CN108667657B - SDN-oriented virtual network mapping method based on local feature information - Google Patents

SDN-oriented virtual network mapping method based on local feature information Download PDF

Info

Publication number
CN108667657B
CN108667657B CN201810401261.2A CN201810401261A CN108667657B CN 108667657 B CN108667657 B CN 108667657B CN 201810401261 A CN201810401261 A CN 201810401261A CN 108667657 B CN108667657 B CN 108667657B
Authority
CN
China
Prior art keywords
virtual
node
physical
mapping
network
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
CN201810401261.2A
Other languages
Chinese (zh)
Other versions
CN108667657A (en
Inventor
曲桦
赵季红
李静
牛浩时
杨绮昊
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.)
Xian Jiaotong University
Original Assignee
Xian Jiaotong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xian Jiaotong University filed Critical Xian Jiaotong University
Priority to CN201810401261.2A priority Critical patent/CN108667657B/en
Publication of CN108667657A publication Critical patent/CN108667657A/en
Application granted granted Critical
Publication of CN108667657B publication Critical patent/CN108667657B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14—Network analysis or design
    • H04L41/145—Network analysis or design involving simulating, designing, planning or modelling of a network
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00—Data switching networks
    • H04L12/28—Data switching networks characterised by path configuration, e.g. LAN [Local Area Networks] or WAN [Wide Area Networks]
    • H04L12/46—Interconnection of networks
    • H04L12/4641—Virtual LANs, VLANs, e.g. virtual private networks [VPN]

Abstract

The invention discloses a virtual network mapping method based on local characteristic information and oriented to SDN, which comprises the steps of firstly, carrying out descending order sorting on importance of virtual nodes and physical nodes according to the local characteristic information, then using a greedy algorithm to complete mapping of the virtual nodes, and finally using a K-Shortest path algorithm to complete mapping of virtual links.

Description

SDN-oriented virtual network mapping method based on local feature information
Technical Field
The invention relates to a virtual network mapping problem in a software defined network, in particular to a virtual network mapping method based on local feature information and oriented to an SDN.
Background
With the rapid development of mobile internet, internet of things and cloud computing, various emerging services are emerging continuously, the traditional network architecture cannot meet the development requirement of new services, the technical rigidity problem of the traditional internet is prominent day by day, and the new services and protocols are difficult to deploy.
Network virtualization is considered to be an effective way to solve the above problems, and it utilizes an abstraction mechanism, a distribution mechanism, and an isolation mechanism to achieve the purpose of network node and network link virtualization. The core idea of network virtualization is to abstract underlying physical resources based on the idea of software, and to satisfy the requirements of various services by creating corresponding virtual networks on the network. Through network virtualization, conventional network Service Providers (ISPs) are divided into Infrastructure Providers (inp), which are responsible for managing and controlling physical resources, and Service Providers (SPs), which provide services to end users by leasing the physical resources of the Infrastructure Providers for creating virtual networks.
Software-defined network (SDN) is a new network architecture, which is composed of five parts, namely a data layer, a control layer, an application layer, a northbound interface and a southbound interface. In conventional networks, both forwarding and control functions are implemented by underlying switches or routers. Different from the traditional network, the software defined network separates the forwarding plane from the control plane, the control function is realized by a uniform controller, and the underlying network equipment only forwards data according to the flow table rule issued by the controller. The characteristic of logic centralized control enables the software defined network to sense the resources of the underlying physical network, and facilitates the allocation and flexible scheduling of underlying network resources as required.
Virtual network mapping, i.e., allocating underlying network resources to a virtual network, is a key technical link in the network virtualization process. The virtual network mapping problem has been identified as the NP-hard problem. The effective virtual network mapping algorithm can improve the utilization rate of physical network resources, can support more virtual networks under the condition of limited physical network resources, and reduces the cost spent on mapping. During mapping, the virtual network comprises node resource constraint and link resource constraint, and virtual nodes of different virtual networks can be mapped to the same physical node. The virtual network mapping oriented to the software defined network is the virtualization of underlying network equipment, so that the virtual network mapping can support more extensible applications, and the same physical infrastructure can run various virtual networks.
The greedy algorithm is an improved hierarchical processing method, when solving a problem, the problem is not considered from the overall optimization, and the selection is a local optimal solution in a certain sense.
The Floyd algorithm is suitable for the shortest path of multiple sources, is a dynamic planning algorithm, can be positive or negative in edge weight, is simple and effective, is easy to understand, and can calculate the shortest distance between any two nodes.
Disclosure of Invention
The invention aims to provide a virtual network mapping method based on local feature information for SDN (software defined network), aiming at the mapping problem of a virtual network in a software defined network.
In order to achieve the purpose, the invention adopts the following technical scheme: an SDN-oriented virtual network mapping method based on local feature information comprises the following steps:
step 1: acquiring related information of a bottom layer physical network and a virtual network, wherein the related information comprises the topology of the physical network and the virtual network, the CPU resources and the flow table resources of a physical node and a virtual node, and the bandwidth resources of a physical link and a virtual link;
step 2: calculating local characteristic information of the physical nodes and the virtual nodes;
and step 3: sequencing the physical nodes and the virtual nodes in the step 2;
and 4, step 4: carrying out virtual network node mapping;
and 5: and carrying out virtual network link mapping.
In step 1, the obtained physical network topology is represented by using a weighted undirected graph: gs＝(Ns，Ls) In which N issRepresenting a collection of physical nodes, LsRepresenting a set of physical links, each physical node ns∈NsHas a total CPU capacity of C (n)s) Flow table space is F (n)s) (ii) a Each physical link ls∈LsHas a total bandwidth size of B (l)s) (ii) a All non-closed loop paths of the physical network are denoted as Ps(ii) a The obtained virtual network topology is also represented by a weighted undirected graph: gv＝(Nv，Lv) In which N isvRepresenting a set of virtual nodes, LvRepresenting a set of virtual links, each virtual node nv∈NvThe required CPU capacity is C (n)v) Flow table space is F (n)v) Each virtual link lv∈LvThe required bandwidth size is B (l)v)。
In the step 2, the specific steps are as follows, and the physical nodes and the virtual nodes are collectively referred to as nodes:
1) calculating local resource information of the nodes:
Figure BDA0001645775940000021
Figure BDA0001645775940000022
wherein REnAs a resource of the home node, CR(n) is the remaining CPU capacity of the node, FR(n) is the remaining flow table space of the node, BR(l) Representing the bandwidth left by the adjacent link of the node; RAnLocal resource information of the node is represented, and neighbor node resources are considered; will RAnThe normalization is expressed as:
Figure BDA0001645775940000031
2) importance of compute nodes in network topology:
Figure BDA0001645775940000032
Figure BDA0001645775940000033
wherein DnRepresenting degrees for node n, m ∈ adj (n) represents neighbor nodes for node n,m,n＝1；Wnrepresenting the importance of the node in the network topology, which takes into account the degree of the neighbor node; w is to benThe normalization is expressed as:
Figure BDA0001645775940000034
3) calculating the local characteristic information of the nodes, wherein the larger the local characteristic information is, the more important the nodes are:
Inf(n)＝RA(n)*W(n)。
in the step 3, the physical node and the virtual node perform processing according to the local feature information respectively calculated in the step 2Sorting in descending order, after sorting, the physical nodes form a set
Figure BDA0001645775940000035
Wherein the content of the first and second substances,
Figure BDA0001645775940000036
Figure BDA0001645775940000037
virtual node forming set
Figure BDA0001645775940000038
Wherein the content of the first and second substances,
Figure BDA0001645775940000039
in the step 4, a greedy algorithm is used when virtual node mapping is performed, that is, in the virtual network, a virtual node with the largest local characteristic information is mapped first, and the mapped physical node is a physical node with the largest local characteristic information which meets the resource requirement of the physical node; the method comprises the following specific steps:
1) input device
Input virtual network Gv＝(Nv，Lv) Physical network Gs＝(Ns，Ls) And a set Sort formed by descending order of the physical nodes and the virtual nodesv，Sorts；
2) Initialization
Because the nodes of the same virtual network can not be mapped to the same physical node, whether the physical node is mapped or not is marked during mapping; firstly, initializing a mark to be a pool, wherein the pool indicates that the virtual node is not mapped, and true indicates that the virtual node is mapped, and in addition, in order to subsequently indicate that the virtual node has completely finished mapping, a pointer pos of the virtual node needs to be initialized to be 0;
3) virtual node mapping
Mix SortvThe virtual nodes in (1) are mapped in sequence, and the mapping is carried out on the virtual node nvJudging whether there is a physical node nsSo that C (n)v)≤CR(ns)，F(nv)≤FR(ns) And bol (n)s) If present, nvMapping to the maximum n of local feature informationsUp, and update the pool (n)s) True, pos + 1; if not, returning the failure of virtual network mapping; and when the pos is equal to the total number of the virtual nodes, finishing the virtual node mapping and returning that the virtual node mapping is successful.
The step 5 of performing virtual link mapping by using a K-short algorithm comprises the following steps:
1) physical link exclusion
Input virtual network Gv＝(Nv，Lv) Physical network Gs＝(Ns，Ls) And a virtual node mapping result; for virtual link lvFirst, B isR(ls)<B(lv) Physical link exclusion of (2);
2) initialization
Initializing a count to be 0, and in order to represent that all virtual link mapping is finished subsequently;
3) virtual link mapping
For connecting two terminal nodes mv，nvVirtual link of (l)vFirst, find the physical node m to which its terminal is mappeds、nsThen find m using Floyd algorithms、nsShortest path p betweens(ii) a If p issIf yes, the count is equal to the count +1, and all the virtual links successfully complete mapping until the count is equal to the total number of the virtual links; otherwise, outputting the failure of virtual network mapping.
The invention also comprises a step 6: when the virtual link mapping in step 5 is successful, updating the physical network resources:
Figure BDA0001645775940000041
Figure BDA0001645775940000042
Figure BDA0001645775940000043
wherein C isR(ns) Representing a physical node nsThe remaining CPU capacity, C (n)s) Representing a physical node nsThe total capacity of the CPU of the system,
Figure BDA0001645775940000044
representation mapping to physical node nsCPU capacity consumed by all virtual nodes on the node; fR(ns) Representing a physical node nsThe remaining flow table space, F (n)s) Representing a physical node nsThe total flow meter space of (a) is,
Figure BDA0001645775940000045
representation mapping to physical node nsFlow table space consumed by all virtual nodes on the network; b isR(ls) Represents a physical link lsThe remaining bandwidth, B (l)s) Represents a physical link lsThe total bandwidth of the network (c) is,
Figure BDA0001645775940000046
representation mapping to physical link/sAll virtual links on the network consume bandwidth resources.
And 6, after the physical network resource updating in the step 6 is finished, performing the mapping process of the next virtual network topology, and repeating the steps 1 to 5.
The invention evaluates the importance of the nodes according to the local characteristic information. The local characteristic information includes the importance of the local resources of the node and the importance of the node in the network topology. The invention also utilizes a greedy algorithm and a shortest path algorithm to respectively complete the mapping of the virtual nodes and the mapping of the virtual links.
1. Related concepts
Concept one: a physical network represented using a weighted undirected graph: gs＝(Ns,Ls) In which N issRepresenting a collection of physical nodes, LsRepresenting a collection of physical links. Each physical node ns∈NsWith a CPU capacity of C (n)s) Flow table space size of F (n)s) Each physical link ls∈LsWith a bandwidth size of B (l)s). All non-closed loop paths of the underlying network are denoted as Ps。
Concept two: a virtual network represented using a weighted undirected graph: gv＝(Nv,Lv) In which N isvRepresenting a set of virtual nodes, LvRepresenting a set of virtual links. Each virtual node nv∈NvThe required CPU capacity is C (n)v) Flow table space is F (n)v) Each virtual link lv∈LvThe required bandwidth resource is B (l)v)。
2. Local feature information
Information one: local resources of the nodes, and resources of each node itself not only consider its CPU resources, flow table space, but also consider bandwidth resources of the links adjacent to it, are defined as follows:
Figure BDA0001645775940000051
wherein REnAs a resource of the home node, CR(n) is the remaining CPU capacity of the node, FR(n) is the remaining flow table space of the node, BR(l) Representing the bandwidth left by the adjacent link of the node. In order to better describe the importance of the node resources, the invention considers the influence of the adjacent node resources on the node. This is because the resource of the node cannot reflect the size of the resource of its neighboring neighbor nodes, and for this reason, the local resource of the node is defined as follows:
Figure BDA0001645775940000052
where adj (n) represents a neighbor of node n.
Its normalization is expressed as:
Figure BDA0001645775940000053
and information II: the importance of the node in the network topology is that the degree of the node cannot well describe the difference between the nodes, and therefore, the degree of the neighbor node is considered, namely the larger the degree of the node is, the larger the degree of the neighbor node of the node is, the more important the node is. The degree of a node is defined as follows:
Figure BDA0001645775940000054
m ∈ adj (n) is a neighbor node of node n,m,n＝1。
importance of nodes in network topology:
Figure BDA0001645775940000055
the normalization is expressed as:
Figure BDA0001645775940000061
the importance of the node is evaluated by integrating the two characteristics of the local characteristic information because the local resource information of the node cannot reflect the importance of the position of the node in the network topology and the importance of the node in the network topology cannot reflect the importance of the resource contained in the node. The local feature information of the defined node is as follows:
Inf(n)＝RA(n)*W(n)
the larger the local feature information of the node is, the more important the node is.
3. Virtual node mapping process
The virtual network mapping process is divided into two stages, the first stage adopts a greedy algorithm to complete the mapping of virtual nodes, and the following is the specific description of the node mapping algorithm:
1) input device
Input virtual network Gv＝(Nv，Lv) And physical network Gs＝(Ns，Ls)，Respectively calculating local characteristic information of the virtual nodes and the physical nodes, and respectively sorting the virtual nodes and the physical nodes in a descending order according to the local characteristic information to form a set
Figure BDA0001645775940000062
Wherein the content of the first and second substances,
Figure BDA0001645775940000063
and
Figure BDA0001645775940000064
wherein the content of the first and second substances,
Figure BDA0001645775940000065
2) initialization
Since nodes of the same virtual network cannot be mapped to the same physical node, whether the physical node is mapped or not needs to be marked during mapping. The mark is first initialized to cool, false indicates that it is not mapped, and true indicates that it is mapped. In addition, a pointer pos of the virtual node needs to be initialized to be 0, so that the end of the mapping of the virtual node is represented for the convenience of follow-up;
3) virtual node mapping
Mix SortvThe virtual nodes in (3) are sequentially mapped to SortsOn the physical node. For virtual node nvJudging whether there is a physical node nsSo that C (n)v)≤CR(ns)，F(nv)≤FR(ns) And bol (n)s) If present, nvMapping to the maximum n of local feature informationsUp, and update the pool (n)s) True, pos + 1; if not, returning the failure of virtual network mapping. And when the pos is equal to the total number of the virtual nodes, finishing the virtual node mapping and returning that the virtual node mapping is successful.
4. Virtual link mapping procedure
In the second stage of virtual network mapping, the mapping of the virtual link is completed by adopting a K-short path algorithm, which specifically comprises the following steps:
1) physical link exclusion
Input virtual network Gv＝(Nv，Lv) Physical network Gs＝(Ns，Ls) And a virtual node mapping result. For virtual link lvFirst, B is excludedR(ls)<B(lv) The physical link of (a);
2) initialization
Initializing a count to be 0, and in order to subsequently represent that the virtual link mapping is finished;
3) virtual link mapping
For connecting two terminal nodes mv，nvVirtual link of (l)vFirst, find the physical node m to which its terminal is mappeds、nsThen find m using Floyd algorithms、nsShortest path p betweens. If p issIf yes, the count is equal to the count +1, and the virtual link mapping is continued until the count is equal to the total number of the virtual links; otherwise, outputting the virtual network mapping failure;
5. updating physical network resources
After each virtual network is mapped successfully, updating physical network resources:
Figure BDA0001645775940000071
Figure BDA0001645775940000072
Figure BDA0001645775940000073
wherein C isR(ns) Representing a physical node nsThe remaining CPU capacity, C (n)s) Representing a physical node nsThe total capacity of the CPU of the system,
Figure BDA0001645775940000074
representation mapping to physical node nsAll virtual node consumption ofThe CPU capacity of (1); fR(ns) Representing a physical node nsThe remaining flow table space, F (n)s) Representing a physical node nsThe total flow meter space of (a) is,
Figure BDA0001645775940000075
representation mapping to physical node nsFlow table space consumed by all virtual nodes on the network; b isR(ls) Represents a physical link lsThe remaining bandwidth, B (l)s) Represents a physical link lsThe total bandwidth of the network (c) is,
Figure BDA0001645775940000076
representation mapping to physical link/sAll virtual links on the network consume bandwidth resources.
Compared with the prior art, the virtual network mapping method based on the local characteristic information for the SDN has the advantages that the resource information and the topological structure of the bottom layer physical network and the virtual network are firstly obtained, then the local characteristic information of the physical nodes and the virtual nodes is calculated, and the importance of the nodes is sorted in a descending order according to the information. This is because the mapping order of the virtual nodes in the virtual network and the physical nodes selected by the virtual node mapping affect whether the virtual network can be successfully mapped. And after the node sequencing is completed, performing virtual node mapping by adopting a greedy algorithm. In order to reduce the cost spent on mapping, the mapping of the virtual link is completed by using a K-Shortest path algorithm. Simulation proves that the virtual network mapping method provided by the invention can effectively improve the acceptance rate of virtual network requests and the profit/cost ratio of virtual network mapping.
Furthermore, the mapping sequence of the virtual nodes in the virtual network and the physical nodes selected by the virtual node mapping affect the mapping result of the virtual network, so that in step 2 of the invention, the importance of the physical nodes and the virtual nodes is evaluated by calculating the local characteristic information of the physical nodes and the virtual nodes, thereby determining the mapping sequence of the virtual nodes and the physical nodes selected by the virtual node mapping, and further improving the acceptance rate of the virtual network request.
Furthermore, in the past data, the computing node resources only include, for example, the CPU capacity of the node, the size of the flow table space, and the bandwidth resources of the adjacent links of the node, but this cannot well describe the difference between the node resources, because the node resources cannot reflect the size of the neighboring node resources, the influence of the neighboring node resources on the node is considered in step 2 of the present invention; in addition, because the degree of the node cannot well describe the difference of the importance of the node in the network topology, for this reason, the invention also considers the degree of the neighbor node; the local feature information of the node includes local resource information of the node and importance information of the node in the network topology, because the local resource information of the node cannot reflect the importance of the position of the node in the network topology, and the importance of the node in the network topology cannot reflect the importance of the resource included in the node, the local feature information integrates these two characteristics, and the importance of the node is evaluated, and the larger the local feature information is, the more important the node is.
Furthermore, the K-short algorithm is used for virtual link mapping, so that physical bandwidth resources occupied by the virtual link mapping can be reduced, and the mapping cost is effectively reduced.
Further, the virtual network mapping proposed by the present invention is divided into two stages: virtual node mapping and virtual link mapping; firstly, virtual node mapping is carried out, if the virtual node mapping fails, the virtual network mapping fails, and virtual link mapping is not needed; if the virtual node mapping is successful, virtual link mapping is carried out, and if the virtual link mapping is failed, virtual network mapping is failed; the successful mapping of both the virtual node and the virtual link means that the virtual network mapping is successful; each time a virtual network is successfully mapped, a portion of the physical network resources are consumed, and step 6 of the present invention updates the physical network resources to determine the remaining physical network resources to prepare for the next virtual network mapping.
Drawings
FIG. 1 is a flow diagram of virtual network mapping;
FIGS. 2a and 2b are graphs of simulation results, wherein FIG. 2a is a comparison line graph of the virtual network request acceptance rate simulated under the same conditions by the method of the present invention and other two prior art methods; fig. 2b is a comparative line graph of the virtual network mapping profit/cost ratio obtained by simulation under the same conditions of the method of the present invention and other two prior art methods.
Detailed Description
The invention is described in detail below with reference to the drawings and the detailed description.
As shown in fig. 1, the SDN-oriented virtual network mapping method based on local feature information provided by the present invention specifically includes the following steps:
step 1, acquiring related information of a bottom layer physical network and a virtual network:
the physical network topology is represented using a weighted undirected graph: gs＝(Ns，Ls) In which N issRepresenting a collection of physical nodes, LsRepresenting a collection of physical links. Each physical node ns∈NsHas a total CPU capacity of C (n)s) Flow table space is F (n)s). Each physical link ls∈LsHas a total bandwidth size of B (l)s) (ii) a The virtual network topology is represented using a weighted undirected graph: gv＝(Nv，Lv) In which N isvRepresenting a set of virtual nodes, LvRepresenting a set of virtual links. Each virtual node nv∈NvThe required CPU capacity is C (n)v) Flow table space is F (n)v). Each virtual link lv∈LvThe required bandwidth size is B (l)v)。
Step 2, calculating local characteristic information of the physical nodes and the virtual nodes:
2.1, calculating local resource information of the nodes:
Figure BDA0001645775940000091
Figure BDA0001645775940000092
wherein REnAs a resource of the home node, CR(n) is the remaining CPU capacity of the node, FR(n) is the remaining flow table space of the node, BR(l) Representing the bandwidth left by the neighbor link of the node; RAnThe local resources of the node are represented, normalized as:
Figure BDA0001645775940000093
2.2, importance of the computing nodes in the network topology:
Figure BDA0001645775940000094
Figure BDA0001645775940000095
wherein DnRepresenting degrees for node n, m ∈ adj (n) represents neighbor nodes for node n,m,n＝1；Wnexpressing the importance of the nodes in the network topology, and normalizing the nodes as follows:
Figure BDA0001645775940000096
2.3, calculating the local characteristic information of the nodes, wherein the larger the local characteristic information is, the more important the nodes are:
Inf(n)＝RA(n)*W(n)
and 3, sequencing the physical nodes and the virtual nodes:
the physical nodes and the virtual nodes are sorted in descending order according to the calculated local characteristic information to form a set
Figure BDA0001645775940000097
Figure BDA0001645775940000098
Wherein the content of the first and second substances,
Figure BDA0001645775940000099
and
Figure BDA00016457759400000910
Figure BDA00016457759400000911
wherein the content of the first and second substances,
Figure BDA00016457759400000912
step 4, virtual network node mapping is carried out:
4.1 input virtual network Gv＝(Nv，Lv) Physical network Gs＝(Ns，Ls) And Sort set Sortv、Sorts；
4.2, initializing a physical node mark, wherein the mark is cool, the cool indicates that the physical node mark is not mapped, and the true indicates that the physical node mark is mapped; initializing pos as 0;
4.3 for virtual node nvJudging whether there is a physical node nsSo that C (n)v)≤CR(ns)，F(nv)≤FR(ns) And bol (n)s) If present, nvMapping to the maximum n of local feature informationsUp, and update the pool (n)s) True, pos + 1; if not, returning the failure of virtual network mapping. And when pos is equal to the total number of the virtual nodes, the virtual nodes successfully complete mapping, and the step 5 is executed.
Step 5, virtual network link mapping is carried out:
5.1 for virtual Link lvFirst, B is excludedR(ls)<B(lv) The physical link of (a);
5.2, initializing a count to be 0;
5.3 for connecting two terminal nodes mv，nvVirtual link of (l)vFirst, find the physics of its terminal mappingNode ms、nsThen find m using Floyd algorithms、nsShortest path p betweens. If p issIf yes, the count is equal to the count +1, the virtual link mapping is continued, and step 6 is executed until the count is equal to the total number of the virtual links; otherwise, outputting the failure of virtual network mapping.
Step 6, updating physical network resources:
Figure BDA0001645775940000101
Figure BDA0001645775940000102
Figure BDA0001645775940000103
wherein C isR(ns) Representing a physical node nsThe remaining CPU capacity, C (n)s) Representing a physical node nsThe total capacity of the CPU of the system,
Figure BDA0001645775940000104
representation mapping to physical node nsCPU capacity consumed by all virtual nodes on the node; fR(ns) Representing a physical node nsThe remaining flow table space, F (n)s) Representing a physical node nsThe total flow meter space of (a) is,
Figure BDA0001645775940000105
representation mapping to physical node nsFlow table space consumed by all virtual nodes on the network; b isR(ls) Represents a physical link lsThe remaining bandwidth, B (l)s) Represents a physical link lsThe total bandwidth of the network (c) is,
Figure BDA0001645775940000106
representation mapping to physical link/sAll virtual links on the network consume bandwidth resources.
The invention is further described with reference to figure 2.
1. Simulation parameters
In simulation, there are 1000 virtual networks in total. The simulation parameter design is shown in table 1:
TABLE 1 simulation parameters design Table
Figure BDA0001645775940000111
2. Analysis of simulation results
The method of the invention is named as Greeny-SP-LC (Greeny Algorithm, short Path and LocalCharacter Based Virtual Network Embedding Algorithm). FIG. 2 is a comparison graph of simulation results of the present invention and the two prior art methods, wherein MARR (Markov Chains with directions ranking) represents a virtual network mapping method (gamma is 0.15) based on Markov chain reward ranking, and RW-SP (random Walk Node mapping and short Path Link mapping) represents a virtual network mapping method based on Markov random Walk virtual Node mapping and Shortest Path virtual Link mapping.
Fig. 2a is a comparison line graph of the virtual network request acceptance rate obtained by simulation under the same conditions between the method of the present invention and other two existing methods, wherein the x-axis represents the number of the arriving virtual networks, and the y-axis represents the virtual network request acceptance rate. It can be seen from the figure that the acceptance rate of the method of the present invention is significantly higher than that of the other two methods, so that the method of the present invention can effectively improve the acceptance rate of the virtual network request.
Fig. 2b is a comparative line graph of the ratio of the virtual network mapping profit/cost obtained by simulation under the same conditions of the method of the present invention and other two existing methods, wherein the x-axis represents the number of the virtual networks arriving, and the y-axis represents the ratio of the virtual network mapping profit/cost. It can be seen from the figure that, compared with the other two methods, the method of the present invention has the highest profit/cost ratio, so that the method of the present invention can effectively improve the profit/cost ratio of the virtual network mapping.

Claims (4)

1. An SDN-oriented virtual network mapping method based on local feature information is characterized in that: the method comprises the following steps:
step 1: acquiring related information of a bottom layer physical network and a virtual network, wherein the related information comprises the topology of the physical network and the virtual network, the CPU resources and the flow table space of a physical node and a virtual node, and the bandwidth resources of a physical link and a virtual link;
in step 1, the obtained physical network topology is represented by using a weighted undirected graph: gs＝(Ns，Ls) In which N issRepresenting a collection of physical nodes, LsRepresenting a set of physical links, each physical node ns∈NsHas a total CPU capacity of C (n)s) Flow table space is F (n)s) (ii) a Each physical link ls∈LsHas a total bandwidth size of B (l)s) (ii) a All non-closed loop paths of the physical network are denoted as Ps(ii) a The obtained virtual network topology is also represented by a weighted undirected graph: gv＝(Nv，Lv) In which N isvRepresenting a set of virtual nodes, LvRepresenting a set of virtual links, each virtual node nv∈NvThe required CPU capacity is C (n)v) Flow table space is F (n)v) Each virtual link lv∈LvThe required bandwidth size is B (l)v)；
Step 2: calculating local characteristic information of the physical nodes and the virtual nodes;
and step 3: sequencing the physical nodes and the virtual nodes in the step 2;
and 4, step 4: carrying out virtual network node mapping;
and 5: carrying out virtual network link mapping;
in the step 2, the concrete steps are as follows:
1) calculating local resource information of the nodes:
Figure FDA0002521024730000011
Figure FDA0002521024730000012
wherein m ∈ adj (n) denotes a neighbor node of node n, REnIs a resource of node n, CR(n) is the remaining CPU capacity of node n, FR(n) is the remaining flow table space of node n, BR(l) Representing the bandwidth left by the adjacent link of the node n; RAnLocal resource information representing a node n, which takes into account neighboring node resources; will RAnThe normalization is expressed as:
Figure FDA0002521024730000013
2) importance of compute nodes in network topology:
Figure FDA0002521024730000014
Figure FDA0002521024730000015
wherein DnWhich represents the degree of the node n,m，n＝1；Wnrepresenting the importance of the node in the network topology, which takes into account the degree of the neighbor node; w is to benThe normalization is expressed as:
Figure FDA0002521024730000021
3) calculating the local characteristic information of the nodes, wherein the larger the local characteristic information is, the more important the nodes are:
Inf(n)＝RA(n)*W(n)；
in the step 3, the physical nodes and the virtual nodes are sorted in a descending order according to the local feature information calculated in the step 2, and after sorting, the physical nodes form a set
Figure FDA0002521024730000022
Wherein the content of the first and second substances,
Figure FDA0002521024730000023
Figure FDA0002521024730000024
virtual node forming set
Figure FDA0002521024730000025
Wherein the content of the first and second substances,
Figure FDA0002521024730000026
the step 5 of performing virtual link mapping by using a K-short algorithm comprises the following steps:
1) physical link exclusion
Input virtual network Gv＝(Nv，Lv) Physical network Gs＝(Ns，Ls) And a virtual node mapping result; for virtual link lvFirst, B isR(ls)＜B(lv) Physical link exclusion of (2);
2) initialization
Initializing a count to be 0, and in order to represent that all virtual link mapping is finished subsequently;
3) virtual link mapping
For connecting two virtual nodes mv，nvVirtual link of (l)vFirst, find the physical node m to which the virtual node mapss、nsThen find m using Floyd algorithms、nsShortest path p betweens(ii) a If p issIf yes, the count is equal to the count +1, and all the virtual links successfully complete mapping until the count is equal to the total number of the virtual links; otherwise, outputting the failure of virtual network mapping.
2. The SDN-oriented local feature information-based virtual network mapping method according to claim 1, wherein: in the step 4, a greedy algorithm is used when virtual node mapping is performed, that is, in the virtual network, a virtual node with the largest local characteristic information is mapped first, and the mapped physical node is a physical node with the largest local characteristic information which meets the resource requirement of the physical node; the method comprises the following specific steps:
1) input device
Input virtual network Gv＝(Nv，Lv) Physical network Gs＝(Ns，Ls) And a set Sort formed by descending order of the physical nodes and the virtual nodesv，Sorts；
2) Initialization
Because the nodes of the same virtual network can not be mapped to the same physical node, whether the physical node is mapped or not is marked during mapping; firstly, initializing a mark to be a pool, wherein the pool indicates that the virtual node is not mapped, and true indicates that the virtual node is mapped, and in addition, in order to subsequently indicate that the virtual node has completely finished mapping, a pointer pos of the virtual node needs to be initialized to be 0;
3) virtual node mapping
Mix SortvThe virtual nodes in (1) are mapped in sequence, and the mapping is carried out on the virtual node nvJudging whether there is a physical node nsSo that C (n)v)≤CR(ns)，F(nv)≤FR(ns) And bol (n)s) If present, nvMapping to the maximum n of local feature informationsUp, and update the pool (n)s) True, pos + 1; if not, returning the failure of virtual network mapping; and when the pos is equal to the total number of the virtual nodes, finishing the virtual node mapping and returning that the virtual node mapping is successful.
3. The SDN-oriented local feature information-based virtual network mapping method according to claim 1, wherein: further comprising the step 6: when the virtual link mapping in step 5 is successful, updating the physical network resources:
Figure FDA0002521024730000031
Figure FDA0002521024730000032
Figure FDA0002521024730000033
wherein C isR(ns) Representing a physical node nsThe remaining CPU capacity, C (n)s) Representing a physical node nsThe total capacity of the CPU of the system,
Figure FDA0002521024730000034
representation mapping to physical node nsCPU capacity consumed by all virtual nodes on the node; fR(ns) Representing a physical node nsThe remaining flow table space, F (n)s) Representing a physical node nsThe total flow meter space of (a) is,
Figure FDA0002521024730000035
representation mapping to physical node nsFlow table space consumed by all virtual nodes on the network; b isR(ls) Represents a physical link lsThe remaining bandwidth, B (l)s) Represents a physical link lsThe total bandwidth of the network (c) is,
Figure FDA0002521024730000036
representation mapping to physical link/sAll virtual links on the network consume bandwidth resources.
4. The SDN-oriented local feature information-based virtual network mapping method according to claim 3, wherein: and after the physical network resource is updated, performing the mapping process of the next virtual network, and repeating the steps 1 to 5.
CN201810401261.2A 2018-04-28 2018-04-28 SDN-oriented virtual network mapping method based on local feature information Active CN108667657B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201810401261.2A CN108667657B (en) 2018-04-28 2018-04-28 SDN-oriented virtual network mapping method based on local feature information

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201810401261.2A CN108667657B (en) 2018-04-28 2018-04-28 SDN-oriented virtual network mapping method based on local feature information

Publications (2)

Publication Number Publication Date
CN108667657A CN108667657A (en) 2018-10-16
CN108667657B true CN108667657B (en) 2020-08-18

Family

ID=63781343

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201810401261.2A Active CN108667657B (en) 2018-04-28 2018-04-28 SDN-oriented virtual network mapping method based on local feature information

Country Status (1)

Country Link
CN (1) CN108667657B (en)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109495300B (en) * 2018-11-07 2020-05-26 西安交通大学 Reliable SDN virtual network mapping method
CN109474464B (en) * 2018-11-13 2020-06-16 电子科技大学 Rapid network updating method based on OpenNF mechanism
CN110224918A (en) * 2019-05-17 2019-09-10 西安交通大学 A kind of cross-domain SDN mapping method of virtual network
CN110213149B (en) * 2019-05-30 2021-02-26 西安邮电大学 Virtual network mapping method based on node entropy
CN111245701B (en) * 2020-01-20 2021-08-31 中国电子科技集团公司第五十四研究所 Link priority virtual network mapping method based on maximum weighted matching
CN111526057B (en) * 2020-04-30 2022-10-25 西安邮电大学 Network slice reliability mapping method based on service type
CN112688848B (en) * 2020-12-17 2021-12-21 北京邮电大学 Heterogeneous virtual network mapping method and system
CN114827273B (en) * 2022-03-30 2023-11-24 南京邮电大学 Virtual network mapping method based on time delay optimization

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103457752A (en) * 2012-05-30 2013-12-18 中国科学院声学研究所 Virtual network mapping method
CN105681153A (en) * 2016-03-16 2016-06-15 北京邮电大学 Virtual network mapping method and device

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10313195B2 (en) * 2015-09-14 2019-06-04 Huawei Technologies Co., Ltd. Dedicated protection for virtual network embedding
CN106100964B (en) * 2016-08-24 2019-06-07 北京邮电大学 A kind of method and apparatus of virtual network mapping

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103457752A (en) * 2012-05-30 2013-12-18 中国科学院声学研究所 Virtual network mapping method
CN105681153A (en) * 2016-03-16 2016-06-15 北京邮电大学 Virtual network mapping method and device

Also Published As

Publication number Publication date
CN108667657A (en) 2018-10-16

Similar Documents

Publication Publication Date Title
CN108667657B (en) SDN-oriented virtual network mapping method based on local feature information
Zhang et al. Adaptive interference-aware VNF placement for service-customized 5G network slices
Yao et al. A novel reinforcement learning algorithm for virtual network embedding
CN112153700B (en) Network slice resource management method and equipment
CN108260169B (en) QoS guarantee-based dynamic service function chain deployment method
CN108566659B (en) 5G network slice online mapping method based on reliability
CN113315700B (en) Computing resource scheduling method, device and storage medium
Chowdhury et al. Vineyard: Virtual network embedding algorithms with coordinated node and link mapping
CN105103506B (en) For the method and system for the non-homogeneous bandwidth request allocation bandwidth in system for cloud computing
WO2020168761A1 (en) Model training method and apparatus
WO2018176385A1 (en) System and method for network slicing for service-oriented networks
CN111538570B (en) Energy-saving and QoS guarantee-oriented VNF deployment method and device
CN113435472A (en) Vehicle-mounted computing power network user demand prediction method, system, device and medium
Kim et al. Multi-agent reinforcement learning-based resource management for end-to-end network slicing
CN108965014A (en) The service chaining backup method and system of QoS perception
CN107196806B (en) Topological proximity matching virtual network mapping method based on sub-graph radiation
CN110365568A (en) A kind of mapping method of virtual network based on deeply study
CN108111335A (en) A kind of method and system dispatched and link virtual network function
CN112073237B (en) Large-scale target network construction method in cloud edge architecture
CN103997515B (en) Center system of selection and its application are calculated in a kind of distributed cloud
Bouzidi et al. Dynamic clustering of software defined network switches and controller placement using deep reinforcement learning
Kanwal et al. A genetic based leader election algorithm for IoT cloud data processing
CN110958192B (en) Virtual data center resource allocation system and method based on virtual switch
CN110290206A (en) A kind of distributed computing system and method for cafe environment
Zerwas et al. Ismael: Using machine learning to predict acceptance of virtual clusters in data centers

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