routes to the VNF using a sex warfare Game (BOS) model. The three solutions described above are described in detail as follows:
(1) the cost of computing resources is minimized (MC scheme for short).
In this solution, definitions are made
Figure BDA0001340546180000071
Maximum mapping cost that can be accepted by the service provider for each virtual link in the service function chain request, i.e.
Figure BDA0001340546180000072
Given by the service provider, it typically does not exceed the charge per virtual link. This optimization model, with the goal of reducing computational resource costs, can be described as a linear program (2) as follows:
Figure BDA0001340546180000073
s.t.
Figure BDA0001340546180000074
Figure BDA0001340546180000075
Figure BDA0001340546180000076
Figure BDA0001340546180000077
Figure BDA0001340546180000078
Figure BDA0001340546180000079
Figure BDA00013405461800000710
Figure BDA00013405461800000711
Figure BDA00013405461800000712
Figure BDA00013405461800000713
that is, when server resources and broadband resources of the underlying network are allocated to each dynamically-arriving service function chain request, an optimal deployment scheme of the current service function chain request is obtained through the following steps:
step 1: virtual network function set N requested from service function chain to be mappedV＝{VNF1,VNF2,...,VNFnThe first VNF of the vns starts, in turn, for each VNF requested by the current service function chaini(i ═ 1, …, n) determines a set of alternative mapping schemes:
(1) determining a VNF currently to be mappediAlternative server set of (2):
set of available servers U from the underlying networkSIn (1), the position constraint and calculation will be satisfiedResource requirements have not been aggregated
Figure BDA0001340546180000081
The server selected by each VNF in (1) is used as the VNFiAlternative server set of (3), i.e. VNFiAlternative servers that may be selected include: servers, VNFs, not selected by other VNFsi-1Selected servers (VNF merge policy), hence
Figure BDA0001340546180000082
Can also be expressed as:
Figure BDA0001340546180000083
(2) determining VNFiAlternative mapping scheme set of (2):
with M (VNF)i) Presentation placement VNFiAnd determines server M (VNF)i) Upper for hosting VNFiThe virtual machine of (2): judging whether a reusable virtual machine exists or not, and if so, enabling the VNFiHosted on a reusable virtual machine; otherwise VNF will beiHosting on a new virtual machine;
under the condition of satisfying ei(connection VNFi-1And VNFiVirtual link of) of a wireless networkiAnd on the premise of time delay requirement, for eiMapping the bottom layer path to obtain eiMapping path of
Figure BDA0001340546180000084
Wherein VNF0Representing a user;
generating a connection VNFiTemporary link te with user terminaliAnd e is combinedi+1(connection VNFiAnd VNFi+1Virtual link of) of a wireless networki+1As teiLink resource requirements of (1); at the time of satisfying teiOn the premise of the link resource demand, for teiMapping the bottom layer path to obtain teiMapping path of
Figure BDA0001340546180000085
I.e. constraining the placement of VNFs by temporary link mapping policiesiDoes not deviate too far from the user terminal;
wherein the mapping path
Figure BDA0001340546180000086
Mapping paths
Figure BDA0001340546180000087
There may be multiple strips.
Will satisfy the conditions
Figure BDA0001340546180000088
Shortest mapping path of
Figure BDA0001340546180000089
As corresponding to current M (VNF)i) E ofiTo ensure pair eiThe link resource cost corresponding to the bottom layer path mapping does not exceed
Figure BDA00013405461800000810
Will correspond to current M (VNF)i) The alternative mapping scheme of (2) is saved into an alternative mapping scheme set, wherein the alternative mapping scheme comprises: m (VNF)i) Hosting VNFiVirtual machine of, eiThe final mapping path of (2);
the combination of different alternative mapping schemes of all VNFs requested by the current service function chain may result in different mapping sets MN；
Finally, according to the formula
Figure BDA0001340546180000091
And respectively calculating the total calculation resource cost of each mapping set, and obtaining the optimal deployment scheme of the current service function chain request from the mapping set corresponding to the minimum total calculation resource cost.
(2) The path length of the service function chain is minimized (SL scheme for short).
At the solution sideIn the scheme, define
Figure BDA0001340546180000092
For the maximum mapping cost that each VNF in a service function chain request can be accepted by the service provider, i.e., the maximum mapping cost
Figure BDA0001340546180000093
Given by the service provider, it typically does not exceed the per VNF charge. This optimization model, aiming at minimizing the path length of the service function chain, can be described as a linear plan (3) as follows:
Figure BDA0001340546180000094
s.t.
Figure BDA0001340546180000095
Figure BDA0001340546180000096
Figure BDA0001340546180000097
Figure BDA0001340546180000098
Figure BDA0001340546180000099
Figure BDA00013405461800000910
Figure BDA00013405461800000911
Figure BDA00013405461800000912
Figure BDA00013405461800000913
Figure BDA00013405461800000914
that is, when server resources and broadband resources of the underlying network are allocated to each dynamically-arriving service function chain request, an optimal deployment scheme of the current service function chain request is obtained through the following steps:
step 1: virtual network function set N requested from service function chain to be mappedV＝{VNF1,VNF2,...,VNFnThe first VNF of the vns starts, in turn, for each VNF requested by the current service function chaini(i ═ 1, …, n) determines a set of alternative mapping schemes:
101: from the set of available servers in the underlay network, the location constraints, computing resource requirements will be met and never aggregated
Figure BDA00013405461800000915
The server selected by each VNF in (1) is used as the VNFiInitial set of alternative servers Ui′；
Separately determining a server on each initial candidate server for hosting a VNFiThe virtual machine of (2): judging whether a reusable virtual machine exists or not, and if so, enabling the VNFiHosted on a reusable virtual machine; otherwise VNF will beiHosting on a new virtual machine;
for each initial alternative server nk∈Ui', respectively determined thereon for hosting a VNFiThe virtual machine of (2): judging whether a reusable virtual machine exists or not, and if so, enabling the VNFiHosted on a reusable virtual machine; otherwise VNF will beiHosted on a new virtual machine. And calculates the resource Cost (VNF)i→nk) And
Figure BDA0001340546180000101
compared with, will be less than or equal to
Figure BDA0001340546180000102
Initial alternative server nkAs VNFiAlternative server n ofmAnd recording the alternative server nmUpper hosting VNFiThe virtual machine of (1); by all alternative servers nmObtaining VNFiAlternative server set U ofi。
102: determining VNFiAlternative mapping scheme set of (2):
for each alternative server nm∈UiIn the presence of a catalyst satisfying eiLink resource requirement xiAnd on the premise of time delay requirement, the slave server nmTo VNFi-1In the bottom layer path of the alternative server, a shortest path is searched as eiMapping path of
Figure BDA0001340546180000103
Wherein VNF0Representing users, i.e. to the same server nmIn particular, VNFi-1How many alternative servers exist, then how many of them exist with respect to eiMapping path of
Figure BDA0001340546180000104
Generating a connection VNFiTemporary link te with user terminaliAnd e is combinedi+1Link resource requirement xi+1As teiLink resource requirements of (1); at the time of satisfying teiOn the premise of link resource demand of (2), the slave server nmFinding a shortest path as te in the bottom path to the user terminaliMapping path of
Figure BDA0001340546180000105
Will correspond to the current nmTo the VNFiWherein the alternative mapping scheme comprises: n ismHosting VNFiVirtual machine, mapping path
Figure BDA0001340546180000106
Step 2: obtaining different mapping sets by the combination of different alternative mapping schemes of all VNFs requested by the current service function chain;
according to the formula
Figure BDA0001340546180000107
And respectively calculating the total link resource cost corresponding to each mapping set, and obtaining the optimized deployment scheme of the current service function chain request by the mapping set corresponding to the minimum total link resource cost.
(3) And (3) computing resource cost and link resource cost (FOCL scheme for short) by using the sex warfare game theory model for fair optimization.
The gender wars game theory model describes such a game scenario: in a game, two players have some common interests, but the common interests have different outcomes and conflicting preferences. For example, the couple would prefer to watch the same television program but would not want to watch their respective television program separately, and the couple would prefer to watch their favorite program.
Thus, in the present invention, computing resource costs and link resource costs are taken as two players in the sexually significant game theory model, which is based on two strategies: I) reuse an existing virtual machine when mapping the VNF, II) use a new virtual machine when mapping the VNF. An existing virtual machine is used to map the VNF, which may reduce the cost of computing resources, but which may result in a longer path for service function chains. A new virtual machine is used to map the VNF, which makes it easier to find a near-optimal path of the service function chain, but which may result in higher cost of computing resources. The mapping process for each VNF is a gaming process where two players, computing resource costs and link resource costs, are played to determine whether to use an existing virtual machine. In the present invention, an existing virtual machine will be reused as the first strategy (i.e., 1-S), a new virtual machine will be used as the second strategy (i.e., 2-S), i.e., computing resource costs as the first player (i.e., 1-P), and link resource costs as the second player (i.e., 2-P). This gaming strategy is represented in table 1.
TABLE 1 Game policy
Figure BDA0001340546180000111
The expressions and parameters referred to in table 1 are annotated as follows:
Figure BDA0001340546180000112
mapping a VNF when using an existing virtual machineiComputing resource revenue in time;
Figure BDA0001340546180000113
mapping a VNF when using an existing virtual machineiLink resource revenue in time;
Cost(ME(VNFi)): mapping a VNF when using an existing virtual machineiA computing resource cost of time;
Cost(pE(ei)): mapping a VNF when using an existing virtual machineiLink resource cost of time;
ME(VNFi): mapping a VNF when using an existing virtual machineiTime VNFiThe mapping scheme of (1);
pE(ei): mapping a VNF when using an existing virtual machineiTime virtual link eiThe mapping path of (2);
Figure BDA0001340546180000114
mapping VNFs when using a new virtual machineiComputing resource revenue in time;
Figure BDA0001340546180000115
mapping VNFs when using a new virtual machineiLink resource revenue in time;
Cost(MN(VNFi)): mapping VNFs when using a new virtual machineiA computing resource cost of time;
Cost(pN(ei)): mapping VNFs when using a new virtual machineiLink resource cost of time;
MN(VNFi): mapping VNFs when using a new virtual machineiTime VNFiThe mapping scheme of (1);
pN(ei): mapping VNFs when using a new virtual machineiTime virtual link eiThe mapping path of (2);
where Cost (M)E(VNFi))，Cost(MN(VNFi))，Cost(pE(ei) And Cost (p)N(ei) Calculated according to the following formula:
Cost(Mλ(VNFi))＝p(Mλ(VNFi))×ε(VNFi)×Ti p，
Figure BDA0001340546180000121
wherein the superscript λ ∈ { E, N }, pE(tei)、pN(tei) Respectively representing the mapping of a VNF when using an existing virtual machine, a new virtual machineiTemporary link te corresponding to the timeiThe mapping path of (2).
In this model, there are two pure policy Nash equilibrium points, i.e.
Figure BDA0001340546180000122
And
Figure BDA0001340546180000123
to obtain VNFiThe pure strategy Nash equilibrium point with the highest total profit is selected as the focus equilibrium point in the invention. Focus balance Point is VNF in the present inventioniThe mapping scheme of (1). The mapping process with each VNF is a process of repeated gaming. The above-mentioned computational resource cost and link resource cost with respect to fair optimization can be described by the following linear program (4):
Figure BDA0001340546180000124
s.t.
Figure BDA0001340546180000125
Figure BDA0001340546180000126
Figure BDA0001340546180000127
Figure BDA0001340546180000128
Figure BDA0001340546180000129
Figure BDA0001340546180000131
Figure BDA0001340546180000132
Figure BDA0001340546180000133
Figure BDA0001340546180000134
λ＝{E,N} (4)
wherein,
Figure BDA0001340546180000135
representing hosted VNFsiVirtual machine of
Figure BDA0001340546180000136
The superscript λ is used to distinguish whether the virtual machine is already present or new, where E denotes already present and N denotes new, the same below.
Figure BDA0001340546180000137
Respectively representing underlying paths
Figure BDA0001340546180000138
Time delay, available bandwidth resources.
Step 1: virtual network function set N requested from service function chain to be mappedV＝{VNF1,VNF2,...,VNFnThe first VNF of the vns starts, in turn, for each VNF requested by the current service function chaini(i ═ 1, …, n) the final mapping scheme determined:
101: determining a VNF currently to be mappediAlternative server set of
Figure BDA0001340546180000139
From the set of available servers in the underlay network, the location constraints, computing resource requirements will be met and never aggregated
Figure BDA00013405461800001310
The server selected by each VNF in (1) is used as the VNFiAlternative server set of
Figure BDA00013405461800001311
102: for alternative server set
Figure BDA00013405461800001312
Each server n inkA server nkAs a hosting VNFiThe virtual machine of (1);
under the condition of satisfying eiLink resource requirement xiAnd placing the VNF on the premise of time delay requirementi-1Server (i.e. M (VNF)i-1) To server nkIn the bottom layer paths, a shortest path is searched as eiMapping path of
Figure BDA00013405461800001313
Wherein M (VNF)0) Representing a user side, namely a physical node where the user is located;
generating a connection VNFiTemporary link te with user terminaliAnd e is combinedi+1Link resource requirement xi+1As teiLink resource requirements of (1); at the time of satisfying teiOn the premise of link resource demand of (2), the slave server nkFinding a shortest path as te in the bottom path to the user terminaliMapping path of
Figure BDA00013405461800001314
By a corresponding server nkCost of computing resources CostVNFN(VNFi→nk) Link resource cost
Figure BDA00013405461800001315
The sum is obtained as the corresponding server nkTotal mapping cost of (c), server n minimizing total mapping costkIs marked as
Figure BDA00013405461800001316
103: for alternative server set
Figure BDA00013405461800001317
Each server n injComputing server njTotal mapping cost of (c), finding n where the total mapping cost is minimaljAnd is marked as
Figure BDA0001340546180000141
Wherein the server njThe total mapping cost calculation method is as follows:
judgment server njWhether a reusable virtual machine exists or not is judged, if not, the server n is judgedjThe total mapping cost of (a) is set to infinity;
if yes, the VNF is startediIs hosted on a reusable virtual machine and is satisfied with connecting with a VNFi-1And VNFiVirtual link e ofiLink resource requirement xiAnd placing the VNF on the premise of time delay requirementi-1Server to server nkIn the bottom layer paths, a shortest path is searched as eiMapping path of
Figure BDA0001340546180000142
Wherein VNF0Representing a user; and generating a connection VNFiTemporary link te with user terminaliAnd will connect to VNFiAnd VNFi+1Virtual link e ofi+1Link resource requirement xi+1As teiLink resource requirements of (1); at the time of satisfying teiOn the premise of link resource demand of (2), the slave server njFinding a shortest path as te in the bottom path to the user terminaliMapping path of
Figure BDA0001340546180000143
By server njCost of computing resources CostVNFE(VNFi→nj) Link resource cost
Figure BDA0001340546180000144
The sum is obtained as the corresponding server njTotal mapping cost of (c);
104: will be provided with
Figure BDA0001340546180000145
The server with the minimum total mapping cost is used as the placing VNFiServer M (VNF)i) Based on server M (VNF)i) Determining VNFiIncluding the server M (VN)Fi) Server M (VNF)i) Hosting VNF oniVirtual machine of (2), corresponding server M (VNF)i) E ofiThe mapping path of (2);
due to the fact that
Figure BDA0001340546180000146
Is a constant value, so is
Figure BDA0001340546180000147
The maximum can be directly obtained and converted into Cost (M)λ(VNFi) Are) and
Figure BDA0001340546180000148
and a mapping scheme with the minimum sum, wherein the lambda belongs to the N, E.
Step 2: and obtaining the optimized deployment scheme requested by the current service function chain according to the final mapping scheme of all VNFs requested by the current service function chain.
In summary, due to the adoption of the technical scheme, the invention has the beneficial effects that:
(1) the application range is wide. Traditional virtual network function or service function chain mapping algorithms are mostly proposed for virtual networks and data center networks, or do not consider a complete service function chain in a complete 5G network. The method provided by the invention can be suitable for the complete service function chain request in the 5G network, so that the method has wider application range compared with the traditional mapping algorithm.
(2) The mapping cost is low. The invention provides three schemes for the multi-target problem of the linear programming (1), and obtains a deployment scheme with low mapping cost for the service function chain request on the basis of combining a VNF merging strategy, a virtual machine reuse strategy and a temporary link mapping strategy, particularly a third scheme, and the cost of the mapping scheme is lower by comprehensively considering the modes of computing resources and link resource cost.
(3) The resource utilization rate is high. When the mapping processing is carried out, the resource consumption can be reduced by the virtual machine reuse strategy, the VNF merging strategy and the temporary link mapping strategy, so that the resource utilization rate can be improved.
(4) The mapping blocking rate is small. Because the virtual machine reuse strategy, the VNF merging strategy and the temporary link mapping strategy used in the mapping process can reduce the resource consumption, the higher the mapping success probability is, the lower the blocking rate is.
Drawings
FIG. 1 is a schematic diagram of a service function chain request in which the numbers in the rectangular box above the virtual network function represent server resource requirements and the numbers above the virtual link represent virtual link resource requirements, delays;
fig. 2 is a schematic diagram and a comparison diagram of a temporary link mapping policy, where fig. 2-a is a mapping scheme without considering the temporary link mapping policy, fig. 2-b and fig. 2-c are mapping schemes with considering the temporary link mapping policy, in the diagram, a to G represent different physical nodes (i.e., servers), a dashed line represents a mapping path of a virtual link, and a dashed dotted line represents a mapping path of a temporary link.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention will be described in further detail with reference to the following embodiments and accompanying drawings.
With a 5G network based on an SDN as an implementation object, a network operator may deploy the optimized deployment scheme (a method for mapping a service function chain) proposed by the present invention on a control layer in a control router of the SDN, and the SDN control router may schedule a control management function carried by the SDN control router to collect information of the whole network, and obtain information of resource conditions of all nodes in the network, and information of resources of links, time delay, and the like. The router can acquire the topology of the whole network and corresponding resource information by the centralized control mode. When a service function chain request comes, the SDN control router may schedule a mapping method based on the service function chain deployed on its control layer according to the whole network information grasped by itself, calculate key parameters such as mapping cost and rejection rate, and feed back the parameters to the operator.
The SDN control router with the FOCL scheme of the invention is deployed and requests a plurality of service function chainsWhen performing the on-line mapping process, firstly, defining the leaving service function chain request set as ExpiredSFC, and defining the arrival queue of the service function chain request as ArrivedSFC. In the queue arrivesfc, each service function chain request is mapped one after the other. Defining a set of service function chain requests blocked due to insufficient underlying network resources as SFCblo(blocked mapping request set SFC for short)blo). The online processing procedure for multiple service function chain requests is:
inputting:
1. underlying network resources, including underlying network GS＝(NS,ES) Resource constraint SC ═ C of underlying networkE,CN,LN) All end-to-end underlay path sets P. Wherein N isSSet of servers, set of routers representing the underlying network, ESSet of physical links representing underlying network, CEIndicating physical link properties such as bandwidth, delay, unit cost, etc., CNThe attributes of the server and the router in the underlying network are represented, wherein the attributes of the server include unit cost of server resources, server resources (such as CPU, memory and storage capacity) and the like, and the attributes of the router mainly refer to the processing capacity of the router, and LNIndicating the positions of the underlying network server and the router;
2. one arriving service function chain request queue, ArrivedSFC, where each service function chain request includes its virtual network (virtual network function set N)VVirtual link set EV) And placement constraints (resource constraint set C of VNFs)NVirtual link resource constraint set CEMaximum time delay constraint set C of virtual linkDVNF placement position constraint set LNThe location L of the userUPosition L of the service terminalT)。
And (3) outputting: mapping costs
Figure BDA0001340546180000161
And blocked set of mapping requests SFCblo。
Step 1: initial
Figure BDA0001340546180000162
Step 2: if it is not
Figure BDA0001340546180000163
Step 3 is executed; otherwise, go to step 10.
And step 3: if it is not
Figure BDA0001340546180000164
The underlying network resources are updated such that
Figure BDA0001340546180000165
Otherwise, go to step 4.
And 4, step 4: service function chain request SFC to pull head of queue from ArrivedVDCkWhere the subscript k is the identifier of the service function chain request.
And 5: invoking SFCM Algorithm to map SFCsk。
Step 6: if SFC is foundkA mapping scheme of
Figure BDA0001340546180000169
Step 7 is executed; otherwise, go to step 8.
And 7: order to
Figure BDA0001340546180000166
And updates underlying network resources and then proceeds to step 9, where
Figure BDA0001340546180000167
Representation mapping scheme
Figure BDA0001340546180000168
Of (2) mapping costs, i.e. SFCkIs calculated by the total mapping cost (computation resource cost + link resource cost) of all VNFs.
And 8: updating SFCblo＝SFCblo∪{SFCk}。
And step 9: updating an authorizedSFC＝ArrivedSFC-SFCkThen go to step 2.
Step 10: return to
Figure BDA0001340546180000171
SFCblo。
The SFCM algorithm involved in step 5 is used to find a mapping scheme for each VNF in a service function chain request, find a mapping path for the service function chain request, and allocate resources to each VNF and each virtual link. The SFCM algorithm finds a mapping scheme for each VNF using one existing virtual machine and one new virtual machine, and then determines the final mapping scheme through the sexually significant game model. The SFCM algorithm finds the mapping scheme with the minimum mapping cost as the final mapping scheme
Figure BDA0001340546180000172
This mapping scheme
Figure BDA0001340546180000173
Including the mapping cost, the mapping record of the VNF (placed server, hosted virtual machine), and the mapping record of the service function chain link (mapping path of the virtual link connecting the current VNF and the last VNF).
Inputting:
1. underlying network GS＝(NS,ES) Resource constraint SC ═ C of underlying networkE,CN,LN) All end-to-end underlay path sets P.
2. One service function chain request GV＝(NV,EV) And place constraint PC ═ CN,CE,CD,LN,LU,LT) As shown in fig. 1.
And (3) outputting: mapping scheme
Figure BDA0001340546180000174
Step 501: store all available servers in USIn (1).
Step 502: traversing N in turn, starting from the first VNFVEach VNF in (1)iExecute the next step, if NVGo to step 512 after each VNF in the set has traversed;
step 503: traversing each server n in the underlay networkk∈USIf the server in the bottom layer network has been traversed, go to step 507;
step 504: if it is not
Figure BDA0001340546180000175
And n iskPerforming steps 505 to 506 if the VNF merge policy is not used by other VNFs in the service function chain request or is satisfied;
step 505: will be the current VNFiMapping to Server nkOn a new virtual machine, and calculates and records the CostVNFN(VNFi→nk) According to equation (5);
step 506: based on the current VNFiPlaced server nkGenerating a slave nkTo the server where the user terminal is located (by the location L of the service terminal)TLearned) temporary link teiThe temporary link teiHas a virtual link resource constraint of xi+1；
Then, a Dijkstra algorithm (shortest path algorithm) is adopted, and a link e is found under the condition of meeting bandwidth and time delay constraints based on a bottom layer path set Pi(connection VNFiAnd VNFi-1Virtual link of) and temporary link teiShortest underlying path of
Figure BDA0001340546180000176
And
Figure BDA0001340546180000177
calculating and recording
Figure BDA0001340546180000181
According to equation (6), VNF is calculated and recordediTotal mapping cost of TCostVNFN(VNFi→nk) According to the formula (9), go to step 503 again;
step 507: traversing each server n in the underlay networkj∈USIf the server in the underlying network has already traversed, go to step 511;
step 508: if it is not
Figure BDA0001340546180000182
And njIf the VNF is not used by other VNFs in the service function chain request or satisfies the VNF merge policy, step 509 to step 510 are performed;
step 509: judgment server njWhether a reusable virtual machine exists or not is judged, if not, the server n is judgedjThe total mapping cost of (c) is set to infinity and goes to step 507; if yes, go to step 510;
step 510: to VNFiHosting on a reusable virtual machine, and computing and recording CostVNFE(VNFi→nj) According to equation (7);
at the same time, based on the current VNFiPlaced server njGenerating a slave njTo the server where the user terminal is located (by the location L of the service terminal)TLearned) temporary link teiThe temporary link teiHas a virtual link resource constraint of xi+1；
Finding link e by using Dijkstra algorithmiAnd a temporary link teiShortest underlying path of
Figure BDA0001340546180000183
And
Figure BDA0001340546180000184
calculating and recording
Figure BDA0001340546180000185
Calculate and record VNF according to equation (8)iTotal mapping cost of TCostVNFE(VNFi→nj) According to the formula (10), go to step 507;
step 511: mapping VNFs on all using new virtual machinesiFind a total mapping cost TCostVNF in the mapping scheme of (1)N(VNFi→nk) A minimum mapping scheme; mapping VNFs on all virtual machines that use presenceiFind a total mapping cost TCostVNF in the mapping scheme of (1)E(VNFi→nj) A minimum mapping scheme; the scheme with the minimum mapping cost is found from the two schemes according to the formula (11) as the VNFiIs finally mapped to scheme M*(VNFi) And updating the mapping scheme of the current service function chain request
Figure BDA0001340546180000186
Returning to step 502, wherein
Figure BDA0001340546180000187
The initial value of (1) is an empty set;
step 512: return to
Figure BDA0001340546180000188
When using server nkOne new virtual machine on VNFiWhen, VNFiCan be calculated according to equation (5), virtual link eiThe mapping cost of (c) can be calculated according to equation (6):
CostVNFN(VNFi→nk)＝P(nk)ε(VNFi)Ti(5)
Figure BDA0001340546180000191
when using server njOne existing virtual machine mapping VNF oniWhen, VNFiCan be calculated according to equation (7), virtual link eiThe mapping cost of (c) can be calculated according to equation (8):
CostVNFE(VNFi→nj)＝P(nj)ε(VNFi)max{Ti-To,0} (7)
Figure BDA0001340546180000192
when using server nkOne new virtual machine on VNFiThen, the minimum total mapping cost can be calculated according to equation (9):
Figure BDA0001340546180000193
when using server njOne existing virtual machine mapping VNF oniThe minimum total mapping cost can be calculated according to equation (10):
Figure BDA0001340546180000194
VNFithe minimum total mapping cost of (c) can be calculated according to equation (11):
TCostVNF(VNFi→nm)＝min{TCostVNFN(VNFi→nk),TCostVNFE(VNFi→nj)} (11)
wherein n ismIs nkOr njIf TCostVNFN(VNFi→nk)、TCostVNFE(VNFi→nj) And if the calculation results are the same, taking any one of the calculation results.
Referring to fig. 2, in a service function chain request including 2 VNFs, a User (User) is placed on a physical node a, a service Terminal (Terminal) is placed on a physical node G, and when the VNF is placed1When, assuming that the alternative physical node is B, D, if only the locally optimal mapping is considered, both alternatives are satisfied, while when selecting physical node B (fig. 2-a), this would result in VNF1Remote from the service terminal, the invention therefore passes through the temporary link teiTo constrain the VNF1Not far away from the service terminal, in calculating the VNF1And when the link resource cost between the user and the user is low, the temporary link te is usediTakes into account the mapping path of e1Link resource cost ofIn other words, it is: t is1X (((unit cost of physical path A → D) × X1+ (cost per unit of physical path D → G) × x2). Thereby making the link resource cost of selecting physical node D (fig. 2-b) smaller for the same unit cost. And then, combining the VNF merging strategy and the virtual reuse strategy to determine the VNF by considering the rule with the minimum total mapping cost2Thus, as shown in fig. 2-c, the mapping scheme of the service function chain request is obtained, and the corresponding bottom layer path is: a → D → E → G. If not passing the temporary link teiThe underlying path of FIG. 2-a, i.e., A → B → C → F → I → H → G, may be obtained, which significantly increases the cost of the link resources.
Therefore, when the server resources and bandwidth resources of the underlying network are allocated to each dynamically-arriving service function chain request, the invention achieves the aim of improving the mapping success rate of the service function chain request and the resource utilization rate of the underlying network and minimizing the total mapping cost as far as possible by considering the deployment modes of the VNF merging strategy, the virtual reuse strategy and the temporary link strategy during deployment.
While the invention has been described with reference to specific embodiments, any feature disclosed in this specification may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise; all of the disclosed features, or all of the method or process steps, may be combined in any combination, except mutually exclusive features and/or steps.

Claims (3)

1. A method for optimizing and deploying service function chains in a 5G mobile network is characterized by comprising the following steps:
step 1: virtual network function set N requested from service function chain to be mappedV＝{VNF1,VNF2,...,VNFnThe first VNF of the vns starts, in turn, for each VNF requested by the current service function chainiDetermining a set of alternative mapping schemes, where i ═ 1, …, n, n denotes the number of VNFs requested by the current service function chain:
101: determining a VNF currently to be mappediAlternative server set of (2):
from the set of available servers in the underlay network, the location constraints, computing resource requirements will be met and never aggregated
Figure FDA0002270595840000011
The server selected by each VNF in (1) is used as the VNFiAlternative server set of (3), i.e. VNFiAlternative servers that may be selected include: VNF0To VNFi-2Unselected servers, VNFi-1Selected server, wherein VNF0Representing a user;
102: determining VNFiAlternative mapping scheme set of (2):
with M (VNF)i) Presentation placement VNFiAny alternative server of (2), M (VNF)0) Representing a user terminal;
determining a server M (VNF)i) Upper for hosting VNFiThe virtual machine of (2): judging whether a reusable virtual machine exists or not, and if so, enabling the VNFiHosted on a reusable virtual machine; otherwise VNF will beiHosting on a new virtual machine; the reusable virtual machine is as follows: hosted at least one and VNF with reuse identifieriVirtual machines of the same type of VNF, and hosted VNF and VNFiDo not belong to the same service function chain request;
meeting the connection VNFi-1And VNFiVirtual link e ofiLink resource requirement xiAnd on the premise of time delay requirement, for eiMapping the bottom layer path to obtain eiMapping path of
Figure FDA0002270595840000012
Generating a connection VNFiTemporary link te with user terminaliAnd will connect to VNFiAnd VNFi+1Virtual link e ofi+1Link resource requirement xi+1As teiLink resource requirements of (1); at the time of satisfying teiOn the premise of the link resource demand, for teiMapping the bottom layer path to obtain teiMapping path of
Figure FDA0002270595840000013
Will satisfy the conditions
Figure FDA0002270595840000014
Shortest mapping path of
Figure FDA0002270595840000015
As corresponding to current M (VNF)i) E ofiOf (d), wherein esRepresenting the physical link of the underlying network, the symbol p (-) representing the unit cost of the object in parentheses, TiRepresenting VNFiThe service time of (a) is set,
Figure FDA0002270595840000016
denotes eiPreset maximum mapping cost;
will correspond to current M (VNF)i) The alternative mapping scheme of (2) is saved into an alternative mapping scheme set, wherein the alternative mapping scheme comprises: m (VNF)i) Hosting VNFiVirtual machine of, eiThe final mapping path of (2);
step 2: obtaining different mapping sets by the combination of different alternative mapping schemes of all VNFs requested by the current service function chain;
according to the formula
Figure FDA0002270595840000021
Respectively calculating the total calculation resource cost of each mapping set, and obtaining the optimized deployment scheme, epsilon (VNF), of the current service function chain request from the mapping set corresponding to the minimum total calculation resource costi) Representing VNFiThe computing resource requirements of (1);
wherein VNFiTime of payment Ti pComprises the following steps: if VNFiThe number of VNFs hosted on the virtual machine on which it is located is equal to 1, then Ti p＝Ti(ii) a If VNFiThe number of VNFs hosted on the virtual machine is greater than 1, then Ti p＝max{Ti-To0}, where T isoRepresenting hosted VNFsiThe original service time of the virtual machine.
2. The method of claim 1, wherein steps 101-102, step 2 are replaced with:
101: from the set of available servers in the underlay network, the location constraints, computing resource requirements will be met and never aggregated
Figure FDA0002270595840000022
The server selected by each VNF in (1) is used as the VNFiThe initial set of alternative servers;
separately determining a server on each initial candidate server for hosting a VNFiThe virtual machine of (2): judging whether a reusable virtual machine exists or not, and if so, enabling the VNFiHosted on a reusable virtual machine; otherwise VNF will beiHosting on a new virtual machine; the reusable virtual machine is as follows: hosted at least one and VNF with reuse identifieriVirtual machines of the same type of VNF, and hosted VNF and VNFiDo not belong to the same service function chain request;
and corresponding each initial alternative server to the VNFiCost of computing resources and VNFiPreset maximum mapping cost of
Figure FDA0002270595840000023
Compared with, will be less than or equal to
Figure FDA0002270595840000024
As VNFiAnd storing the hosted VNF on the alternative serveriThe virtual machine of (1);
wherein each initial alternative server corresponds to a VNFiThe cost of computing resources of (a) is: cost per unit x VNF of initial alternative serveriComputing resource requirement of x VNFiPay time T on initial alternative serveri p；
Time of payment Ti pComprises the following steps: if VNFiThe number of VNFs hosted on the virtual machine on which it is located is equal to 1, then Ti p＝Ti(ii) a If VNFiThe number of VNFs hosted on the virtual machine is greater than 1, then Ti p＝max{Ti-To0}, where T isiRepresenting VNFiService time of, ToRepresenting hosted VNFsiThe original service time of the virtual machine of (1);
102: determining VNFiAlternative mapping scheme set of (2):
with M (VNF)i) Presentation placement VNFiIn satisfying the connection VNFi-1And VNFiVirtual link e ofiLink resource requirement xiAnd on the premise of time delay requirement, the slave server M (VNF)i-1) To M (VNF)i) In the bottom layer paths, a shortest path is searched as eiMapping path of
Figure FDA0002270595840000031
Wherein M (VNF)0) Representing a user terminal;
generating a connection VNFiTemporary link te with user terminaliAnd will connect to VNFiAnd VNFi+1Virtual link e ofi+1Link resource requirement xi+1As teiLink resource requirements of (1); at the time of satisfying teiOn the premise of link resource demand of (c), the slave server M (VNF)i) Finding a shortest path as te in the bottom path to the user terminaliMapping path of
Figure FDA0002270595840000032
Will correspond to current M (VNF)i) The alternative mapping scheme of (2) is saved into an alternative mapping scheme set, wherein the alternative mapping scheme comprises: m (VNF)i) Hosting VNFiVirtual machine, mapping path
Figure FDA0002270595840000033
Step 2: obtaining different mapping sets by the combination of different alternative mapping schemes of all VNFs requested by the current service function chain;
according to the formula
Figure FDA0002270595840000034
Respectively calculating the total link resource cost corresponding to each mapping set, and obtaining the optimized deployment scheme of the current service function chain request by the mapping set corresponding to the minimum total link resource cost, wherein EVSet of virtual links representing current service function chain request, esRepresenting the physical links of the underlying network.
3. A method for optimizing and deploying service function chains in a 5G mobile network is characterized by comprising the following steps:
step 1: virtual network function set N requested from service function chain to be mappedV＝{VNF1,VNF2,...,VNFnThe first VNF of the vns starts, in turn, for each VNF requested by the current service function chainiThe determined final mapping scheme, where i ═ 1, …, n, n denotes the number of VNFs requested by the current service function chain:
101: determining a VNF currently to be mappediAlternative server set of
Figure FDA0002270595840000035
From the set of available servers in the underlay network, the location constraints, computing resource requirements will be met and never aggregated
Figure FDA0002270595840000036
The server selected by each VNF in (1) is used as the VNFiAlternative server set of
Figure FDA0002270595840000037
102: for alternative server set
Figure FDA0002270595840000041
Each server n inkA server nkAs a hosting VNFiThe virtual machine of (1);
meeting the connection VNFi-1And VNFiVirtual link e ofiLink resource requirement xiAnd placing the VNF on the premise of time delay requirementi-1Server to server nkIn the bottom layer paths, a shortest path is searched as eiMapping path of
Figure FDA0002270595840000042
Wherein VNF0Representing a user;
generating a connection VNFiTemporary link te with user terminaliAnd will connect to VNFiAnd VNFi+1Virtual link e ofi+1Link resource requirement xi+1As teiLink resource requirements of (1); at the time of satisfying teiOn the premise of link resource demand of (2), the slave server nkFinding a shortest path as te in the bottom path to the user terminaliMapping path of
Figure FDA0002270595840000043
By a corresponding server nkCost of computing resources CostVNFN(VNFi→nk) Link resource cost
Figure FDA0002270595840000044
The sum is obtained as the corresponding server nkTotal mapping cost of (c), server n minimizing total mapping costkIs marked as
Figure FDA0002270595840000045
Wherein CostVNFN(VNFi→nk)＝P(nk)ε(VNFi)Ti，
Figure FDA0002270595840000046
The symbol p (-) denotes the unit cost of the object in parentheses ε (VNF)i) Representing VNFiComputing resource requirement of, TiRepresenting VNFiService time of esA physical link representing an underlying network;
103: for alternative server set
Figure FDA0002270595840000047
Each server n injComputing server njTotal mapping cost of (c), finding n where the total mapping cost is minimaljAnd is marked as
Figure FDA0002270595840000048
Wherein the server njThe total mapping cost calculation method is as follows:
judgment server njWhether a reusable virtual machine exists or not is judged, if not, the server n is judgedjThe total mapping cost of (a) is set to infinity; the reusable virtual machine is as follows: hosted at least one and VNF with reuse identifieriVirtual machines of the same type of VNF, and hosted VNF and VNFiDo not belong to the same service function chain request;
if yes, the VNF is startediIs hosted on a reusable virtual machine and is satisfied with connecting with a VNFi-1And VNFiVirtual link e ofiLink resource requirement xiAnd placing the VNF on the premise of time delay requirementi-1Server to server nkIn the bottom layer paths, a shortest path is searched as eiMapping path of
Figure FDA0002270595840000049
Wherein VNF0Representing a user; generating a connection VNFiTemporary with user terminalLink teiAnd will connect to VNFiAnd VNFi+1Virtual link e ofi+1Link resource requirement xi+1As teiLink resource requirements of (1); at the time of satisfying teiOn the premise of link resource demand of (2), the slave server njFinding a shortest path as te in the bottom path to the user terminaliMapping path of
Figure FDA0002270595840000051
By server njCost of computing resources CostVNFE(VNFi→nj) Link resource cost
Figure FDA0002270595840000052
The sum is obtained as the corresponding server njTotal mapping cost of (c);
wherein CostVNFE(VNFi→nj)＝P(nj)ε(VNFi)max{Ti-To,0}，
Figure FDA0002270595840000053
esPhysical link, T, representing the underlying networkoRepresenting hosted VNFsiThe original service time of the virtual machine of (1);
104: will be provided with
Figure FDA0002270595840000054
The server with the minimum total mapping cost is used as the placing VNFiServer M (VNF)i) Based on server M (VNF)i) Determining VNFiIncluding the server M (VNF)i) Server M (VNF)i) Hosting VNF oniVirtual machine of (2), corresponding server M (VNF)i) E ofiThe mapping path of (2);
step 2: and obtaining the optimized deployment scheme requested by the current service function chain according to the final mapping scheme of all VNFs requested by the current service function chain.
CN201710536013.4A 2017-07-04 2017-07-04 Optimized deployment method of service function chain in 5G mobile network Active CN107332913B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710536013.4A CN107332913B (en) 2017-07-04 2017-07-04 Optimized deployment method of service function chain in 5G mobile network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710536013.4A CN107332913B (en) 2017-07-04 2017-07-04 Optimized deployment method of service function chain in 5G mobile network

Publications (2)

Publication Number Publication Date
CN107332913A CN107332913A (en) 2017-11-07
CN107332913B true CN107332913B (en) 2020-03-27

Family

ID=60197782

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710536013.4A Active CN107332913B (en) 2017-07-04 2017-07-04 Optimized deployment method of service function chain in 5G mobile network

Country Status (1)

Country Link
CN (1) CN107332913B (en)

Families Citing this family (50)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108322333B (en) * 2017-12-28 2021-03-02 广东电网有限责任公司电力调度控制中心 Virtual network function placement method based on genetic algorithm
CN108259238B (en) * 2018-01-10 2021-06-18 重庆邮电大学 Method for mapping network function service chain in data center
CN108134843B (en) * 2018-01-26 2020-07-31 重庆邮电大学 Service function chain deployment method under 5G-C-RAN scene
CN108200202B (en) * 2018-02-06 2019-11-12 电子科技大学 A kind of service function chain security deployment method calculating network applied to cloud and mist
CN108400940B (en) * 2018-02-27 2019-11-05 西南交通大学 A kind of multicast virtual network function dispositions method based on Estimation of Distribution Algorithm
CN108471389B (en) * 2018-03-12 2020-06-12 电子科技大学 Switch system based on service function chain
CN108494596B (en) * 2018-03-23 2020-02-21 西安电子科技大学 Collaborative construction and SFC (Small form-factor computing) mapping method for dependency among multiple VNFs (virtual network configuration functions)
CN108684046B (en) * 2018-04-23 2021-07-06 重庆邮电大学 Random learning-based access network service function chain deployment method
CN108429685B (en) * 2018-04-28 2020-08-04 电子科技大学 Service function chain routing method based on segmented routing technology
CN108616425B (en) * 2018-04-28 2021-06-01 电子科技大学 Method for relieving cascade failure risk of service function chain
CN108696596B (en) * 2018-05-29 2021-03-02 南京财经大学 Resource providing method for service combination problem in cloud integrated wireless-optical fiber hybrid access network
CN108923961B (en) * 2018-06-25 2020-02-21 电子科技大学 Multi-entry network service function chain optimization method
CN108965020A (en) * 2018-07-27 2018-12-07 北京邮电大学 Cross-domain virtual network mapping method and device thereof, computer-readable medium
US10819743B2 (en) * 2018-11-05 2020-10-27 Nanning Fugui Precision Industrial Co., Ltd. Anti-replay processing method and device utilizing the same
CN109547241B (en) * 2018-11-14 2021-07-30 南京大学 Virtual network function deployment method facing NUMA (non Uniform memory Access) architecture based on Markov algorithm
CN109831319B (en) * 2018-12-27 2021-10-26 南京大学 Network function deployment method considering multidimensional resource constraints
US10944647B2 (en) * 2019-01-24 2021-03-09 Vmware, Inc. Dynamic inter-cloud placement of virtual network functions for a slice
CN109617811B (en) * 2019-01-29 2020-06-30 电子科技大学 Rapid migration method for mobile application in cloud network
CN110018832B (en) * 2019-02-20 2022-07-12 中国人民解放军海军工程大学 Radar software component deployment strategy based on improved dynamic programming
US12004077B2 (en) 2019-03-12 2024-06-04 Nokia Technologies Oy Method, device and computer readable medium for service chain
CN110022230B (en) * 2019-03-14 2021-03-16 北京邮电大学 Deep reinforcement learning-based service chain parallel deployment method and device
CN109842528B (en) * 2019-03-19 2020-10-27 西安交通大学 Service function chain deployment method based on SDN and NFV
CN110446121B (en) * 2019-04-07 2021-05-18 西安电子科技大学 Virtual network function service chain mapping method based on betweenness centrality
CN110049130B (en) * 2019-04-22 2020-07-24 北京邮电大学 Service deployment and task scheduling method and device based on edge computing
CN110190987B (en) * 2019-05-08 2022-02-01 南京邮电大学 Virtual network function reliability deployment method based on backup income and remapping
US11012288B2 (en) 2019-05-14 2021-05-18 Vmware, Inc. Congestion avoidance in a slice-based network
US11588733B2 (en) 2019-05-14 2023-02-21 Vmware, Inc. Slice-based routing
US10892994B2 (en) 2019-05-14 2021-01-12 Vmware, Inc. Quality of service in virtual service networks
CN110120978B (en) * 2019-05-17 2021-05-14 电子科技大学 Safety protection method for elastic user cloud computing resources
CN110298381B (en) * 2019-05-24 2022-09-20 中山大学 Cloud security service function tree network intrusion detection system
CN110311862B (en) * 2019-06-14 2020-09-25 北京邮电大学 Service chain mapping method and device
EP3987396A1 (en) * 2019-06-18 2022-04-27 Telefonaktiebolaget LM Ericsson (publ) Method and system for deploying a new service function chain (sfc) based on similarities with previously deployed sfcs
CN110311816B (en) * 2019-06-28 2021-08-13 上海交通大学 VNF (virtual network function) placement method for VNF co-location interference sensing in NFV (network virtual function) environment
CN110380898B (en) * 2019-07-08 2022-11-15 紫光云技术有限公司 Hyperhedger Fabric-based multi-operator edge service chain arrangement strategy
CN110545228B (en) * 2019-07-26 2020-11-20 北京邮电大学 Service function chain request processing method and system
CN110535705B (en) * 2019-08-30 2022-04-26 西安邮电大学 Service function chain construction method capable of adapting to user time delay requirement
CN110851235B (en) * 2019-11-04 2022-06-03 中国人民解放军战略支援部队信息工程大学 Virtual network function deployment method suitable for multidimensional resource optimization configuration
CN111385202B (en) * 2020-03-17 2022-03-11 重庆邮电大学 Route distribution method based on virtual network function
CN111526070B (en) * 2020-04-29 2022-06-03 重庆邮电大学 Service function chain fault detection method based on prediction
CN111800291B (en) * 2020-05-27 2021-07-20 北京邮电大学 Service function chain deployment method and device
CN112083933B (en) * 2020-08-27 2023-07-04 重庆大学 Service function chain deployment method based on reinforcement learning
CN113032096B (en) * 2021-03-17 2023-03-31 西安电子科技大学 SFC mapping method based on node importance and user demand dual perception
CN113225211B (en) * 2021-04-27 2022-09-02 中国人民解放军空军工程大学 Fine-grained service function chain extension method
CN113535388B (en) * 2021-06-25 2023-09-29 广西综合交通大数据研究院 Task-oriented service function aggregation method
CN114205317B (en) * 2021-10-21 2023-07-21 北京邮电大学 SDN and NFV-based service function chain SFC resource allocation method and electronic equipment
CN114124796A (en) * 2021-10-26 2022-03-01 清华大学 Service chain deployment processing method and device, computer storage medium and terminal
CN114650225B (en) * 2022-03-14 2023-04-18 西安电子科技大学 Service function chain SFC mapping method based on particle swarm optimization algorithm
CN115767325B (en) * 2022-10-31 2024-09-13 苏州大学 Service function chain profit maximization mapping method and system
GB202305487D0 (en) * 2023-04-14 2023-05-31 Samsung Electronics Co Ltd Method and apparatus for efficient service function chain placement in service based 5G/6G core architecture
CN117792874A (en) * 2023-11-20 2024-03-29 中国人民解放军61660部队 Service chain efficient backup method based on cloud environment

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105141617A (en) * 2015-09-14 2015-12-09 上海华为技术有限公司 Deploying and adjusting method for service functions among data centers and deploying and adjusting device for service functions among data centers
CN105430051A (en) * 2015-10-30 2016-03-23 浙江工商大学 Service function chaining construction method for SDN
CN106411678A (en) * 2016-09-08 2017-02-15 清华大学 Bandwidth guarantee type virtual network function (VNF) deployment method
CN106487556A (en) * 2015-08-28 2017-03-08 中兴通讯股份有限公司 The dispositions method of business function SF and device

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9998563B2 (en) * 2015-10-12 2018-06-12 Fujitsu Limited Vertex-centric service function chaining in multi-domain networks

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106487556A (en) * 2015-08-28 2017-03-08 中兴通讯股份有限公司 The dispositions method of business function SF and device
CN105141617A (en) * 2015-09-14 2015-12-09 上海华为技术有限公司 Deploying and adjusting method for service functions among data centers and deploying and adjusting device for service functions among data centers
CN105430051A (en) * 2015-10-30 2016-03-23 浙江工商大学 Service function chaining construction method for SDN
CN106411678A (en) * 2016-09-08 2017-02-15 清华大学 Bandwidth guarantee type virtual network function (VNF) deployment method

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
Scalable and coordinated allocation of service function chains;Michael Till Beck等;《ELSEVIER》;20170401;第102卷;全文 *
数据中心虚拟机的迁移方法研究;赵东成;《中国优秀硕士学位论文全文数据库 信息科技辑（月刊）》;20170215;全文 *
虚拟网络的映射技术研究;孙罡;《中国博士学位论文全文数据库 信息科技辑（月刊）》;20130515;全文 *

Also Published As

Publication number Publication date
CN107332913A (en) 2017-11-07

Similar Documents

Publication Publication Date Title
CN107332913B (en) Optimized deployment method of service function chain in 5G mobile network
CN108260169B (en) QoS guarantee-based dynamic service function chain deployment method
CN109818865B (en) SDN enhanced path boxing device and method
Cicconetti et al. A decentralized framework for serverless edge computing in the internet of things
CN114338504B (en) Micro-service deployment and routing method based on network edge system
WO2023024219A1 (en) Joint optimization method and system for delay and spectrum occupancy in cloud-edge collaborative network
CN108881207B (en) Network security service realization method based on security service chain
CN108809857A (en) A method of the traffic monitoring based on SDN and service quality securing strategy
CN104170335A (en) Congestion control and resource allocation in split architecture networks
US20130132542A1 (en) Method and System for Effective BGP AS-Path Pre-pending
CN109412963B (en) Service function chain deployment method based on stream splitting
Wang et al. A reinforcement learning approach for online service tree placement in edge computing
CN114827284B (en) Service function chain arrangement method and device in industrial Internet of things and federal learning system
CN111866623A (en) High-efficiency virtual optical network survivability mapping method for service reliability
Zhu et al. Traffic flow optimization for UAVs in multi-layer information-centric software-defined FANET
CN115766722A (en) Computing power network task scheduling method and device based on information center network
Seid et al. Blockchain-empowered resource allocation in Multi-UAV-enabled 5G-RAN: a multi-agent deep reinforcement learning approach
CN109495565A (en) High concurrent service request processing method and equipment based on distributed ubiquitous computation
Nguyen et al. Efficient virtual network embedding with node ranking and intelligent link mapping
Meneguette et al. An efficient green-aware architecture for virtual machine migration in sustainable vehicular clouds
Ebrahimzadeh et al. h-horizon sequential look-ahead greedy algorithm for VNF-FG embedding
Wang et al. Low-latency service chaining with predefined NSH-based multipath across multiple datacenters
US11483210B2 (en) Interdomain path calculation based on an abstract topology
CN113094857A (en) Energy-saving controller layout method for software defined vehicle network
Liu et al. A method for adaptive resource adjustment of dynamic service function chain

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant