CN105743763A - Virtual network mapping method based on learning and reasoning - Google Patents

Abstract

The invention provides a virtual network mapping method based on learning and reasoning. The method comprises the following steps: (1) generating a dependency matrix M between underlying physical network nodes according to historical data; (2) ordering virtual nodes according to the demands of the virtual nodes to CPU computing resource from large to small; (3) selecting an optimal underlying physical node corresponding to the current virtual node to be mapped by Bias reasoning decision according to the dependency matrix M, the mapping sequence of the virtual nodes and a topological relation of the current virtual node to be mapped and the mapped virtual nodes, and mapping the virtual node to the underlying physical node; and (4) realizing the link mapping from a virtual link between the virtual nodes on a virtual network to a physical path on an underlying physical network according to a set link mapping algorithm.

Description

A kind of mapping method of virtual network based on study and reasoning

Technical field

The present invention relates to a kind of method realizing virtual network mapping, belong to technical field of the computer network, especially belong to network virtualization technical field.

Background technology

Network virtualization refer to the infrastructure network of a shared physics is logically divided into multiple mutually isolated, heterogeneous, there is the virtual network of respective special topological structure.Virtual network is generally formed logical topology by multiple dummy nodes and a plurality of virtual link, each dummy node and every virtual link have each different demands, such as dummy node, to the requirement in the demand of CPU computing capability and residing geographical position, (dummy node needs to be mapped on the bottom physical node being within certain geographical position range, to meet certain topological structure or to provide a range of service), the virtual link demand to bandwidth ability.Service provider is by renting the virtual resource section on the bottom physical network that infrastructure provider provides, make full use of the access control of bottom physical network infrastructure, can without carrying out the related physical network hardware input when, the self-defining procotol of rapid deployment or framework (i.e. virtual network), provide diversified service to terminal use.

At upper-level virtual network in the mapping process of bottom physical network, owing to needing to meet the resource requirement of node and link simultaneously, causing that virtual network mapping problems is a NP difficult problem.How effectively and efficiently to realize the key challenge that virtual network mapping is network virtualization technical development.Corresponding virtual network mapping solution is mostly based on heuritic approach and designs at present, but current heuristic virtual network mapping scheme there are the following problems: (1) commonly uses greedy matching strategy is that each node to be mapped selects the current the highest physical node that sorts to map, and have ignored relation and the impact of dummy node and the current dummy node to be mapped previously mapped, do not consider the topology of virtual network, therefore higher bandwidth resource consumption is very easily caused, increase the degree of fragmentation of Internet resources, and ultimately result in the reduction of network mapping performance；(2) performance that node sequencing algorithm maps for virtual network, the performance mapped especially with the virtual network of greedy matching strategy has material impact, and in current node sequencing algorithm, commonly used resource assessment index is that the CPU computing capability of node self is multiplied by its adjacent link bandwidth sum criterion as a node resource ability.This resource assessment index is not comprehensively accurately, it is possible to select in node maps that CPU ability is strong but physical node that link bandwidth resource is not enough, and then causes the failure that following link maps.Therefore, in the process carrying out virtual network mapping, how better the resource capability of bottom physical network nodes to be evaluated, how with reference to mapping complete dummy node and topological structure thereof, it is achieved the optimized choice that node maps is one, current Computer Network Project field technical barrier urgently to be resolved hurrily.

Summary of the invention

In view of this, it is an object of the invention to a kind of method realizing virtual network mapping of invention, the former virtual network request historical data that success maps on bottom physical network can be utilized from statistical angle, realize the scientific evaluation to bottom physical network resource ability, and then combine the topological relation having completed maps virtual node and dummy node to be mapped, use study and inference technology to realize node and map the optimized choice with link maps.

In order to achieve the above object, the present invention proposes the mapping method of virtual network based on study and reasoning, and described method includes following operative step:

(1) according to the virtual network requests historical data set that success maps on bottom physical network, statistics generates the dependence matrix M between bottom physical network nodes, and detailed process includes following operator step:

(101) all physical nodes of bottom physical network being numbered, number and start from 1 until n, n are natural numbers, its size is equal to the physical node sum of this bottom physical network；

(102) from the historical data set that virtual network requests is successfully mapped on bottom physical network, take out each map record, all construct the node mapping matrix I of n row n row_k, wherein each element initial value is null value, and subscript k represents kth bar map record；In this map record, if the bottom physical node being numbered i was successfully mapped by a dummy node in this map record, then allow described node mapping matrix I_kThe element a of the i-th row i-th row_iiValue is updated to the CPU computational resource requirements value equal to this dummy node, and i is greater than the natural number equal to 1, less than or equal to n here；The all of node mapping matrix being previously obtained carry out Matrix Calculating and, obtain node map global matrix I, namelyWherein Q represents the size of the historical data set that described virtual network requests successfully mapped on bottom physical network；

(103) from the historical data set that virtual network requests is successfully mapped on bottom physical network, take out each map record, all construct the link maps matrix R of n row n row_k, wherein each element initial value is null value, and subscript k represents kth bar map record；In this map record, if be numbered between the bottom physical node of i and the bottom physical node being numbered j physical pathway was successfully mapped by a virtual link in this map record, then allow described link maps matrix R_kI-th row jth row element b_ijValue is the node degree numerical value jumping figure value divided by this physical pathway of physical node i, allows described link maps matrix R_kThe element b of jth row i-th row_jiValue is the node degree numerical value jumping figure value divided by this physical pathway of node j；Wherein i and j is greater than the natural number equal to 1, less than or equal to n, and i and j can not be equal；The all of link maps matrix being previously obtained carry out Matrix Calculating and, obtain link maps global matrix R, namelyWherein Q represents the size of the historical data set that described virtual network requests successfully mapped on bottom physical network；

(104) described node is mapped global matrix I and link maps global matrix R to be normalized, obtain the dependence matrix M between bottom physical network nodes；The concrete mode of normalized is: the element M of the i-th row jth row of dependence matrix M_ijValue isI in formula_ijRepresent that node maps the i-th row jth column element of global matrix I, R_ijRepresenting link maps global matrix R the i-th row jth column element, i and j is greater than the natural number equal to 1, less than or equal to n；So when i is equal to j, element M_iiIt is the node significance level value of the physical node being numbered i in bottom physical network, and when i is not equal to j, element M_ijIt it is the physical node being numbered i in the bottom physical network link tightness degree value to the physical node being numbered j；

(2) for currently waiting for the virtual network requests mapped, according to dummy node in this virtual network, CPU is calculated the demand size of resource, from big to small all dummy nodes of this virtual network being ranked up, this order is the sequencing that dummy node maps；

(3) sequencing mapped according to aforementioned generated dependence matrix M and described dummy node, topological relation according to current dummy node to be mapped with the dummy node completing mapping, adopt Bayesian inference judgement to select best bottom physical node corresponding to current dummy node to be mapped, carry out dummy node and map to the node of bottom physical node；

(4), after node maps and is all successfully completed, the link maps between physical pathway is realized in virtual network on the virtual link layer physical network on earth between dummy node according to the link maps algorithm set.

The particular content of described step 3 is to include following operator step:

(301) sequencing mapped according to described dummy node, takes out the dummy node coming foremost currently also not carrying out mapping；

(302) if this dummy node does not have father node, then find out the CPU being currently able to meet this dummy node and calculate all candidate's bottom physical nodes of resource requirement, then that bottom physical node therefrom selecting node significance level value the highest carries out node mapping, namely selects the M corresponding to candidate's bottom physical node from dependence matrix M_iiBeing worth that the highest physical node, be mapped on this physical node by this dummy node, wherein subscript i represents the numbering of described candidate's bottom physical node；Go to step (304)；

(303) if this dummy node has m father node, then first determine there be m bottom physical node of correspondence mappings relation with all father nodes of this dummy node, and find out all candidate's bottom physical nodes of the CPU calculating resource requirement being currently able to meet this dummy node, then from described candidate's bottom physical node, select that physical node making it maximum with the connection product of the link tightness degree value of the bottom physical node corresponding to m father node of this dummy node carry out node mapping, namely from described candidate's bottom physical node, select that bottom physical node W meeting following formula, this dummy node is mapped on this bottom physical node W:

$W=\arg \underset{j}{max}\underset{i=1}{\overset{m}{\Π}}{M}_{{\mathrm{jP}}^i}$

In above formula, pⁱRepresenting the numbering of the corresponding bottom physical network nodes mapped of father node of described dummy node, i value is from 1 to m；J represents described candidate's bottom physical node；M is a natural number be more than or equal to 1；The father node of described dummy node refers to dummy node that is adjacent with this dummy node in virtual network and that successfully mapped prior to this dummy node；

(304) repeating above step, until all dummy nodes complete to map, now node mapping phase completes.

The link maps method of the setting described in described step 4 refers to k shortest path k-shortestpath method.

The beneficial effects of the present invention is: make full use of the historical data that virtual network successfully maps, achieve the scientific evaluation to bottom physical network resource ability, in conjunction with considering topological relation between mapping node and node to be mapped comprehensively, the present invention is capable of node and maps the optimized choice with link maps, reduce unnecessary bandwidth resources expense, alleviate the degree of fragmentation of Internet resources, finally significantly improve include long-term average yield, long-term receptance and long-term gain expense than every virtual network mapping performance.

Accompanying drawing explanation

Fig. 1 is a kind of flow chart based on study and the mapping method of virtual network of reasoning that the present invention proposes.

Fig. 2 is the schematic diagram of bottom physical network in embodiments of the invention.

Fig. 3 is the schematic diagram of the virtual network of a success mapping in embodiments of the invention.

Fig. 4 is that in embodiments of the invention, virtual network shines schematic diagram with bottom physical network mapping pair.

Fig. 5 is a virtual network schematic diagram to be mapped in embodiments of the invention.

Fig. 6 is the node mapping process schematic diagram of virtual network to be mapped in embodiments of the invention.

Detailed description of the invention

For making the object, technical solutions and advantages of the present invention clearly, below in conjunction with accompanying drawing, the present invention is described in further detail.

Referring to Fig. 1, introducing a kind of mapping method of virtual network based on study and reasoning that the present invention proposes, described method includes following operative step:

(1) according to the virtual network requests historical data set that success maps on bottom physical network, statistics generates the dependence matrix M between bottom physical network nodes, and detailed process includes following operator step:

(101) all physical nodes of bottom physical network being numbered, number and start from 1 until n, n are natural numbers, its size is equal to the physical node sum of this bottom physical network；

It show a bottom physical network referring to Fig. 2, Fig. 2, include 6 physical nodes (in figure, circle represents physical node, the numbering that digitized representation physical node in circle is corresponding) and 8 physical links (shown in solid in figure) altogether.The numeral on each physical node side represents that the CPU that this physical node has calculates resource, in Fig. 2, physical node 1 has the CPU calculating resource of 20 units, physical node 2 has the CPU of 15 units and calculates resource, physical node 3 has the CPU of 20 units and calculates resource, physical node 4 has the CPU of 8 units and calculates resource, physical node 5 has the CPU of 10 units and calculates resource, and the CPU that physical node 6 has 30 units calculates resource.nullBetween node, the numeral in the other bracket of physical link represents the bandwidth resources that this physical link has，Here the physical link between physical node and the physical node of numbering 2 of numbering 1 has the bandwidth resources of 20 units，Physical link between physical node and the physical node of numbering 5 of numbering 1 has the bandwidth resources of 20 units，Physical link between physical node and the physical node of numbering 3 of numbering 2 has the bandwidth resources of 10 units，Physical link between physical node and the physical node of numbering 5 of numbering 2 has the bandwidth resources of 10 units，Physical link between physical node and the physical node of numbering 4 of numbering 3 has the bandwidth resources of 20 units，Physical link between physical node and the physical node of numbering 6 of numbering 3 has the bandwidth resources of 15 units，Physical link between physical node and the physical node of numbering 6 of numbering 4 has the bandwidth resources of 20 units，Physical link between physical node and the physical node of numbering 6 of numbering 5 has the bandwidth resources of 10 units.

(102) from the historical data set that virtual network requests is successfully mapped on bottom physical network, take out each map record, all construct the node mapping matrix I of n row n row_k, wherein each element initial value is null value, and subscript k represents kth bar map record；In this map record, if the bottom physical node being numbered i was successfully mapped by a dummy node in this map record, then allow described node mapping matrix I_kThe element a of the i-th row i-th row_iiValue is updated to the CPU computational resource requirements value equal to this dummy node, and i is greater than the natural number equal to 1, less than or equal to n here；

Referring to Fig. 3, Fig. 3 is the schematic diagram of the virtual network that a success maps, this virtual network includes 3 dummy nodes a, b, c that hexagon represents, the numeral on node side represents the CPU computational resource requirements size of this dummy node, namely dummy node a needs the CPU of 10 units to calculate resource, dummy node b needs the CPU of 7 units to calculate resource, and the CPU that dummy node c needs 9 units calculates resource.Numeral in the other bracket of virtual link (representing with solid line in Fig. 3) between dummy node represents the bandwidth resources demand size of this virtual link, and dummy node a, b, c virtual link between any two is required for the bandwidth resources of 8 units here.According to dummy node to the CPU demand size calculating resource, from big to small tri-dummy nodes of a, b, c being ranked up, ranking results is: a, c, b.

Assume that the virtual network shown in Fig. 3 constitutes Article 1 history map record, this virtual network is mapped to the bottom physical network shown in Fig. 2, concrete mapping situation is as shown in Figure 4, namely dummy node a is mapped to physical node 6, dummy node c is mapped to physical node 3, dummy node b is mapped in physical node 1, Fig. 4 virtual link and is represented by dotted lines, and the node mapping matrix result of calculation that then Article 1 history map record is corresponding is as follows:

$I_1=\left[\begin{array}{cccccc}7& 0& 0& 0& 0& 0\\ 0& 0& 0& 0& 0& 0\\ 0& 0& 9& 0& 0& 0\\ 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 0& 0& 10\end{array}\right]$

The all of node mapping matrix being previously obtained carry out Matrix Calculating and, obtain node map global matrix I, namelyWherein Q represents the size of the historical data set that described virtual network requests successfully mapped on bottom physical network；

(103) from the historical data set that virtual network requests is successfully mapped on bottom physical network, take out each map record, all construct the link maps matrix R of n row n row_k, wherein each element initial value is null value, and subscript k represents kth bar map record；In this map record, if be numbered between the bottom physical node of i and the bottom physical node being numbered j physical pathway was successfully mapped by a virtual link in this map record, then allow described link maps matrix R_kI-th row jth row element b_ijValue is the node degree numerical value jumping figure value divided by this physical pathway of physical node i, allows described link maps matrix R_kThe element b of jth row i-th row_jiValue is the node degree numerical value jumping figure value divided by this physical pathway of node j；Wherein i and j is greater than the natural number equal to 1, less than or equal to n, and i and j can not be equal；

The Article 1 history map record that virtual network according to above figure 3 is constituted, and this virtual network according to Fig. 4 is mapped to the concrete mapping situation of the bottom physical network shown in Fig. 2, namely virtual link a-b is mapped to physical pathway 6-5-1, virtual link b-c is mapped to physical pathway 1-2-3, virtual link c-a is mapped to physical pathway 3-6, and the link maps matrix calculation result that then Article 1 history map record is corresponding is as follows:

$R_1=\left[\begin{array}{cccccc}0& 0& \frac{2}{2}& 0& 0& \frac{2}{2}\\ 0& 0& 0& 0& 0& 0\\ \frac{3}{2}& 0& 0& 0& 0& \frac{3}{1}\\ 0& 0& 0& 0& 0& 0\\ 0& 0& 0& 0& 0& 0\\ \frac{3}{2}& 0& \frac{3}{1}& 0& 0& 0\end{array}\right]$

The all of link maps matrix being previously obtained carry out Matrix Calculating and, obtain link maps global matrix R, namelyWherein Q represents the size of the historical data set that described virtual network requests successfully mapped on bottom physical network；

(104) described node is mapped global matrix I and link maps global matrix R to be normalized, obtain the dependence matrix M between bottom physical network nodes；The concrete mode of normalized is: the element M of the i-th row jth row of dependence matrix M_ijValue isI in formula_ijRepresent that node maps the i-th row jth column element of global matrix I, R_ijRepresenting link maps global matrix R the i-th row jth column element, i and j is greater than the natural number equal to 1, less than or equal to n；So when i is equal to j, element M_iiIt is the node significance level value of the physical node being numbered i in bottom physical network, and when i is not equal to j, element M_ijIt it is the physical node being numbered i in the bottom physical network link tightness degree value to the physical node being numbered j；

(2) for currently waiting for the virtual network requests mapped, according to dummy node in this virtual network, CPU is calculated the demand size of resource, from big to small all dummy nodes of this virtual network being ranked up, this order is the sequencing that dummy node maps；

(3) sequencing mapped according to aforementioned generated dependence matrix M and described dummy node, topological relation according to current dummy node to be mapped with the dummy node completing mapping, adopt Bayesian inference judgement to select best bottom physical node corresponding to current dummy node to be mapped, carry out dummy node and map to the node of bottom physical node；

(4), after node maps and is all successfully completed, the link maps between physical pathway is realized in virtual network on the virtual link layer physical network on earth between dummy node according to the link maps algorithm set.

The particular content of described step 3 is to include following operator step:

(301) sequencing mapped according to described dummy node, takes out the dummy node coming foremost currently also not carrying out mapping；

(302) if this dummy node does not have father node, then find out the CPU being currently able to meet this dummy node and calculate all candidate's bottom physical nodes of resource requirement, then that bottom physical node therefrom selecting node significance level value the highest carries out node mapping, namely selects the M corresponding to candidate's bottom physical node from dependence matrix M_iiBeing worth that the highest physical node, be mapped on this physical node by this dummy node, wherein subscript i represents the numbering of described candidate's bottom physical node；Go to step (304)；

(303) if this dummy node has m father node, then first determine there be m bottom physical node of correspondence mappings relation with all father nodes of this dummy node, and find out all candidate's bottom physical nodes of the CPU calculating resource requirement being currently able to meet this dummy node, then from described candidate's bottom physical node, select that physical node making it maximum with the connection product of the link tightness degree value of the bottom physical node corresponding to m father node of this dummy node carry out node mapping, namely from described candidate's bottom physical node, select that bottom physical node W meeting following formula, this dummy node is mapped on this bottom physical node W:

$W=\arg \underset{j}{max}\underset{i=1}{\overset{m}{\Π}}{M}_{{\mathrm{jP}}^i}$

In above formula, pⁱRepresenting the numbering of the corresponding bottom physical network nodes mapped of father node of described dummy node, i value is from 1 to m；J represents described candidate's bottom physical node；M is a natural number be more than or equal to 1；The father node of described dummy node refers to dummy node that is adjacent with this dummy node in virtual network and that successfully mapped prior to this dummy node；

Referring to Fig. 5, it it is a virtual network to be mapped shown in Fig. 5, this virtual network includes 4 dummy node a, b, c and the d that hexagon represents, the numeral on node side represents the CPU computational resource requirements size of this dummy node, namely dummy node a needs the CPU of 6 units to calculate resource, dummy node b needs the CPU of 8 units to calculate resource, and dummy node c needs the CPU of 10 units to calculate resource, and the CPU that dummy node d needs 5 units calculates resource.In Fig. 5, the virtual link omission between dummy node does not mark.

The node mapping process of the virtual network to be mapped shown in Fig. 5 is illustrated referring to Fig. 6, Fig. 6, specific as follows:

The first step such as dummy node c shown in Fig. 6 (a) has the highest CPU computational resource requirements, sort and the highest first map, now it does not have father node, therefore first find out the CPU being currently able to meet this dummy node and calculate all candidate's bottom physical nodes of resource requirement, then therefrom select that physical node (that assumes this physical node is numbered x) that its node significance level value corresponding in dependence matrix M is the highest, dummy node c is mapped on this physical node；

Second step such as dummy node b shown in Fig. 6 (b) has second highest CPU computational resource requirements, ranked second position to map, now it is also without father node, therefore first find out the CPU being currently able to meet this dummy node and calculate all candidate's bottom physical nodes of resource requirement, then therefrom select that physical node (that assumes this physical node is numbered y) that its significance level value corresponding in dependence matrix M is the highest, dummy node b is mapped on this physical node；

3rd step such as dummy node a shown in Fig. 6 (c) and then maps, and its father node is dummy node c and b, and the bottom physical node that therefore dummy node a maps is one and is currently able to meet its CPU calculating resource requirement, and meetsPhysical node, that namely assumes this physical node is numbered z；

4th step such as dummy node d shown in Fig. 6 (d) finally maps, and its father node is dummy node a, and the bottom physical node that therefore dummy node d maps is one and is currently able to meet its CPU calculating resource requirement, and meetsPhysical node.

(304) repeating above step, until all dummy nodes complete to map, now node mapping phase completes.

The link maps method of the setting described in described step 4 refers to k shortest path k-shortestpath method.

Method proposed by the invention has been carried out a large amount of emulation experiment by inventor.In emulation experiment, we use general topology generator GT-ITM software building one to have the physical network of 100 nodes and about 500 links and all virtual network requests.It is all be connected by link with the probability of 50% between each two physical node.In physical network, the CPU of node calculates the bandwidth resources numerical value of resource and link and obeys being uniformly distributed of [50,100].The arrival rate of each virtual network mapping request is obeyed every 100 unit of time and is arrived the Poisson distribution of 5 virtual network mapping request.Each virtual network comprises [5,20] and is uniformly distributed individual dummy node, and the virtual link connection probability between each two dummy node is 50% equally.The CPU of virtual network interior joint calculates the bandwidth resources numerical value of resource and link and obeys being uniformly distributed of [1,50].The persistent period of each virtual network requests meets the exponential of average out to 500 unit of time.In each emulation experiment, we allow bottom physical network accept about 2500 virtual network mapping request (being about 50000 unit of time), and using the arithmetic mean of instantaneous value of ten simulation results as last the simulation experiment result.The method of the results show present invention is effective, can significantly improve include long-term average yield, long-term receptance and long-term gain expense than every virtual network mapping performance.

Claims (3)

1. the mapping method of virtual network based on study and reasoning, it is characterised in that: described method includes following operative step:

(1) according to the virtual network requests historical data set that success maps on bottom physical network, statistics generates the dependence matrix M between bottom physical network nodes, and detailed process includes following operator step:

(101) all physical nodes of bottom physical network being numbered, number and start from 1 until n, n are natural numbers, its size is equal to the physical node sum of this bottom physical network；

(102) from the historical data set that virtual network requests is successfully mapped on bottom physical network, take out each map record, all construct the node mapping matrix I of n row n row_k, wherein each element initial value is null value, and subscript k represents kth bar map record；In this map record, if the bottom physical node being numbered i was successfully mapped by a dummy node in this map record, then allow described node mapping matrix I_kThe element a of the i-th row i-th row_iiValue is updated to the CPU computational resource requirements value equal to this dummy node, and i is greater than the natural number equal to 1, less than or equal to n here；The all of node mapping matrix being previously obtained carry out Matrix Calculating and, obtain node map global matrix I, namelyWherein Q represents the size of the historical data set that described virtual network requests successfully mapped on bottom physical network；

(103) from the historical data set that virtual network requests is successfully mapped on bottom physical network, take out each map record, all construct the link maps matrix R of n row n row_k, wherein each element initial value is null value, and subscript k represents kth bar map record；In this map record, if be numbered between the bottom physical node of i and the bottom physical node being numbered j physical pathway was successfully mapped by a virtual link in this map record, then allow described link maps matrix R_kI-th row jth row element b_ijValue is the node degree numerical value jumping figure value divided by this physical pathway of physical node i, allows described link maps matrix R_kThe element b of jth row i-th row_jiValue is the node degree numerical value jumping figure value divided by this physical pathway of node j；Wherein i and j is greater than the natural number equal to 1, less than or equal to n, and i and j can not be equal；The all of link maps matrix being previously obtained carry out Matrix Calculating and, obtain link maps global matrix R, namelyWherein Q represents the size of the historical data set that described virtual network requests successfully mapped on bottom physical network；

(104) described node is mapped global matrix I and link maps global matrix R to be normalized, obtain the dependence matrix M between bottom physical network nodes；The concrete mode of normalized is: the element M of the i-th row jth row of dependence matrix M_ijValue isI in formula_ijRepresent that node maps the i-th row jth column element of global matrix I, R_ijRepresenting link maps global matrix R the i-th row jth column element, i and j is greater than the natural number equal to 1, less than or equal to n；So when i is equal to j, element M_iiIt is the node significance level value of the physical node being numbered i in bottom physical network, and when i is not equal to j, element M_ijIt it is the physical node being numbered i in the bottom physical network link tightness degree value to the physical node being numbered j；

(2) for currently waiting for the virtual network requests mapped, according to dummy node in this virtual network, CPU is calculated the demand size of resource, from big to small all dummy nodes of this virtual network being ranked up, this order is the sequencing that dummy node maps；

(3) sequencing mapped according to aforementioned generated dependence matrix M and described dummy node, topological relation according to current dummy node to be mapped with the dummy node completing mapping, adopt Bayesian inference judgement to select best bottom physical node corresponding to current dummy node to be mapped, carry out dummy node and map to the node of bottom physical node；

(4), after node maps and is all successfully completed, the link maps between physical pathway is realized in virtual network on the virtual link layer physical network on earth between dummy node according to the link maps algorithm set.

2. a kind of mapping method of virtual network based on study and reasoning according to claim 1, it is characterised in that: the particular content of described step 3 is to include following operator step:

(301) sequencing mapped according to described dummy node, takes out the dummy node coming foremost currently also not carrying out mapping；

(302) if this dummy node does not have father node, then find out the CPU being currently able to meet this dummy node and calculate all candidate's bottom physical nodes of resource requirement, then that bottom physical node therefrom selecting node significance level value the highest carries out node mapping, namely selects the M corresponding to candidate's bottom physical node from dependence matrix M_iiBeing worth that the highest physical node, be mapped on this physical node by this dummy node, wherein subscript i represents the numbering of described candidate's bottom physical node；Go to step (304)；

(303) if this dummy node has m father node, then first determine there be m bottom physical node of correspondence mappings relation with all father nodes of this dummy node, and find out all candidate's bottom physical nodes of the CPU calculating resource requirement being currently able to meet this dummy node, then from described candidate's bottom physical node, select that physical node making it maximum with the connection product of the link tightness degree value of the bottom physical node corresponding to m father node of this dummy node carry out node mapping, namely from described candidate's bottom physical node, select that bottom physical node W meeting following formula, this dummy node is mapped on this bottom physical node W:

$W=\arg \underset{j}{max}\underset{i=1}{\overset{m}{\Π}}{M}_{{\mathrm{jP}}^i}$

In above formula, pⁱRepresenting the numbering of the corresponding bottom physical network nodes mapped of father node of described dummy node, i value is from 1 to m；J represents described candidate's bottom physical node；M is a natural number be more than or equal to 1；The father node of described dummy node refers to dummy node that is adjacent with this dummy node in virtual network and that successfully mapped prior to this dummy node；

(304) repeating above step, until all dummy nodes complete to map, now node mapping phase completes.

3. a kind of mapping method of virtual network based on study and reasoning according to claim 1, it is characterised in that: the link maps method of the setting described in described step 4 refers to k shortest path k-shortestpath method.