CN105871606B - Mapping method for enhancing survivability of virtual network based on divide-and-conquer strategy - Google Patents

Mapping method for enhancing survivability of virtual network based on divide-and-conquer strategy Download PDF

Info

Publication number
CN105871606B
CN105871606B CN201610188764.7A CN201610188764A CN105871606B CN 105871606 B CN105871606 B CN 105871606B CN 201610188764 A CN201610188764 A CN 201610188764A CN 105871606 B CN105871606 B CN 105871606B
Authority
CN
China
Prior art keywords
node
network
resources
sub
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
CN201610188764.7A
Other languages
Chinese (zh)
Other versions
CN105871606A (en
Inventor
赵夙
王艳军
朱晓荣
黄正超
王振
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.)
Nanjing University of Posts and Telecommunications
Original Assignee
Nanjing University of Posts and Telecommunications
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 Nanjing University of Posts and Telecommunications filed Critical Nanjing University of Posts and Telecommunications
Priority to CN201610188764.7A priority Critical patent/CN105871606B/en
Publication of CN105871606A publication Critical patent/CN105871606A/en
Application granted granted Critical
Publication of CN105871606B publication Critical patent/CN105871606B/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
    • 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]
    • 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
    • H04L41/06—Management of faults, events, alarms or notifications
    • H04L41/0654—Management of faults, events, alarms or notifications using network fault recovery
    • H04L41/0663—Performing the actions predefined by failover planning, e.g. switching to standby network elements

Landscapes

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

Abstract

A mapping method for enhancing the survivability of a virtual network based on a divide-and-conquer strategy is generally divided into a physical network initialization stage, an initial mapping stage and a remapping stage, and the specific process comprises the following steps: (1) constructing a sub-network based on a logic sub-network division method of a heuristic algorithm; (2) dividing node resources based on a node candidate resource pool generation method of a divide-and-conquer strategy; (3) dividing link resources based on a generation method of a link candidate resource pool of a divide-and-conquer strategy; (4) the virtual network initial mapping method based on the restorability reduces the failure probability of the virtual network; (5) calculating the similarity between node resources based on the similarity measurement of the fault nodes of the similarity function; (6) and optimizing the virtual network remapping by using the mixed integer programming model of the virtual network remapping. The method effectively solves the problem of poor survivability of the traditional virtual network, provides a stable and efficient survivability guarantee for the virtual network, and has important practical significance and good application prospect.

Description

Mapping method for enhancing survivability of virtual network based on divide-and-conquer strategy
Technical Field
The invention particularly relates to a mapping method for enhancing the survivability of a virtual network based on a divide-and-conquer strategy, and belongs to the technical field of network virtualization virtual network mapping.
Background
With the rapid development of wireless communication technology and the increasing demand for diversified mobile services, a wireless network will exhibit deployment density, traffic diversity and network heterogeneity in the future, and new technologies and new applications that are continuously emerging also place higher and more strict requirements on the network, such as microservice providers, personal networks, resource customization services, etc., however, the current network architecture cannot bear these new network features, which will seriously hinder the network architecture and technical innovation, and will also delay the growth of emerging networks.
The network virtualization is regarded as the most potential technology for solving the current network rigidity problem and decoupling the future network architecture, the complex network management and control function can be separated from hardware, and the upper layer is extracted to perform unified coordination management and diversified configuration, so that the network management cost is reduced, the network management and control efficiency is improved, the abstraction, the unified representation, the resource sharing and the efficient multiplexing of network resources are realized, and a feasible scheme is provided for the coexistence and the fusion of heterogeneous wireless networks. Network virtualization has gained increasing attention and research depth as a core paradigm for future networks, such as the X-Bone project and the GENI project. The X-Bone constructs a virtual network through an encapsulation technology, and supports the functions of discovery, deployment and monitoring of dynamic resources; the GENI is a worldwide network virtualization project initiated by the National Science Foundation (NSF), which partitions resources from the time and space aspects to realize virtualization based on the original network virtualization technical achievements, and further constructs a network test platform characterized by openness and large scale to provide conditions and basis for exploring the next generation of internet.
Virtual network mapping is an important branch of network virtualization and is responsible for effectively mapping virtual network resources onto physical network resources. However, most of the current virtual network mapping mechanisms are based on the assumption that the network normally operates, and in an actual network environment, a physical network can encounter severe scenes such as natural disasters, artificial attacks and the like at different degrees, so that the network presents more complex burstiness and dynamics. In order to make the virtual network have higher reliability, the virtual network must have a stronger survivability and robustness mechanism, and further, the normal operation of the virtual network is still ensured when the physical network fails, which is also a key problem to be solved by the present invention.
Disclosure of Invention
The technical problem is as follows: the invention aims to provide a mapping method for enhancing the survivability of a virtual network based on a divide-and-conquer strategy, mainly solves the problem of poor survivability of the virtual network caused by instability of a physical network, and provides survivability guarantee for normal operation of the virtual network.
The technical scheme is as follows: the invention provides a mapping method for enhancing the survivability of a virtual network based on a divide-and-conquer strategy, which comprises the following steps: (1) a logical sub-network partitioning method based on heuristic algorithm, which is used for logically partitioning a physical network; (2) the node candidate resource pool generation method based on the divide-and-conquer strategy is used for dividing the node resources of each sub-network into working resources and backup resources so as to provide candidate nodes for fault nodes; (3) the method for generating the link candidate resource pool based on the divide-and-conquer strategy is used for dividing the link resources of each sub-network into working resources and backup resources so as to provide candidate links for a fault link and an affected link; (4) the virtual network initial mapping method based on the recoverability is used for mapping important virtual resources to physical resources with high recoverability preferentially in an initial mapping stage; (5) the similarity measurement of the fault nodes based on the similarity function is used for describing and calculating the similarity of the fault nodes and the candidate nodes so as to improve the success rate of remapping; (6) the mixed integer programming model of the virtual network remapping is used for optimizing the virtual network remapping so as to improve the efficiency of the remapping mechanism.
(1) Logical subnetwork division method based on heuristic algorithm for given physical network GsAnd a number of zones K, defining an evaluation function f (i, n) ≦ g (i, n) + h (i, n) (1 ≦ i ≦ K), wherein i represents the ith sub-network, g (i, n) represents the actual cost from the initial node to the current node n, i.e., the actual step size (node hop count) from the initial node to the current node, and h (i, n) represents the estimated cost of the best path from the current node n to the destination node, i.e., the pre-generation sub-network
Figure GDA0002315074820000021
The maximum step length from the initial node is used for providing heuristic information required by the search process. And selecting a path which most possibly reaches the destination node from the state space according to the valuation function, thereby logically partitioning the physical network. In the process of heuristic search, once the searched node is the node which is marked, the node is indicated to be positioned on the boundary of the sub-network, so the heuristic search needs to be repeated when the heuristic search is finishedStarting a search until all nodes connected with the initial node in the current sub-network are marked as visited, and then judging whether the current sub-network is in a visited state
Figure GDA0002315074820000022
The division is successful. i denotes an index of the ith sub-network,
Figure GDA0002315074820000023
representing the ith pre-generated subnetwork.
(2) Node candidate resource pool generation method based on divide-and-conquer strategy, aiming at each pre-generated sub-network generated by the process (1)
Figure GDA0002315074820000031
According to its node backup proportion
Figure GDA0002315074820000032
And pre-generation of sub-networks
Figure GDA0002315074820000033
Node resource collection of
Figure GDA0002315074820000034
Number of computing node resource backups
Figure GDA0002315074820000035
And dividing the node resources of the sub-network into working resources and backup resources, and further constructing a node candidate pool of the sub-network so as to provide candidate nodes for the failed node. In order to solve the problems of insufficient node backup resources and node backup resource redundancy as much as possible, the generation method of the node candidate resource pool adopts historical fault data
Figure GDA0002315074820000036
The resource backup proportion is updated, so that the candidate resource pool is more coordinated with the current network condition, and the problems are solved,
Figure GDA0002315074820000037
representing ith sub-network with N physical network nodes in T period
Figure GDA0002315074820000038
The total number of failed nodes in the group.
(3) Method for generating a pool of link candidates based on a divide-and-conquer strategy, for each pre-generated sub-network G generated by said process (1)iAccording to its link backup ratio
Figure GDA0002315074820000039
And pre-generation of sub-networks
Figure GDA00023150748200000310
Calculating the number of link resource backups
Figure GDA00023150748200000311
And dividing the link resources of the sub-network into working resources and backup resources, and further constructing a link candidate pool of the sub-network so as to provide candidate paths for the failed link and the affected link.
(4) The method comprises calculating the node resource influence degree of the physical network to the virtual network
Figure GDA00023150748200000312
And the degree of link resource impact
Figure GDA00023150748200000313
Secondly, the restorability of the physical node resource is obtained
Figure GDA00023150748200000314
And degree of recoverability of physical link resources
Figure GDA00023150748200000315
Since the recoverability of the resource is inversely related to the influence degree of the resource, the important virtual resource is preferentially mapped to the physical resource with high recovery degree in the initial mapping stage, so that the resource recovery method can be applied to the resource recovery systemProviding a certain degree of survivability for the virtual network during initial mapping; with NSAnd ESRespectively representing a set of node resources and a set of link resources of the physical network, nsRepresenting a node, i.e.
Figure GDA00023150748200000317
esIndicating a link, i.e.
Figure GDA00023150748200000318
(5) The similarity measurement of the fault nodes based on the similarity function is used for measuring the similarity of the fault nodes based on the processes (2) and (3) by using ciDenotes CPU processing capability of node i, miIndicating the storage capability of node i,/iRepresenting the geographical location of node i, the process first utilizes attributes of the physical node resources, such as CPU processing capacity ciStorage capacity miGeographic location liEtc. constructing a node attribute column vector
Figure GDA00023150748200000316
For any node n of backup resource pooljConstructed column vector
Figure GDA0002315074820000041
For a failed node nfConstructed column vector
Figure GDA0002315074820000042
Further describing and calculating the similarity of the fault node resource and the candidate node resource
Figure GDA0002315074820000043
And screening candidate node resources with the most similar physical characteristics to the fault node to improve the success rate of remapping. Aiming at the problems of different requirements of different network environments and application scenes on network performance, different weights c 'are given to different attributes of resources during specific remapping'f＝αcf，m'f＝βmf，l'f＝λlfIn order to represent the need for different features,the flexibility of survivability mapping is increased.
(6) A mixed integer programming model OF virtual network remapping is characterized in that a resource capacity constraint equation, a node mapping constraint equation, a flow conservation constraint equation and a variable value constraint equation are established on the basis OF a virtual node set FN (n, x) which is influenced by a fault node x and needs to be migrated and a virtual link set FL (l, x) which is influenced by the fault node x and needs to be migrated, the candidate resource capacity is respectively guaranteed to be not less than the total virtual resource requirement, the uniqueness OF virtual resource mapping to be remapped, the network flow conservation and the legal value OF an integer variable are respectively guaranteed, and an objective function OF is established on the basis OF the process (5) and used for optimizing virtual network remapping so as to improve the efficiency OF a remapping mechanism.
Advantageous effects
The method effectively solves the problem of poor survivability of the virtual network in the traditional virtual network mapping, provides a stable and efficient survivability method for the virtual network, and has important practical significance and good application prospect.
Drawings
FIG. 1 is an overall system flow diagram of an embodiment of the present invention;
FIG. 2 is a schematic diagram of logical partitioning of a physical network according to an embodiment of the present invention;
fig. 3 is a schematic diagram of a node resource similarity measurement according to an embodiment of the present invention.
Detailed Description
The mapping method for enhancing the survivability of the virtual network based on the divide-and-conquer strategy according to the present invention will be described in detail with reference to the accompanying drawings and the detailed description.
As shown in fig. 1, a mapping method for enhancing the survivability of a virtual network based on a divide-and-conquer strategy mainly includes the following processes:
(1) a logical sub-network partitioning method based on heuristic algorithm, which is used for logically partitioning a physical network;
(2) the node candidate resource pool generation method based on the divide-and-conquer strategy is used for dividing the node resources of each sub-network into working resources and backup resources so as to provide candidate nodes for fault nodes;
(3) the method for generating the link candidate resource pool based on the divide-and-conquer strategy is used for dividing the link resources of each sub-network into working resources and backup resources so as to provide candidate links for a fault link and an affected link;
(4) the virtual network initial mapping method based on the recoverability is used for mapping important virtual resources to physical resources with high recoverability preferentially in an initial mapping stage;
(5) the similarity measurement of the fault nodes based on the similarity function is used for describing and calculating the similarity of the fault nodes and the candidate nodes so as to improve the success rate of remapping;
(6) the mixed integer programming model of the virtual network remapping is used for optimizing the virtual network remapping so as to improve the efficiency of the remapping mechanism.
The system operation flow of the present invention is described with reference to fig. 1:
1. process (1) logical sub-network partitioning method based on heuristic algorithm, for a given physical network GsAnd a number of zones K, defining an evaluation function f (i, n) ≦ g (i, n) + h (i, n) (1 ≦ i ≦ K), wherein i represents the ith sub-network, g (i, n) represents the actual cost from the initial node to the current node n, i.e., the actual step size (node hop count) from the initial node to the current node, and h (i, n) represents the estimated cost of the best path from the current node n to the destination node, i.e., the pre-generation sub-network
Figure GDA0002315074820000051
The maximum step length from the initial node is used for providing heuristic information required by the search process. And selecting a path which most possibly reaches the destination node from the state space according to the valuation function, thereby logically partitioning the physical network. In the process of heuristic search, once the searched node is the marked node, the node is indicated to be positioned on the boundary of the sub-network, therefore, after the heuristic search is finished, the search needs to be started again until the nodes connected with the initial node in the current sub-network are all marked to be in the visited state, and the current sub-network is in the visited stateThe division is successful. Now, taking the generation of the ith logical sub-network as an example, a heuristic algorithm for constructing the logical sub-networks is described.
(1) With an initial node niStructuring sub-networks
Figure GDA0002315074820000053
And a set of OPEN, the OPEN being,
Figure GDA0002315074820000054
OPEN←{nilet CLOSE be an empty set, let niMarked as accessed.
(2) The loop is opened and if OPEN is the empty set, the algorithm ends with a failure.
(3) Take node n with the minimum f (i, n) from OPEN and let OPEN ← OPEN- { n }, CLOSE ← CLOSE ∪ { n }.
(4) If n has been marked by another subnet, the search process ends and a heuristic search is restarted, beginning at step ①.
(5) Expanding n and making M be the child node of n and not the node set of its parent node, then
Figure GDA0002315074820000061
(6) For each node M ∈ M:
① if
Figure GDA00023150748200000610
And is
Figure GDA00023150748200000611
OPEN ← { m }, while estimating h (i, m) and calculating f (i, m) ═ g (i, m) + h (i, m).
② if m ∈ OPEN or m ∈ CLOSE, adjust its backtracking pointer to the path that gives the minimum g (i, m) value.
③ if m's trace back pointer is adjusted and m ∈ CLOSE, then OPEN ← { m }, again.
(7) And returning to the circulation state.
Through the above steps of constructing logical sub-networks, the physical network
Figure GDA0002315074820000062
Is divided into K regions, where NSAnd ESRespectively representing a set of node resources and a set of link resources of the physical network,
Figure GDA0002315074820000063
representing a collection of physical network node attributes, i.e. nodes
Figure GDA00023150748200000612
For the attribute of
Figure GDA0002315074820000064
The representation of the set is represented by a set,
Figure GDA0002315074820000065
representing a set of physical network link attributes, i.e. node isAnd
Figure GDA00023150748200000613
link e (i) betweens,js)∈ESFor the attribute of
Figure GDA0002315074820000066
And (4) representing a set. Is provided with
Figure GDA0002315074820000067
For i sub-network (meaning of each variable is similar to G)S) It is obvious that
Figure GDA0002315074820000068
Is GSThe constraint is as follows:
Figure GDA0002315074820000069
referring specifically to FIG. 2, a physical network G is shownSIs logically divided into 4 sub-networks of G1-G4 and the like. (note: for the sake of convenience of presentation,
Figure GDA0002315074820000071
equivalent to GiThe same below)
2. Process (2) node candidate resource pool generation method based on divide-and-conquer strategy, for each sub-network generated by the process (1)
Figure GDA0002315074820000072
According to its node backup proportion
Figure GDA0002315074820000073
Number of computing node resource backups
Figure GDA0002315074820000074
The node resources of the sub-network are divided into working resources and backup resources. Definition of
Figure GDA0002315074820000075
Representing the degree of resources of node n in network G, where C1、C2The balance factors, CPU (n) and mem (n), respectively, greater than zero represent the CPU capacity and memory capacity of the physical node n, respectively, and this function characterizes the amount of resource capacity that the node n possesses in the network G. Is provided with
Figure GDA0002315074820000076
Is according to D (G) by the node in the ith sub-networkiN) front in descending order
Figure GDA0002315074820000077
The node set formed by each element is the node backup set of the ith sub-network
Figure GDA0002315074820000078
Thereby building a candidate pool of nodes for the sub-network to provide candidate nodes for the failed node. Referring specifically to fig. 2, 4 subnetworks are each assigned a candidate resource pool.
In order to solve the problems of insufficient node backup resources and node backup resource redundancy as much as possible, the node candidate resource pool is generatedThe method adopts historical fault data to update the resource backup proportion, so that the candidate resource pool is more coordinated with the current network condition. The specific implementation scheme is as follows: definition of
Figure GDA0002315074820000079
Indicating the ith sub-network within T periods of time
Figure GDA00023150748200000710
Of failed nodes, where nfailIndicating a failed node. Since network node failures occur randomly, it is not necessary to determine whether a network node failure occurred randomly
Figure GDA00023150748200000711
Is a random variable, then the random variable is first mapped to before the next VN mapping
Figure GDA00023150748200000712
Calculating a statistical average
Figure GDA00023150748200000713
According to the formula
Figure GDA00023150748200000714
Updating sub-networks
Figure GDA00023150748200000715
Node backup ratio of
Figure GDA00023150748200000716
εNRepresenting a node limiting factor for ensuring that the new node backup ratio is not greater than 1 when a large network fault occurs,
Figure GDA00023150748200000717
representing sub-networks
Figure GDA00023150748200000718
The backup ratio of old nodes.
3. Process (3) generation method of link candidate resource pool based on divide and conquer strategy, and its programFor each sub-network generated by the process (1)
Figure GDA00023150748200000719
According to its link backup ratio
Figure GDA00023150748200000720
Calculating the number of link resource backups
Figure GDA0002315074820000081
The link resources of the sub-network are divided into working resources and backup resources. Modeling in the same process (2) to construct the ith sub-network
Figure GDA00023150748200000811
Is a link backup set of
Figure GDA0002315074820000082
A candidate pool of links for the sub-network is generated to provide candidate paths for the failed link and the affected links.
4. The method comprises (4) calculating the node resource influence degree of the physical network to the virtual network
Figure GDA0002315074820000083
And the degree of link resource impact
Figure GDA0002315074820000084
The specific calculation is as follows:
node resource influence degree:
Figure GDA0002315074820000085
wherein S represents the mapping in the sub-network
Figure GDA0002315074820000086
The total number of virtual resources in (c); at a physical node nsIn the case of a cut point, P represents the number of clusters into which the current subnetwork is to be divided when it fails, sjRepresenting the total number of virtual resources in the jth cluster network;at a physical node nsIn the case of non-secant points, Mig (n)s) Represents the total number of virtual resources, Mig, that need to be migrated in the event of its failureN(ns) Represents the total number of virtual node resources, Mig, that need to be migrated in the event of its failureE(ns) Representing the total number of virtual link resources that need to be migrated when it fails.
The link resource influence degree:
Figure GDA0002315074820000087
wherein S represents the mapping in the sub-network
Figure GDA0002315074820000088
The total number of virtual resources in (c); on physical link esUnder the condition of edge cutting, if the edge is broken, the current sub-network is divided into two clusters; on physical link esIn the case of non-cutting edges, Mig (e)s) Representing a physical link esTotal number of virtual resources, Mig, to be migrated in case of failureE(es) Representing a physical link esThe number of virtual link resources that need to be migrated in the event of a failure. The recoverability of the physical node resource and the physical link resource are respectively
Figure GDA0002315074820000089
And
Figure GDA00023150748200000810
since the recoverability of the resource is negatively correlated with the influence degree of the resource, important virtual resources are preferentially mapped to physical resources with high recoverability in the initial mapping stage, so that a certain degree of survivability can be provided for the virtual network in the initial mapping stage.
5. Process (5) is based on the fault node similarity measure of the similarity function, which on the basis of processes (2) and (3) first utilizes the attributes of the physical node resources: CPU processing capacity ciStorage capacity miGeographic location liEtc. constructing a node attribute column vector
Figure GDA0002315074820000091
For any node n of backup resource pooljConstructed column vector
Figure GDA0002315074820000092
Then it communicates with the failed node nfConstructed column vector
Figure GDA0002315074820000093
Similarity between them
Figure GDA0002315074820000094
Defined as (to ensure that the similarity is not negative, take the absolute value of the similarity function):
Figure GDA0002315074820000095
wherein the content of the first and second substances,
Figure GDA0002315074820000096
Figure GDA0002315074820000097
Figure GDA0002315074820000098
Figure GDA0002315074820000099
by passing
Figure GDA00023150748200000910
And screening candidate nodes with the most similar physical characteristics with the fault node to improve the success rate of remapping. Referring specifically to fig. 3, A, B, C, D are all candidate nodes, and if n1 node fails, then candidate node A, B becomes a candidate node for n1, and SIM (n1, a)>SIM (n1, B). From the foregoing analysis, it can be seen that A should be taken as the actual substitute node for n1, since B is more similar to n2 than n1High, so it would be more reasonable to have B as a substitute node for n2 (in the event of a failure of n 2). Similarly, if n3 fails, D should be taken as its replacement node instead of the C node. Aiming at the problems of different requirements of different network environments and application scenes on network performance, different weights c 'are given to different attributes of resources during specific remapping'f＝αcf，m'f＝βmf，l'f＝λlfThe method has the advantages that the requirements for different characteristics are shown, the flexibility of the survivability mapping is increased, wherein α is larger than or equal to 1, the requirement of higher CPU processing capacity and storage capacity is shown, and the lambda is generally equal to 1, and the geographic position is better as the geographic position is closer to the geographic position of the fault node.
6. And (6) establishing a resource capacity constraint equation, a node mapping constraint equation, a flow conservation constraint equation and a variable value constraint equation on the basis of the virtual node set FN (n, x) which is influenced by the fault node x and needs to be migrated and the virtual link set FL (l, x) which is influenced by the fault node x and needs to be migrated.
① resource capability constraints
Figure GDA0002315074820000101
Figure GDA0002315074820000102
Wherein the content of the first and second substances,
Figure GDA0002315074820000103
representing a virtual node n in the kth subnetworki(niWhether e.g. FN (n, x)) remaps to physical node nyIf so, then
Figure GDA0002315074820000104
Otherwise
Figure GDA0002315074820000105
Figure GDA0002315074820000106
Indicating that all links affected by failed node x are remapped at a physical link
Figure GDA0002315074820000107
The total flow rate of the upper part of the tank,
Figure GDA0002315074820000108
representing remapped virtual links in a kth subnetwork
Figure GDA0002315074820000109
In a physical link
Figure GDA00023150748200001010
The upper occupied flow rate;
Figure GDA00023150748200001011
representing node n in a candidate pool of nodesyThe remaining resource capacity of;
Figure GDA00023150748200001012
representing links in a link candidate pool
Figure GDA00023150748200001013
The remaining bandwidth of (c).
② node mapping constraints
Figure GDA00023150748200001014
Wherein the content of the first and second substances,
Figure GDA00023150748200001015
representation of VNiThe nodes in the remapping phase map the results.
③ flow conservation constraint
Figure GDA00023150748200001016
Figure GDA00023150748200001017
Figure GDA00023150748200001018
A (x) represents a set of nodes adjacent to the physical node x. The first type shows that the link adjacent to the fault node does not participate in remapping and also shows the flow conservation; the second formula is represented by
Figure GDA00023150748200001019
As a bearer virtual link
Figure GDA0002315074820000111
Is the end point of the physical link
Figure GDA0002315074820000112
When the inflow of the node is equal to
Figure GDA0002315074820000113
Otherwise
Figure GDA0002315074820000114
When the inflow flow of the node is equal to the outflow flow so as to keep the flow conservation; the third expression represents if the virtual node niRemapping to a physical node nyThen n will beyAs a bearer virtual link
Figure GDA0002315074820000115
At the start of the physical link, node nyIs equal to
Figure GDA0002315074820000116
Otherwise node nyThe outflow rate of (a) is equal to the inflow rate to maintain flow conservation.
④ variable value constraint
Figure GDA0002315074820000117
Figure GDA0002315074820000118
Figure GDA0002315074820000119
And respectively ensuring that the candidate resource capacity is not less than the total virtual resource demand, the mapping uniqueness of the virtual resource to be remapped, the network flow conservation and the integer variable value validity.
Establishing an objective function on the basis of the above analysis and said process (5)
Figure GDA00023150748200001110
For optimizing virtual network remapping. The similarity function in the first item not only ensures that the fault resource and the candidate resource have the maximum physical characteristics, but also optimizes the connectivity of the residual candidate resource pool, and the second item represents the network flow occupied by the minimized virtual link remapping; σ, τ are weighting factors that balance the first and second terms, respectively.

Claims (3)

1. A mapping method for enhancing the survivability of a virtual network based on a divide-and-conquer strategy is characterized in that: dividing a logic sub-network based on a heuristic algorithm, constructing a candidate resource pool based on a divide-and-conquer strategy, perfecting initial mapping of a virtual network based on recoverability and remapping the virtual network based on a similar function, and the specific process comprises the following steps:
(1) a logical sub-network partitioning method based on heuristic algorithm, which is used for logically partitioning a physical network;
(2) the node candidate resource pool generation method based on the divide-and-conquer strategy is used for dividing the node resources of each sub-network into working resources and backup resources so as to provide candidate nodes for fault nodes;
(3) the method for generating the link candidate resource pool based on the divide-and-conquer strategy is used for dividing the link resources of each sub-network into working resources and backup resources so as to provide candidate links for a fault link and an affected link;
(4) the virtual network initial mapping method based on the recoverability is used for mapping important virtual resources to physical resources with high recoverability preferentially in an initial mapping stage;
(5) the similarity measurement of the fault nodes based on the similarity function is used for describing and calculating the similarity of the fault nodes and the candidate nodes so as to improve the success rate of remapping;
(6) the mixed integer programming model of the virtual network remapping is used for optimizing the virtual network remapping so as to improve the efficiency of the remapping mechanism;
said process (1) is based on a heuristic method of logical sub-network partitioning for a given physical network GsAnd a number of zones K, defining an evaluation function f (i, n) ≦ g (i, n) + h (i, n) (1 ≦ i ≦ K), wherein i represents an index of the ith sub-network, g (i, n) represents an actual cost from the initial node to the current node n, i.e., an actual step size from the initial node to the current node, i.e., a node hop count, and h (i, n) represents an estimated cost of the best path from the current node n to the destination node, i.e., a pre-generated sub-network
Figure FDA0002249919470000011
The maximum step length from the initial node is used for providing heuristic information required by the search process; selecting a path which is most likely to reach a destination node from the state space according to the valuation function, and carrying out logic partitioning on the physical network; in the process of heuristic search, once the searched node is the marked node, the node is indicated to be positioned on the boundary of the sub-network, therefore, after the heuristic search is finished, the search needs to be started again until the nodes connected with the initial node in the current sub-network are all marked to be in the accessed state, and the sub-network is pre-generated currently
Figure FDA0002249919470000012
Successfully dividing;
Figure FDA0002249919470000013
representing the ith pre-generation subnetwork;
the process (2) is a node candidate resource pool generation method based on a divide-and-conquer strategy, and each sub-network generated by the process (1)
Figure FDA0002249919470000021
According to its node backup proportion
Figure FDA0002249919470000022
And pre-generation of sub-networks
Figure FDA0002249919470000023
Node resource collection of
Figure FDA0002249919470000024
Number of computing node resource backups
Figure FDA0002249919470000025
Dividing the node resources of the sub-network into working resources and backup resources, further constructing a node candidate pool of the sub-network so as to provide candidate nodes for the fault node, and simultaneously adopting historical fault data
Figure FDA0002249919470000026
The resource backup proportion is updated, the problems of insufficient node backup resources and node backup resource redundancy are solved,
Figure FDA0002249919470000027
representing ith sub-network with N physical network nodes in T period
Figure FDA0002249919470000028
The total number of failed nodes in the set;
the process (3) is a divide-and-conquer policy-based link candidate resource pool generation method, and each pre-generation generated by the process (1) isAdult network
Figure FDA0002249919470000029
According to its link backup ratio
Figure FDA00022499194700000210
And pre-generation of sub-networks
Figure FDA00022499194700000211
Link resource set of
Figure FDA00022499194700000212
Calculating the number of link resource backups
Figure FDA00022499194700000213
Dividing the link resources of the sub-network into working resources and backup resources, and further constructing a link candidate pool of the sub-network so as to provide candidate paths for the failed link and the affected link;
the process (4) is a virtual network initial mapping method based on the restorability degree, and firstly, the node resource influence degree of the physical network to the virtual network is calculated
Figure FDA00022499194700000214
And the degree of link resource impact
Figure FDA00022499194700000215
Secondly, the restorability of the physical node resource is obtained
Figure FDA00022499194700000216
And degree of recoverability of physical link resources
Figure FDA00022499194700000217
The important virtual resources are mapped to the physical resources with high recovery degree preferentially in the initial mapping stage; with NSAnd ESSet of node resources and set of link resources representing a physical network, respectivelyN is a radical ofsRepresenting a node, i.e.
Figure FDA00022499194700000218
esIndicating a link, i.e.
Figure FDA00022499194700000219
2. The divide-and-conquer strategy-based mapping method for enhancing survivability of a virtual network according to claim 1, wherein: the process (5) is based on the fault node similarity measurement of the similarity function, and is based on the processes (2) and (3) by using ciDenotes CPU processing capability of node i, miIndicating the storage capability of node i,/iRepresenting the geographical location of a node i, the process first constructs a node attribute column vector V using attributes of the physical node resourcesi N＝(ci,mi,li)TFor any node n of the backup resource pooljConstructed column vector
Figure FDA0002249919470000031
For a failed node nfConstructed column vector
Figure FDA0002249919470000032
Further describing and calculating the similarity of the fault node resource and the candidate node resource
Figure FDA0002249919470000033
Screening candidate node resources with the most similar physical characteristics to the fault node resources to improve the success rate of remapping; aiming at the problems of different requirements of different network environments and application scenes on network performance, different weights c 'are given to different attributes of resources during specific remapping'f＝αcf，m'f＝βmf，l'f＝λlfTo represent the need for different features and to increase the flexibility of survivability mapping.
3. The divide-and-conquer strategy-based mapping method for enhancing survivability of a virtual network according to claim 1, wherein: the mixed integer programming model OF virtual network remapping in the process (6) is based on a virtual node set FN (n, x) which is influenced by a fault node x and needs to be migrated and a virtual link set FL (l, x) which is influenced by the fault node x and needs to be migrated, a resource capacity constraint equation, a node mapping constraint equation, a flow conservation constraint equation and a variable value constraint equation are established, candidate resource capacity is respectively guaranteed to be not less than total virtual resource demand, mapping uniqueness OF virtual resources to be remapped, network flow conservation and integer variable value validity are respectively guaranteed, and an objective function OF is established on the basis OF the process (5) and used for optimizing virtual network remapping so as to improve the efficiency OF a remapping mechanism.
CN201610188764.7A 2016-03-29 2016-03-29 Mapping method for enhancing survivability of virtual network based on divide-and-conquer strategy Active CN105871606B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201610188764.7A CN105871606B (en) 2016-03-29 2016-03-29 Mapping method for enhancing survivability of virtual network based on divide-and-conquer strategy

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201610188764.7A CN105871606B (en) 2016-03-29 2016-03-29 Mapping method for enhancing survivability of virtual network based on divide-and-conquer strategy

Publications (2)

Publication Number Publication Date
CN105871606A CN105871606A (en) 2016-08-17
CN105871606B true CN105871606B (en) 2020-04-10

Family

ID=56626684

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201610188764.7A Active CN105871606B (en) 2016-03-29 2016-03-29 Mapping method for enhancing survivability of virtual network based on divide-and-conquer strategy

Country Status (1)

Country Link
CN (1) CN105871606B (en)

Families Citing this family (3)

* 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
CN110191382B (en) * 2019-06-27 2020-03-27 北京邮电大学 Virtual link priority mapping method based on path sorting
JP6982601B2 (en) * 2019-07-24 2021-12-17 Kddi株式会社 Coordinated virtual network allocation method and device

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102710488A (en) * 2012-06-07 2012-10-03 北京邮电大学 Method for realizing virtual network mapping
CN103179052A (en) * 2011-12-20 2013-06-26 中国科学院声学研究所 Virtual resource allocation method and system based on proximity centrality
CN103475504A (en) * 2013-08-23 2013-12-25 北京邮电大学 Virtual network remapping method based on topology awareness
CN103856385A (en) * 2013-12-11 2014-06-11 北京邮电大学 Virtual network mapping method based on link priority

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9350632B2 (en) * 2013-09-23 2016-05-24 Intel Corporation Detection and handling of virtual network appliance failures

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103179052A (en) * 2011-12-20 2013-06-26 中国科学院声学研究所 Virtual resource allocation method and system based on proximity centrality
CN102710488A (en) * 2012-06-07 2012-10-03 北京邮电大学 Method for realizing virtual network mapping
CN103475504A (en) * 2013-08-23 2013-12-25 北京邮电大学 Virtual network remapping method based on topology awareness
CN103856385A (en) * 2013-12-11 2014-06-11 北京邮电大学 Virtual network mapping method based on link priority

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
基于Ｓｔａｃｋｅｌｂｅｒｇ博弈的网络虚拟化资源分配方法;赵夙等;《光通信研究》;20150928;63-66 *
基于网络中心性分析的虚拟网络映射算法*;王文钊等;《计算机应用研究》;20150215;第32卷(第2期);565-568 *

Also Published As

Publication number Publication date
CN105871606A (en) 2016-08-17

Similar Documents

Publication Publication Date Title
CN111147307B (en) Service function chain reliable deployment method based on deep reinforcement learning
CN103699606B (en) A kind of large-scale graphical partition method assembled with community based on summit cutting
CN112738820A (en) Dynamic deployment method and device of service function chain and computer equipment
CN105871606B (en) Mapping method for enhancing survivability of virtual network based on divide-and-conquer strategy
Zhu et al. Exploring the road to 6G: ABC—foundation for intelligent mobile networks
EP2348678B1 (en) Network topology method, device and system
CN109743215B (en) Ant colony optimization virtual network mapping method based on disaster prediction under multi-region fault
CN109446385B (en) Method for establishing network resource equipment map and using method of equipment map
CN103259744A (en) Method for mapping mobile virtual network based on clustering
CN108600101A (en) A kind of network for the optimization of end-to-end time delay performance services cross-domain method of combination
US8738559B2 (en) Graph partitioning with natural cuts
CN101321134A (en) Service quality routing selection method under dynamic network condition
CN105681153A (en) Virtual network mapping method and device
Ghosh et al. A cognitive routing framework for reliable communication in IoT for industry 5.0
Ren et al. Adaptive recovery mechanism for SDN controllers in Edge-Cloud supported FinTech applications
CN109995580A (en) VN mapping method based on GA_PSO hybrid algorithm in 5G network slice
Li et al. Dynamic SDN controller association mechanism based on flow characteristics
Zhang Storage optimization algorithm design of cloud computing edge node based on artificial intelligence technology
WO2024022185A1 (en) Route generation method and apparatus and storage medium
Zhang Reliable virtual network mapping algorithm with network characteristics and associations
Feng et al. COVE: Co-operative virtual network embedding for network virtualization
Csorba et al. Cost-efficient deployment of collaborating components
CN113392280B (en) Cross-region-oriented multi-master-model distributed graph calculation method
Fang et al. A multi-controller deployment method of SDN network based on FPGA
CN104038400B (en) A kind of mapping method of virtual network across data center

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant