CN105262663A - Cross-domain mapping method for hybrid virtual network (HVN) - Google Patents

Abstract

The invention discloses a cross-domain mapping method for a hybrid virtual network (HVN). In the event of having one known request of a bottom multi-domain network and the HVN, node calculation resources and link bandwidth resources in the bottom network are distributed to the HVN request based on an HVNMMD-D algorithm and a spectrogram decomposition-based mapping method; therefore, a mapping scheme, which not only satisfies the QoS (Quality of Service) but also saves the mapping cost, is found; secondly, the specificities of multicast parts in the bottom multi-domain network and the HVN are comprehensively considered; common bandwidth resources and calculation resources are optimized; and thus, cross-domain mapping development requirements of the current HVN can be satisfied.

Description

A kind of cross-domain mapping method mixing virtual network

Technical field

The invention belongs to Internet communication technology field, more specifically say, relate to a kind of cross-domain mapping method mixing virtual network.

Background technology

The Internet has achieved huge achievement as the acquisition of information of the world today and the effective means of exchange.In the past few decades, the development of network technology and the multifarious increase of internet, applications make network environment become better and better, and confirm that it is worth.But the fast development of the Internet also causes very large burden to network, the existing network architecture is made can not well to carry so many application.And due to the multi-provider characteristic of network, it is difficult that each supplier of the coordinating and unifying does new adjustment to bottom-layer network structure.Therefore, network of today can only structure can only carry out slowly and simply changing by the restriction of multi-provider, can not carry out reforming fast on a large scale.

Network virtualization technology, as the effective means solving the Internet bottleneck problem, is subject to the extensive concern of scholar and enterprise recently.Network virtualization technology allow share network bottom layer deploy isomery network application and do not need changed network bottom architecture.Therefore enrich network application, maintain the overall architecture of existing network simultaneously.The essence of network virtualization is exactly can the virtual resource of dynamic assignment by being abstracted into by the computational resource of bottom physical network nodes and the bandwidth resources of link, and each network application of distributing to of isolation makes different virtual network frameworks can share the physical network resource of bottom.Under network virtualization environment, resource has following features following characteristics:

(1), isomerism: the virtual resource in network virtualization environment is of a great variety, Various Functions, and the difference between the operation of access configuration mode, local management system, shared rule is very large.

(2), distributivity: the difference that virtual resource distributes on position is in the ground local, is under the jurisdiction of multiple infrastructure supplier.

(3), autonomy: infrastructure supplier, as the owner of virtual resource, has five-star administration authority to resource, and has autonomous managerial ability.

(4), extensibility: on the one hand, due to the needs of Facilities Construction, existing infrastructure provider can increase new network equipment resource, expands network size; On the other hand, due to the demand of business development, virtual net may need new infrastructure provider to be included in management framework.

(5), dynamic: in network virtualization environment, virtual resource can dynamically add or leave system, As time goes on the dynamic change such as present position, service provision capacity, load of resource, also likely occurs that equipment physical fault causes the situation of resource inaccessible.

Extensively sharing of bottom-layer network resource is the main purpose of network virtualization.Therefore the mapping of virtual network resource is the emphasis of network virtualization technology is also difficult point.Virtual resource mapping algorithm is as one of the key issue of network virtualization technology, it achieve the process reasonably mapped to by the virtual network requests of user on the physical resource of bottom physical network, wherein how efficient allocation physical network resource, to meet link bandwidth and the joint behavior requirement of each virtual network, is the key of virtual network mapping problems.

In existing network application, communication form has the forms such as clean culture, multicast, broadcast, mixing virtual network.Wherein clean culture and multicast are widely used in the application of the real-time of many needs high QoS, but in virtual network, existing most virtual resource mapping algorithm only to unicast service or multicast service effective.Single unicast or multicast mapping algorithm is not also suitable for the situation mixing virtual network, simultaneously the resource of a bottom physical network supplier does not likely carry whole virtual network requests or needs several physical network suppliers cooperation to dispose virtual network requests due to position constraint, therefore needs the consideration carrying out multiple domain.

Summary of the invention

The object of the invention is to overcome the deficiencies in the prior art, a kind of cross-domain mapping method mixing virtual network is provided, when a known bottom multiple-domain network and mixing virtual network HVN request, find out one and meet QoS, and reduce the mapping scheme mapping cost.

For achieving the above object, a kind of mixing virtual network cross-domain mapping method based on HVNMMD-D algorithm of the present invention, is characterized in that, comprise the following steps:

(1) HVN request, is split

According to the attribute of dummy node and virtual link in mixing virtual network, HVN request is split into multicast network request MVN and unicast networks request UVN;

(2), the cross-domain mapping of multicast network request MVN

(2.1), the mapping cost of MVN in each physical domain is estimated

(2.1.1) mapping cost in the territory, estimating multicast root node

Cost(v→n ^s)＝(cst(n ^s)+(MC-Con(n ^s)))*req(v)(1)

Wherein, Con (n ^s) be physical node n ^sadjacent node number, MC is Con (n maximum in all nodes of bottom-layer network in territory ^s), req (v) is MVN root node v resource requirement quantity, cst (n ^s) be physical node n ^sresource unit price;

According to formula (1), choose the candidate mappings node of physical node as MVN root node v of cost minimization;

(2.1.2) mapping cost in the territory, estimating multicast leaf node

Cost(v _L→n ^k)+Cost(p)(2)

Wherein

Cost(v _L→n ^k)＝cst(n ^k)*req(v _L)(3)

$Cost\left(p\right)=\underset{l^s\∈p}{\Σ}cst\left(l^s\right)*req\left(l^v\right)---\left(4\right)$

Wherein, cst (n ^k) be physical node n ^kresource unit price, req (v _l) be multicast leaf node v _lresource requirement quantity, Cost (v _l→ n ^k) be multicast leaf node v _lat physical node n ^kon mapping cost;

L ^vfor connecting multicast root node v and multicast leaf node v _lvirtual link, p is virtual link l ^vmapping path in physical network, cst (l ^s) be physical link l ^sresource unit price, req (l ^v) be virtual link resource requirement, Cost (p) is virtual link l ^vmapping cost;

According to formula (2), choose the physical node n mapping least cost ^sas MVN leaf node v _lcandidate mappings node, the path p between corresponding node is virtual link l ^vcandidate mappings path;

(2.2), the cross-domain mapping of multicast network request MVN

(2.2.1), global view is built to multicast network

First initialization overall situation figure is sky, then the both candidate nodes in each territory and path candidate is added in overall figure, and increases cross-domain path candidate according to the both candidate nodes in not same area according to the virtual link annexation in virtual network;

(2.2.2), optimization modeling is carried out according to global view

Optimization aim is: $Min\left\{\underset{n^d\∈N_M^D}{\Σ}x\left(n^d\right)*f\left(n^d\right)+\underset{l^d\∈E_M^D}{\Σ}y\left(l^d\right)*f\left(l^d\right)\right\}---\left(5\right)$

Constraints is: $\left\{\begin{array}{c}\underset{n^d:cad\left(n^d\right)=n_m^v}{\Σ}x\left(n^d\right)=1,\∀n_m^v\∈N_M^V\\ \underset{l^d:cad\left(l^d\right)=l_m^v}{\Σ}y\left(l^d\right)=1,\∀l_m^v\∈E_M^V\\ y\left(l^d\right)\≤x\left(n^d\right),\∀l^d=(n^d,m^d)\\ x\left(n^d\right)\∈\{0,1\},\∀n^d\∈N_M^D\\ y\left(l^d\right)\∈\{0,1\},\∀l^d\∈E_M^D\end{array}\right.---\left(6\right)$

Wherein, l ^dfor the path candidate in overall figure, n ^d, m ^dfor l ^dtwo end points.X (n ^d) be binary variable, as both candidate nodes n ^dbeing selected, is 1, otherwise is 0, y (l ^d) be binary variable, as path candidate l ^dbeing selected, is 1, otherwise is 0, cad (n ^d) for being mapped in n ^don dummy node, cad (l ^d) for being mapped in l ^don virtual link, f (n ^d) be the mapping cost f (n of both candidate nodes ^d)=Cost (cad (n ^d) → n ^d), f (l ^d) be the mapping cost f (l of path candidate ^d)=Cost (l ^d), for multicast virtual node, for multicast virtual link, for the both candidate nodes set in overall figure, for the alternative path set in overall figure, for multicast node set, for multicast link set;

(2.2.3), model solution

First to the variable y (l in model ^d), x (n ^d) carry out integer and relax, recycling cplex solves, and in the result solved, traversal all-multicast dummy node is each multicast virtual sensor selection problem x (n ^d) maximum both candidate nodes n ^das final mapping node, select n ^dbetween path candidate be final mapping path, and carry out the cross-domain mapping of multicast network request MVN with this;

(3), the cross-domain mapping of unicast networks request UVN

(3.1), unmapped clean culture dummy node v in UVN is estimated _aterritory in map cost

Unmapped clean culture dummy node v is chosen from UVN _a, then estimate clean culture dummy node v _aterritory in map cost;

Cost(v _A→n ^k)+Cost(D)(7)

Wherein

$Cost\left(D\right)=\underset{p\∈P}{\Σ}Cost\left(p\right)---\left(8\right)$

Cost (v _a→ n ^k) can calculate according to formula (2);

Wherein, Cost (D) is clean culture dummy node v _aand the mapping being mapped in virtual link set between the dummy node in Same Physical territory in unicast networks spends, P is the mapping path set that virtual link is integrated into physical network, and p is the paths in P;

The physical node n of minimum mapping cost is chosen according to formula (7) ^kas clean culture dummy node v _aboth candidate nodes, the path between corresponding node is the path candidate of virtual link;

(3.2), the cross-domain mapping of unicast networks request UVN

(3.2.1), global view is built to unicast networks

First initialization overall situation figure is empty, again the both candidate nodes in each territory selected in step (3.1) and path candidate are added in overall figure, and increase cross-domain path candidate according to the both candidate nodes in not same area according to the virtual link annexation in virtual network;

(3.2.2), optimization modeling is carried out according to global view

Optimization aim is: $Min\left\{\underset{n^d\∈N_U^D}{\Σ}x\left(n^d\right)*f\left(n^d\right)+\underset{l^d\∈E_U^D}{\Σ}y\left(l^d\right)*f\left(l^d\right)\right\}---\left(9\right)$

Constraints is: $\left\{\begin{array}{c}\underset{n^d:cad\left(n^d\right)=n_u^v}{\Σ}x\left(n^d\right)=1,\∀n_u^v\∈N_U^V\\ \underset{l^d:cad\left(l^d\right)=l_u^v}{\Σ}y\left(l^d\right)=1,\∀l_u^v\∈E_U^V\\ y\left(l^d\right)\≤x\left(n^d\right),\∀l^d=(n^d,m^d)\\ x\left(n^d\right)\∈\{0,1\},\∀n^d\∈N_U^D\\ y\left(l^d\right)\∈\{0,1\},\∀l^d\∈E_U^D\end{array}\right.---\left(10\right)$

Wherein, l ^dfor the path candidate in overall figure, n ^d, m ^dfor the both candidate nodes in overall figure and be l ^dtwo end points, x (n ^d) be binary variable, as both candidate nodes n ^dbeing selected, is 1, otherwise is 0, y (l ^d) be binary variable, as path candidate l ^dbeing selected, is 1, otherwise is 0, cad (n ^d) for being mapped in n ^don dummy node, cad (l ^d) for being mapped in l ^don virtual link, f (n ^d) be the mapping cost f (n of both candidate nodes ^d)=Cost (cad (n ^d) → n ^d), f (l ^d) be the mapping cost f (l of path candidate ^d)=Cost (l ^d), for clean culture dummy node, for clean culture virtual link, for the node set in overall figure, for the set of paths in overall figure. for unicast networks node set, for unicast networks link set;

(3.2.3), model solution

First to the variable y (l in model ^d), x (n ^d) carry out integer and relax, recycling cplex solves, and in the result solved, travels through all clean culture dummy nodes, is that each clean culture dummy node selects x (n ^d) maximum both candidate nodes n ^das final mapping node, select n ^dbetween path candidate be final mapping path, and carry out the cross-domain mapping of unicast networks request UVN with this.

Further, the present invention also provides a kind of mixing virtual network cross-domain mapping method based on pattern decomposition, it is characterized in that, comprises the following steps:

(1), the spectrum segmentation of HVN request

(1.1), adjacency matrix is set up to HVN request

If A _{n × n}for the adjacency matrix that single HVN asks, wherein n is dummy node number in HVN request, and carry nodal community in each dummy node, i.e. multicast or clean culture, A (i, j) represents the virtual link bandwidth demand between i-th dummy node and a jth dummy node, wherein, i, j ∈ [1, n];

(1.2), matrix D is built _{n × n}, and meet:

$\left\{\begin{array}{c}D(i,j)=\underset{j}{\Σ}A\left(i,j\right),\left(i=j\right)\\ D(i,j)=0,(i\≠j)\end{array}\right.---\left(11\right)$

(1.3), matrix B is built _{n × n}, B _{n × n}=D _{n × n}-A _{n × n};

(1.4), compute matrix B _{n × n}front k characteristic value and characteristic of correspondence vector, wherein, the maximum occurrences of k is matrix B _{n × n}the total number of characteristic value; A feature space is formed again by characteristic vector, each vector in feature space represents some dummy nodes by the position that it occurs in characteristic vector, to the vectorial K-means cluster in feature space, cluster acquired results is exactly the dummy node set of each subgraph;

(2) candidate domain of subgraph, is chosen

(2.1), subgraph is sorted

The attribute of all dummy nodes in traversal subgraph, made number one by the subgraph containing multicast root node, the subgraph containing multicast leaf node comes thereafter successively, finally comes finally successively by the subgraph only containing clean culture node;

(2.2), estimate subgraph territory in mapping cost

If containing multicast root node in subgraph, then estimate its cost according to formula (1), then choose map Least-cost physical node as mapping node in the territory of multicast root node;

If containing multicast leaf node in subgraph, then estimate its territory according to formula (2) and map link maps cost in cost, territory, then the physical node choosing mapping Least-cost is as mapping node in the territory of multicast leaf node;

If containing clean culture dummy node in subgraph, first calculate in its territory according to formula (7) and map cost, recycling formula (10) calculates current subgraph virtual unicast node v _uand the link maps cost between the subgraph mapped.

$Cost\left(M\right)=\underset{p\∈L}{\Σ}Cost\left(p\right)---\left(10\right)$

Wherein L is and node v _ulink set between the subgraph having an annexation.

Then dummy node v _utotal mapping cost be

Const(v _u→n ^k)+Cost(M)+Cost(D)(11)

The physical node of mapping cost minimization is chosen as mapping node in the territory of clean culture dummy node according to formula (11).

(3), the cross-domain mapping of subgraph

According to the mapping process of step (2.2), calculate the mapping cost sum of all nodes in each subgraph respectively, in each subgraph, choose the final mapping territory of territory as this subgraph of mapping cost minimization respectively, corresponding mapping node and mapping path are final mapping node and path, thus complete the cross-domain mapping of all subgraphs.

Goal of the invention of the present invention is achieved in that

A kind of cross-domain mapping method mixing virtual network, when a known bottom multiple-domain network and mixing virtual network (HVN) are asked, by based on HVNMMD-D algorithm and the mapping method based on pattern decomposition, distribute node calculate resource in bottom-layer network and link bandwidth resource to mixing virtual network requests, thus have found and a kind ofly not only meet QoS but also save the mapping scheme mapping cost.Secondly, considered the particularity of multicast portions in multiple domain bottom-layer network and hybrid network in the present invention, common bandwidth resources and computational resource have been done to the development need optimizing to meet when forward slip value virtual network cross-domain mapping.

Meanwhile, the cross-domain mapping method that the present invention mixes virtual network also has following beneficial effect:

(1), applied widely

Traditional multiple domain mapping policy is propose for a kind of situation in clean culture virtual network requests or multicast virtual network request mostly, cannot be applicable to the situation mixing virtual network requests under normal circumstances; And this method is not only applicable to clean culture virtual network requests and the simultaneous mixing virtual network requests of multicast virtual network request, and be applicable to the service request of clean culture virtual network requests or multicast virtual network request.Therefore, compared with traditional mapping policy, the scope of application of this method is wider.

(2), resource utilization is high

Because the present invention is when mapping HVN request, multicast portions in preferential mapping HVN, and in the mapping process guiding root node, employ the number of degrees of bottom-layer network node, the surplus resources of node and resource unit price as inducible factor simultaneously, bottom-layer network load so just can be made more balanced, and accept more HVN and ask, add the utilance of bottom-layer network resource, namely add income.

Accompanying drawing explanation

Fig. 1 is the fractionation schematic diagram of HVN request;

Fig. 2 is the multicast network request mapping schematic diagram based on HVNMMD-D algorithm;

Fig. 3 is the unicast networks request mapping schematic diagram based on HVNMMD-D algorithm;

Fig. 4 is the HVN request mapping schematic diagram based on pattern decomposition.

Embodiment

Below in conjunction with accompanying drawing, the specific embodiment of the present invention is described, so that those skilled in the art understands the present invention better.Requiring particular attention is that, in the following description, when perhaps the detailed description of known function and design can desalinate main contents of the present invention, these are described in and will be left in the basket here.

Embodiment

For convenience of description, first the relevant speciality term occurred in embodiment is described:

HVNMMD-D (decomposition-basedalgorithmforhybridvirtualnetworkmappi ngacrossmultipledomains) is based on the mixing virtual network mapping algorithm decomposing thought;

HVN (hybridvirtualnetwork) mixes virtual network;

MVN (multicastvirtualnetwork) multicast virtual network;

UVN (unicastvirtualnetwork) clean culture virtual network;

In the present invention, a kind of cross-domain mapping method mixing virtual network mainly comprises two kinds, that is: based on HVNMMD-D algorithm and the mixing virtual network cross-domain mapping method based on pattern decomposition, at this, we first set forth the dummy node in two kinds of algorithms: dummy node comprises multicast virtual node and clean culture dummy node, and multicast virtual node comprises again multicast root node and multicast leaf node.Below two kinds of methods are described in detail.

One, based on the mixing virtual network cross-domain mapping method of HVNMMD-D algorithm, its concrete steps implemented comprise:

S1, fractionation HVN request

According to the attribute of dummy node and virtual link in mixing virtual network, HVN request is split into multicast network request MVN and unicast networks request UVN;

In the present embodiment, as shown in Fig. 1 (b), node a, b, c and inter-node link (a, b), (a, c) have the characteristic of unicast networks and multicast network simultaneously, are multicast network request MVN; As shown in Fig. 1 (c), remaining part only has unicast networks characteristic, is unicast networks request UVN; As shown in Fig. 1 (a), jointly form mixing virtual network by multicast network request MVN and unicast networks request UVN.

The cross-domain mapping of S2, multicast network request MVN

S2.1, first MVN is carried out preliminary mapping in each physical domain

Cost is mapped in the territory of S2.1.1, estimation multicast root node

Cost(v→n ^s)＝(cst(n ^s)+(MC-Con(n ^s)))*req(v)(1)

Wherein, Con (n ^s) be physical node n ^sadjacent node number, MC is Con (n maximum in all nodes of bottom-layer network in territory ^s), req (v) is MVN root node v resource requirement quantity, cst (n ^s) be physical node n ^sresource unit price;

According to formula (1), choose the candidate mappings node of physical node as MVN root node v of cost minimization;

In the present embodiment, as shown in Fig. 2 (a), choose three physical domain and carry out preliminary mapping, by the mapping cost of said method estimation multicast root node a in three physical domain, as shown in Fig. 2 (c), in territory 1, territory 2, territory 3, find a candidate mappings node respectively;

Cost is mapped in the territory of S2.1.2, estimation multicast leaf node

Cost(v _L→n ^k)+Cost(p)(2)

Wherein

Cost(v _L→n ^k)＝cst(n ^k)*req(v _L)(3)

$Cost\left(p\right)=\underset{l^s\∈p}{\Σ}cst\left(l^s\right)*req\left(l^v\right)---\left(4\right)$

Wherein, cst (n ^k) be physical node n ^kresource unit price, req (v _l) be multicast leaf node v _lresource requirement quantity, Cost (v _l→ n ^k) be multicast leaf node v _lat physical node n ^kon mapping cost;

L ^vfor connecting multicast root node v and multicast leaf node v _lvirtual link, p is virtual link l ^vmapping path in physical network, cst (l ^s) be physical link l ^sresource unit price, req (l ^v) be virtual link resource requirement, Cost (p) is virtual link l ^vmapping cost;

According to formula (2), choose the physical node n mapping least cost ^sas MVN leaf node v _lcandidate mappings node, the path p between corresponding node is virtual link l ^vcandidate mappings path;

In the present embodiment, estimate multicast leaf node b, c mapping cost in three physical domain in Fig. 2 (b) according to the method described above respectively, as shown in Fig. 2 (c), be respectively multicast leaf node b, c and find a candidate mappings node in territory 1, territory 2, territory 3, the path between corresponding node is candidate mappings path;

The cross-domain mapping of S2.2, multicast network request MVN

S2.2.1, to multicast network build global view

First initialization overall situation figure is sky, then the both candidate nodes in each territory and path candidate is added in overall figure, and increases cross-domain path candidate according to the both candidate nodes in not same area according to the virtual link annexation in virtual network;

In this embodiment, according to said method, first the both candidate nodes in Fig. 2 (c) and link are added in overall figure, wherein, both candidate nodes in 1 ~ territory, territory 3 corresponding points in overall figure are a1 ~ a3, other both candidate nodes also do identical conversion, then increase cross-domain path candidate according to the annexation of 2 (b) interior joint, and its result is as shown in Fig. 2 (d);

S2.2.2, carry out optimization modeling according to global view

Optimization aim is: $Min\left\{\underset{n^d\∈N_M^D}{\Σ}x\left(n^d\right)*f\left(n^d\right)+\underset{l^d\∈E_M^D}{\Σ}y\left(l^d\right)*f\left(l^d\right)\right\}---\left(5\right)$

Constraints is: $\left\{\begin{array}{c}\underset{n^d:cad\left(n^d\right)=n_m^v}{\Σ}x\left(n^d\right)=1,\∀n_m^v\∈N_M^V\\ \underset{l^d:cad\left(l^d\right)=l_m^v}{\Σ}y\left(l^d\right)=1,\∀l_m^v\∈E_M^V\\ y\left(l^d\right)\≤x\left(n^d\right),\∀l^d=(n^d,m^d)\\ x\left(n^d\right)\∈\{0,1\},\∀n^d\∈N_M^D\\ y\left(l^d\right)\∈\{0,1\},\∀l^d\∈E_M^D\end{array}\right.---\left(6\right)$

Wherein, l ^dfor the path candidate in overall figure, n ^d, m ^dfor l ^dtwo end points.X (n ^d) be binary variable, as both candidate nodes n ^dbeing selected, is 1, otherwise is 0, y (l ^d) be binary variable, as path candidate l ^dbeing selected, is 1, otherwise is 0, cad (n ^d) for being mapped in n ^don dummy node, cad (l ^d) for being mapped in l ^don virtual link, f (n ^d) be the mapping cost f (n of both candidate nodes ^d)=Cost (cad (n ^d) → n ^d), f (l ^d) be the mapping cost f (l of path candidate ^d)=Cost (l ^d), for multicast virtual node, for multicast virtual link, for the both candidate nodes set in overall figure, for the alternative path set in overall figure, for multicast node set, for multicast link set;

S2.2.3, model solution

First to the variable y (l in model ^d), x (n ^d) carry out integer and relax, recycling cplex solves, and in the result solved, traversal all-multicast dummy node is each multicast virtual sensor selection problem x (n ^d) maximum both candidate nodes n ^das final mapping node, select n ^dbetween path candidate be final mapping path, and carry out the cross-domain mapping of multicast network request MVN with this;

In the present embodiment, according to said method, model solution is carried out to the overall figure in Fig. 2 (d), shown in its solving result Fig. 2 (e), by multicast network request MVN cross-domain mapping in territory 2;

The cross-domain mapping of S3, unicast networks request UVN

Unmapped clean culture dummy node v in S3.1, estimation UVN _aterritory in map cost

Unmapped clean culture dummy node v is chosen from UVN _a, then estimate clean culture dummy node v _aterritory in map cost;

Cost(v _A→n ^k)+Cost(D)(7)

Wherein

$Cost\left(D\right)=\underset{p\∈P}{\Σ}Cost\left(p\right)---\left(8\right)$

Cost (v _a→ n ^k) can calculate according to formula (2);

Wherein, Cost (D) is clean culture dummy node v _aand the mapping being mapped in virtual link set between the dummy node in Same Physical territory in unicast networks spends, P is the mapping path set that virtual link is integrated into physical network, and p is the paths in P;

The physical node n of minimum mapping cost is chosen according to formula (7) ^kas clean culture dummy node v _aboth candidate nodes, the path between corresponding node is the path candidate of virtual link;

In the present embodiment, three physical domain are chosen equally according to said method, non-mapping node d, e mapping cost in territory 1, territory 2, territory 3 in estimation clean culture virtual network graph 3 (b), and in territory 1, territory 2, find the candidate mappings node of node d, e and internodal mapping path, as shown in Fig. 3 (c);

The cross-domain mapping of S3.2, unicast networks request UVN

S3.2.1, to unicast networks build global view

First initialization overall situation figure is sky, then the both candidate nodes in each territory and path candidate is added in overall figure, and increases cross-domain path candidate according to the both candidate nodes in not same area according to the virtual link annexation in virtual network;

In the present embodiment, according to said method, both candidate nodes in Fig. 3 (c) and alternative link are added in overall figure, wherein, both candidate nodes in territory 1 and territory 3 in overall figure corresponding points for being respectively d1, e1 and d2, e2, then increase cross-domain path candidate according to the annexation of 3 (b) interior joint, its result is as shown in Fig. 3 (d);

S3.2.2, carry out optimization modeling according to global view

Optimization aim is: $Min\left\{\underset{n^d\∈N_U^D}{\Σ}x\left(n^d\right)*f\left(n^d\right)+\underset{l^d\∈E_U^D}{\Σ}y\left(l^d\right)*f\left(l^d\right)\right\}---\left(9\right)$

Constraints is: $\left\{\begin{array}{c}\underset{n^d:cad\left(n^d\right)=n_u^v}{\Σ}x\left(n^d\right)=1,\∀n_u^v\∈N_U^V\\ \underset{l^d:cad\left(l^d\right)=l_u^v}{\Σ}y\left(l^d\right)=1,\∀l_u^v\∈E_U^V\\ y\left(l^d\right)\≤x\left(n^d\right),\∀l^d=(n^d,m^d)\\ x\left(n^d\right)\∈\{0,1\},\∀n^d\∈N_U^D\\ y\left(l^d\right)\∈\{0,1\},\∀l^d\∈E_U^D\end{array}\right.---\left(10\right)$

Wherein, l ^dfor the path candidate in overall figure, n ^d, m ^dfor the both candidate nodes in overall figure and be l ^dtwo end points, x (n ^d) be binary variable, as both candidate nodes n ^dbeing selected, is 1, otherwise is 0, y (l ^d) be binary variable, as path candidate l ^dbeing selected, is 1, otherwise is 0, cad (n ^d) for being mapped in n ^don dummy node, cad (l ^d) for being mapped in l ^don virtual link, f (n ^d) be the mapping cost f (n of both candidate nodes ^d)=Cost (cad (n ^d) → n ^d), f (l ^d) be the mapping cost f (l of path candidate ^d)=Cost (l ^d), for clean culture dummy node, for clean culture virtual link, for the node set in overall figure, for the set of paths in overall figure. for unicast networks node set, for unicast networks link set;

S3.2.3, model solution

First to the variable y (l in model ^d), x (n ^d) carry out integer and relax, recycling cplex solves, and in the result solved, travels through all clean culture dummy nodes, is that each clean culture dummy node selects x (n ^d) maximum both candidate nodes n ^das final mapping node, select n ^dbetween path candidate be final mapping path, and carry out the cross-domain mapping of unicast networks request UVN with this;

In the present embodiment, according to said method, model solution is carried out to the overall figure in Fig. 3 (d), shown in its solving result Fig. 3 (e), by unicast networks request UVN cross-domain mapping in territory 1.

Two, based on the mixing virtual network cross-domain mapping method of pattern decomposition, its concrete steps implemented comprise:

The spectrum segmentation of T1, HVN request

T1.1, to HVN request set up adjacency matrix

If A _{n × n}for the adjacency matrix that single HVN asks, wherein n is dummy node number in HVN request, and carry nodal community in each dummy node, i.e. multicast or clean culture, A (i, j) represents the virtual link bandwidth demand between i-th dummy node and a jth dummy node, wherein, i, j ∈ [1, n];

T1.2, structure matrix D _{n × n}, and meet:

$\left\{\begin{array}{c}D(i,j)=\underset{j}{\Σ}A\left(i,j\right),\left(i=j\right)\\ D(i,j)=0,(i\≠j)\end{array}\right.---\left(11\right)$

T1.3, structure matrix B _{n × n}, B _{n × n}=D _{n × n}-A _{n × n};

T1.4, compute matrix B _{n × n}front k characteristic value and characteristic of correspondence vector, wherein, the maximum occurrences of k is matrix B _{n × n}the total number of characteristic value; A feature space is formed again by characteristic vector, each vector in feature space represents some dummy nodes by the position that it occurs in characteristic vector, to the vectorial K-means cluster in feature space, cluster acquired results is exactly the dummy node set of each subgraph;

In the present embodiment, according to the method described above the multicast hybrid network in Fig. 1 is split into subgraph (a) as shown in Figure 4 and subgraph (b);

T2, choose the candidate domain of subgraph

T2.1, subgraph to be sorted

The attribute of all dummy nodes in traversal subgraph, made number one by the subgraph containing multicast root node, the subgraph containing multicast leaf node comes thereafter successively, finally comes finally successively by the subgraph only containing clean culture node;

Cost is mapped in the territory of T2.2, estimation subgraph

If containing multicast root node in subgraph, then estimate its cost according to formula (1), then choose map Least-cost physical node as mapping node in the territory of multicast root node;

If containing multicast leaf node in subgraph, then estimate its territory according to formula (2) and map link maps cost in cost, territory, then the physical node choosing mapping Least-cost is as mapping node in the territory of multicast leaf node;

If containing clean culture dummy node in subgraph, first calculate in its territory according to formula (7) and map cost, recycling formula (10) calculates current subgraph virtual unicast node v _uand the link maps cost between the subgraph mapped.

$Cost\left(M\right)=\underset{p\∈L}{\Σ}Cost\left(p\right)---\left(10\right)$

Wherein L is and node v _ulink set between the subgraph having an annexation.

Then dummy node v _utotal mapping cost be

Const(v _u→n ^k)+Cost(M)+Cost(D)(11)

The physical node of mapping cost minimization is chosen as mapping node in the territory of clean culture dummy node according to formula (11);

In the present embodiment, as Fig. 4 (a), choose three physical domain and carry out preliminary mapping, node a, b, c mapping cost in three physical domain in subgraph (a) is estimated by said method, respectively in territory 1, its candidate mappings node and internodal candidate mappings path is found, as shown in Fig. 4 (d) in territory 2 and territory 3

The cross-domain mapping of T3, subgraph

According to the mapping process of step T2.2, calculate the mapping cost sum of all nodes in each subgraph respectively, in each subgraph, choose the final mapping territory of territory as this subgraph of mapping cost minimization respectively, corresponding mapping node and mapping path are final mapping node and path, thus complete the cross-domain mapping of all subgraphs.

In the present embodiment, as shown in Fig. 4 (d), choose according to said method and map the mapping territory of the minimum territory 2 of cost sum as subgraph (a), its mapping result is as shown in Fig. 4 (e); Repeat the mapping that step T2 and T3 completes subgraph 4 (b), as shown in Fig. 4 (f).

Although be described the illustrative embodiment of the present invention above; so that those skilled in the art understand the present invention; but should be clear; the invention is not restricted to the scope of embodiment; to those skilled in the art; as long as various change to limit and in the spirit and scope of the present invention determined, these changes are apparent, and all innovation and creation utilizing the present invention to conceive are all at the row of protection in appended claim.

Claims (3)

1., based on a mixing virtual network cross-domain mapping method for HVNMMD-D algorithm, it is characterized in that, comprise the following steps:

(1) HVN request, is split

According to the attribute of dummy node and virtual link in mixing virtual network, HVN request is split into multicast network request MVN and unicast networks request UVN;

(2), the cross-domain mapping of multicast network request MVN

(2.1), the mapping cost of MVN in each physical domain is estimated

(2.1.1) mapping cost in the territory, estimating multicast root node

Cost(v→n ^s)＝(cst(n ^s)+(MC-Con(n ^s)))*req(v)(1)

Wherein, Con (n ^s) be physical node n ^sadjacent node number, MC is Con (n maximum in all nodes of bottom-layer network in territory ^s), req (v) puts v resource requirement quantity, cst (n successively for MVN root ^s) be physical node n ^sresource unit price;

According to formula (1), choose the candidate mappings node of physical node as MVN root node v of cost minimization;

(2.1.2) mapping cost in the territory, estimating multicast leaf node

Cost(v _L→n ^k)+Cost(p)(2)

Wherein

Cost(v _L→n ^k)＝cst(n ^k)*req(v _L)(3)

$Cost\left(p\right)=\underset{l^s\∈p}{\Σ}cst\left(l^s\right)*req\left(l^v\right)---\left(4\right)$

Wherein, cst (n ^k) be physical node n ^kresource unit price, req (v _l) be multicast leaf node v _lresource requirement quantity, Cost (v _l→ n ^k) be multicast leaf node v _lat physical node n ^kon mapping cost;

L ^vfor connecting multicast root node v and multicast leaf node v _lvirtual link, p is virtual link l ^vmapping path in physical network, cst (l ^s) be physical link l ^sresource unit price, req (l ^v) be virtual link resource requirement, Cost (p) is virtual link l ^vmapping cost;

According to formula (2), choose the physical node n mapping least cost ^sas MVN leaf node v _lcandidate mappings node, the path p between corresponding node is virtual link l ^vcandidate mappings path;

(2.2), the cross-domain mapping of multicast network request MVN

(2.2.1), global view is built to multicast network

First initialization overall situation figure is sky, then the both candidate nodes in each territory and path candidate is added in overall figure, and increases cross-domain path candidate according to the both candidate nodes in not same area according to the virtual link annexation in virtual network;

(2.2.2), optimization modeling is carried out according to global view

Optimization aim is: $Min\left\{\underset{n^d\∈N_M^D}{\Σ}x\left(n^d\right)*f\left(n^d\right)+\underset{l^d\∈E_M^D}{\Σ}y\left(l^d\right)*f\left(l^d\right)\right\}---\left(5\right)$

Constraints is: $\left\{\begin{array}{c}\underset{n^d:cad\left(n^d\right)=n_m^v}{\Σ}x\left(n^d\right)=1,\∀n_m^v\∈N_M^V\\ \underset{l^d:cad\left(l^d\right)=l_m^v}{\Σ}y\left(l^d\right)=1,\∀l_m^v\∈E_M^V\\ y\left(l^d\right)\≤x\left(n^d\right),\∀l^d=(n^d,m^d)\\ x\left(n^d\right)\∈\{0,1\},\∀n^d\∈N_M^D\\ y\left(l^d\right)\∈\{0,1\},\∀l^d\∈E_M^D\end{array}\right.---\left(6\right)$

Wherein, l ^dfor the path candidate in overall figure, n ^d, m ^dfor l ^dtwo end points.X (n ^d) be binary variable, as both candidate nodes n ^dbeing selected, is 1, otherwise is 0, y (l ^d) be binary variable, as path candidate l ^dbeing selected, is 1, otherwise is 0, cad (n ^d) for being mapped in n ^don dummy node, cad (l ^d) for being mapped in l ^don virtual link, f (n ^d) be the mapping cost f (n of both candidate nodes ^d)=Cost (cad (n ^d) → n ^d), f (l ^d) be the mapping cost f (l of path candidate ^d)=Cost (l ^d), for multicast virtual node, for multicast virtual link, for the both candidate nodes set in overall figure, for the alternative path set in overall figure, for multicast node set, for multicast link set.

(2.2.3), model solution

First to the variable y (l in model ^d), x (n ^d) carry out integer and relax, recycling cplex solves, and in the result solved, traversal all-multicast dummy node is each multicast virtual sensor selection problem x (n ^d) maximum both candidate nodes n ^das final mapping node, select n ^dbetween path candidate be final mapping path, and carry out the cross-domain mapping of multicast network request MVN with this;

(3), the cross-domain mapping of unicast networks request UVN

(3.1), unmapped clean culture dummy node v in UVN is estimated _aterritory in map cost

Unmapped clean culture dummy node v is chosen from UVN _a, then estimate clean culture dummy node v _aterritory in map cost;

Cost(v _A→n ^k)+Cost(D)(7)

Wherein

$Cost\left(D\right)=\underset{p\∈P}{\Σ}Cost\left(p\right)---\left(8\right)$

Cost (v _a→ n ^k) can calculate according to formula (2);

Wherein, Cost (D) is clean culture dummy node v _aand the mapping being mapped in virtual link set between the dummy node in Same Physical territory in unicast networks spends, P is the mapping path set that virtual link is integrated into physical network, and p is the paths in P;

The physical node n of minimum mapping cost is chosen according to formula (7) ^kas clean culture dummy node v _aboth candidate nodes, the path between corresponding node is the path candidate of virtual link;

(3.2), the cross-domain mapping of unicast networks request UVN

(3.2.1), global view is built to unicast networks

First initialization overall situation figure is empty, again the both candidate nodes in each territory selected in step (3.1) and path candidate are added in overall figure, and increase cross-domain path candidate according to the both candidate nodes in not same area according to the virtual link annexation in virtual network;

(3.2.2), optimization modeling is carried out according to global view

Optimization aim is: $Min\left\{\underset{n^d\∈N_U^D}{\Σ}x\left(n^d\right)*f\left(n^d\right)+\underset{l^d\∈E_U^D}{\Σ}y\left(l^d\right)*f\left(l^d\right)\right\}---\left(9\right)$

Constraints is: $\left\{\begin{array}{c}\underset{n^d:cad\left(n^d\right)=n_u^v}{\Σ}x\left(n^d\right)=1,\∀n_u^v\∈N_U^V\\ \underset{l^d:cad\left(l^d\right)=l_u^v}{\Σ}y\left(l^d\right)=1,\∀l_u^v\∈E_U^V\\ y\left(l^d\right)\≤x\left(n^d\right),\∀l^d=(n^d,m^d)\\ x\left(n^d\right)\∈\{0,1\},\∀n^d\∈N_U^D\\ y\left(l^d\right)\∈\{0,1\},\∀l^d\∈E_U^D\end{array}\right.---\left(10\right)$

Wherein, l ^dfor the path candidate in overall figure, n ^d, m ^dfor the both candidate nodes in overall figure and be l ^dtwo end points, x (n ^d) be binary variable, as both candidate nodes n ^dbeing selected, is 1, otherwise is 0, y (l ^d) be binary variable, as path candidate l ^dbeing selected, is 1, otherwise is 0, cad (n ^d) for being mapped in n ^don dummy node, cad (l ^d) for being mapped in l ^don virtual link, f (n ^d) be the mapping cost f (n of both candidate nodes ^d)=Cost (cad (n ^d) → n ^d), f (l ^d) be the mapping cost f (l of path candidate ^d)=Cost (l ^d), for clean culture dummy node, for clean culture virtual link, for the node set in overall figure, for the set of paths in overall figure. for unicast networks node set, for unicast networks link set;

(3.2.3), model solution

First to the variable y (l in model ^d), x (n ^d) carry out integer and relax, recycling cplex solves, and in the result solved, travels through all clean culture dummy nodes, is that each clean culture dummy node selects x (n ^d) maximum both candidate nodes n ^das final mapping node, select n ^dbetween path candidate be final mapping path, and carry out the cross-domain mapping of unicast networks request UVN with this.

2., based on a mixing virtual network cross-domain mapping method for pattern decomposition, it is characterized in that, comprise the following steps:

(1), the spectrum segmentation of HVN request

(1.1), adjacency matrix is set up to HVN request

If A _{n × n}for the adjacency matrix that single HVN asks, wherein n is dummy node number in HVN request, and carry nodal community in each dummy node, i.e. multicast or clean culture, A (i, j) represents the virtual link bandwidth demand between i-th dummy node and a jth dummy node, wherein, i, j ∈ [1, n];

(1.2), matrix D is built _{n × n}, and meet:

$\left\{\begin{array}{c}D(i,j)=\underset{j}{\Σ}A\left(i,j\right),\left(i=j\right)\\ D(i,j)=0,(i\≠j)\end{array}\right.---\left(11\right)$

(1.3), matrix B is built _{n × n}, B _{n × n}=D _{n × n}-A _{n × n};

(1.4), compute matrix B _{n × n}front k characteristic value and characteristic of correspondence vector, wherein, the maximum occurrences of k is matrix B _{n × n}the total number of characteristic value; A feature space is formed again by characteristic vector, each vector in feature space represents some dummy nodes by the position that it occurs in characteristic vector, to the vectorial K-means cluster in feature space, cluster acquired results is exactly the dummy node set of each subgraph;

(2) candidate domain of subgraph, is chosen

(2.1), subgraph is sorted

The attribute of all dummy nodes in traversal subgraph, made number one by the subgraph containing multicast root node, the subgraph containing multicast leaf node comes thereafter successively, finally comes finally successively by the subgraph only containing clean culture node;

(2.2), estimate subgraph territory in mapping cost

If containing multicast root node in subgraph, then estimate its cost according to formula (1), then choose map Least-cost physical node as mapping node in the territory of multicast root node;

If containing multicast leaf node in subgraph, then estimate its territory according to formula (2) and map link maps cost in cost, territory, then the physical node choosing mapping Least-cost is as mapping node in the territory of multicast leaf node;

If containing clean culture dummy node in subgraph, first calculate in its territory according to formula (7) and map cost, recycling formula (10) calculates current subgraph virtual unicast node v _uand the link maps cost between the subgraph mapped.

$Cost\left(M\right)=\underset{p\∈L}{\Σ}Cost\left(p\right)---\left(10\right)$

Wherein L is and node v _ulink set between the subgraph having an annexation.

Then dummy node v _utotal mapping cost be

Const(v _u→n ^k)+Cost(M)+Cost(D)(11)

The physical node of mapping cost minimization is chosen as mapping node in the territory of clean culture dummy node according to formula (11).

(3), the cross-domain mapping of subgraph

According to the mapping process of step (2.2), calculate the mapping cost sum of all nodes in each subgraph respectively, in each subgraph, choose the final mapping territory of territory as this subgraph of mapping cost minimization respectively, corresponding mapping node and mapping path are final mapping node and path, thus complete the cross-domain mapping of all subgraphs.

3. the mixing virtual network cross-domain mapping method according to claim 1 and 2, it is characterized in that, described dummy node comprises multicast virtual node and clean culture dummy node, and multicast virtual node comprises again multicast root node and multicast leaf node.