CN103825963A - Virtual service transition method - Google Patents

Abstract

The invention relates to a virtual service transition method which comprises the following steps: selecting transition evaluation parameters according to service types and user requirements; monitoring virtual nodes in a network by a monitor; obtaining state information of the virtual nodes in regard to the evaluation parameters when a new request is reached or a user request is changed; calculating the service transition cost of each virtual node according to the state information; acquiring the virtual node with the lowest virtual service transition cost as a service node from the virtual nodes through the transition cost calculation method; selecting the service node to transit the virtual service. The virtual service transition method provided by the invention can be used for judging and executing the transition by effectively combining the influences of each factor on the transition, and realizing the management of network resources and improving the user service experience quality by utilizing the service transition.

Description

Virtual Service moving method

Technical field

The present invention relates to mobile Internet field, relate in particular to a kind of Virtual Service moving method.

Background technology

Growing along with science and technology, mobile Internet has become one of main trend of internet development.According to statistics, by the end of in March, 2013, China's fixed the Internet broadband user number is 1.81 hundred million users, and mobile Internet user has reached 8.17 hundred million users.Along with the enhancing of Network Mobility sexual demand, the change of user behavior, the enriching constantly of information interaction type, the increase day by day of data traffic, ensures and the service experience quality that improves user becomes the new challenge that internet arena will be faced.And the problem that develops into solution mobility demand of Intel Virtualization Technology provides possibility.Virtualized target is exactly without consideration bottom physical network attribute in the situation that, to realize the level and smooth movement of Virtual Service, realizes the distribution according to need to Internet resources, improves user's Quality of experience.Fig. 1 is network virtualization schematic diagram in prior art.As shown in Figure 1, Intel Virtualization Technology is by abstract, separation, isolation mech isolation test, at a public bottom physical network (Substrate Network, SN) upper multiple virtual networks (Virtual Network that supports, VN) each virtual network can be used separate protocol architecture, and according to the user's request of dynamic change, the node resource in whole network and link circuit resource are carried out to reasonable disposition, greatly to bring into play the advantage of resource-sharing, improve to greatest extent network resource utilization, obtain maximized income.

Fig. 2 is that in prior art, Flowvisor realizes virtualized schematic diagram.As shown in Figure 2, network virtualization platform FlowVisor produces independently network burst by dividing flow table space, effectively realizes network virtualization, and physical network is divided into multiple network bursts.Network traffics on each network burst are isolated mutually, and managing bandwidth, CPU use and the configuration of stream table etc.User can carry out the experimental studies such as non-interfering various discharge model and agreement innovation on each burst.FlowVisor realizes and disposing in some Large-scale Campus such as Stanford Univ USA at present, in famous future network experimental bed GENI and Internet 2 projects, is also using FlowVisor to carry out virtualized management.

Fig. 3 is the schematic diagram that in prior art, network virtualization hierarchical service supplies a model.As shown in Figure 3, in virtualized environment, service provider (Service Provider, SP) according to the demand of user (User), the resource of needs (Internet resources, computational resource, storage resources etc.) is described to the infrastructure provider (Infrastructure Provider, InP) to bottom with certain forms.Infrastructure provider disposes and management bottom physical resource, selects corresponding resource, to complete the establishment to virtual network from virtual resource pond.Service provider implements provider's leasehold network resource to basis, for user provides needed service.

Because some new users add, some old users shift out, and the change of some users position in network, the variation of number of users, user behavior, user preference in network, or the change of some bottom-layer networks, all need according to changing, Virtual Service to be adjusted and moved.How to change the scale to virtual network, resource distribution timely according to these adjusts, ensure network service quality (Quality of Service, and user experience quality (Quality of Experience QoS), QoE), obtaining reliable and stable network service with minimum delay, is one of challenge realizing Virtual Service.Virtual main finger server/main frame in the present invention virtual, shares reusable software and hardware resources and information offers computer and other equipment as required by virtualized.

Fig. 4 is that in prior art, user moves schematic diagram.As shown in Figure 4, services migrating need to be considered the balance of various costs.When with the nearer dummy node of user during as ISP, the delay of service may be less, and service quality and Quality of experience are higher.But migration also can bring other expense, has a negative impact simultaneously, during as migration, mass data is transmitted the pressure to network, even may cause service disruption.Services migrating problem is exactly how reasonably to adjust the position of network service, more effectively saves Internet resources, reduces service response time and cost, improves the problem of user's experience.

In the prior art, at present under virtualized environment, utilize services migrating to realize network resource management and reduce the research of energy consumption and apply still more preliminary, as J Grassler, the people such as S Schmid are at " at 32nd IEEE Conference on Computer Communications (INFOCOM Demo), Turin, Italy, April2013. " research that " Move-with-the-Sun or Move-with-the-Moon Wide-Area CloudNet Migrations Under Latency and Resource Constraints.Demo " makes.Services migrating problem under virtualized environment is mainly divided into single domain and two kinds of application scenarioss of multiple domain.

Under single domain environment, less in the aspect such as type, quality of resource difference between node, move needed cost less.But in the time that the node in territory cannot meet user's demand, need to move to node in other territories so that service to be provided.Between node in multiple territories, not only there are differences, also can be due to the extra roaming cost of cross-domain generation.

Below to the M Bienkowski relevant with this patent, the people such as A Feldmann are at " In Proc.ACM SIGCOMM VISA, 2010. " " Competitive analysis for service migration in vnets " proposes Virtual Service balance migration algorithm (MIG) and D Arora, the people such as M Bienkowski are at " Proceedings of the5th International Conference on Principles, Systems and Applications of IP Telecommunication, 2011. " the cross-domain balanced algorithm (MIX that " Online strategies for intra and inter provider service migration in virtual networks " proposes_k) briefly introduce.

(1) Virtual Service balance migration algorithm (MIG)

Virtual Service balance migration algorithm (MIG) quantizes the parameters of migration cost and income, by the method for Dynamic comparison, and the opportunity of judgement migration.The basic thought of MIG algorithm is in the time that migration occurs, and reaches migration cost Cost_migwith income Cost_accbetween balance, select feasible, a more excellent service provider.

In the time that user's position changes, the delay from virtual server end to client will increase, and this will affect the service quality of some services.By migration, can realize the optimization to service delay, serving time delay is migration income Cost_acca part.In addition, can migration realize also relevant with the disposable load of server.If potential node can not meet user's demand, migration cannot be carried out.And in the time moving, under identical condition, should select the larger node of disposable load of server to move.Therefore in the t moment, to request sequence R_t, carry out the income Cost that migration can produce_acccan be expressed as

${\mathrm{Cost}}_{\mathrm{acc}}\left(t\right)=\underset{R_t}{\mathrm{\Σf}}(\mathrm{delay}\left(r_t\right),\mathrm{load}\left(r_t\right))---\left(1\right)$

In literary composition, suppose that the disposable load of Servers-all can both meet user's demand, to Cost_accsimplify, obtain

${\mathrm{Cost}}_{\mathrm{acc}}\left(t\right)=\underset{R_t}{\mathrm{\Σ}}f\left(\mathrm{delay}\left(r_t\right)\right)---\left(2\right)$

Due to the attend to the basic or the fundamental impact of body size of the bandwidth kimonos on migration path, need careful consideration whether carry out services migrating.The big or small size (s) of service itself and the bandwidth w (p) of migration path have determined the needed time of migration jointly.In literary composition by the migration cost Cost of the service s moving_migbe expressed as

${\mathrm{Cost}}_{\mathrm{mig}}\left(t\right)=\underset{s\∈S}{\mathrm{\Σ}}f(w\left(p\right),k,\mathrm{size}\left(s\right))---\left(3\right)$

In Virtual Service balance migration algorithm, only considering provides the problem of the services migrating in territory, therefore the cost Cost of services migrating an infrastructure_migcan be reduced to

Cost_mig(u,v)＝max_esize(s)/w(e) （4）

Virtual Service balance migration algorithm will be divided into multiple time slots the time, suppose that the initial service provider that provides is node v.In the time that request arrives, time slot starts, and calculates the cost Cost producing due to the variation of request_acc(v).If by other virtual network node, as node u, as service provider, there is services migrating, can produce migration cost Cost_mig(v, u).If β=max_u{ Cost_mig(u, v) }.If meet Cost_acc(v) > β, moves, from meeting inequality Cost_acc(u) solution of < β is concentrated the random service provider of a node as this request that select.If there is no such node u, does not need to move, and time slot finishes.In the time that next one request arrives, new time slot starts, and recalculates Cost_acc(v).

(2) cross-domain balanced algorithm (MIX_k)

Cross-domain balanced algorithm (MIX_k) be improvement and the optimization to Virtual Service balance migration algorithm, consider that service providing node is positioned at the situation of multiple virtual networks.The time delay of request is Cost_acc(v), migration cost β=max_u{ Cost_mig(u, v) }.Suppose that the extra cost that territory of leap produces is π (wherein, π>=β²), the roaming cost of crossing over k territory is k* π.

Suppose that the initial service provider that provides is node v, in the time that service request sequence arrives, first consider the migration situation in a territory.Calculate the income Cost being produced by request_acc(v), and with migration cost β=max_u{ Cost_mig(u v) } compares.If meet Cost_acc(v) > β, from meeting inequality Cost_acc(u) solution of < β is concentrated the random destination node of a node as migration of selecting.If there is not such node u in territory, consider the services migrating in cross-domain situation.If income Cost_acc(v) the extra cost that is better than cross-domain generation is moved.

In sum, in the moving method of prior art, exist such as following problem: MIG migration algorithm has only simply been considered service time delay and two kinds of influencing factors of link bandwidth, and the selection of impact migration node is a lot of because have.Service time delay is simply quantified as service request access node to the jumping figure between service provider.The quantification of two kinds of factors is too simple, inflexible, is unfavorable for the dynamic adjustment on the opportunity of moving.And, between each factor that determines to move, be the impact that interacts, the simple big or small correlation that can not reflect between each factor.Although MIX_kin algorithm, increase the roaming cost producing due to cross-domain, but still existed above-mentioned quantification and correlation to portray simple question.

Summary of the invention

The object of the invention is for the problems referred to above, a kind of Virtual Service moving method based on fair, just and dynamic QoS computation model is provided.

For achieving the above object, the invention provides a kind of Virtual Service moving method, described method comprises:

Select migration evaluating according to COS and user's request, the dummy node in monitor monitoring network, in the time that new request arrival or user's request change, obtains the state information of described dummy node about evaluating;

Calculate the services migrating cost of each dummy node according to described state information;

The dummy node that obtains Virtual Service migration Least-cost by the services migrating cost of the each dummy node of described calculating from dummy node is as service node;

Select described service node to move described Virtual Service.

Preferably, described state information comprises cost, ISP's credit and the strike price of size, service disruption and the recovery of bandwidth on disposable load, the migration path of time delay, server of service, service itself.

Preferably, the described services migrating cost of calculating each dummy node according to described state information specifically comprises:

Set up the computation model generating based on QoS according to described state information.

Preferably, the described computation model generating based on QoS according to described state information foundation specifically comprises:

According to described state information, obtain matrix Q:

$Q=\left(\begin{array}{cccc}{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{\mathrm{1,1}}& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{1,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{1,\mathrm{<figure-callout\; label="m\; q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">m</figure-callout>}}& \\ q_{}{}_{\mathrm{2,1}}& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{2,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{2,m}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \text{\&CenterDot;}& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ q_{n,1}& q_{\text{n,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& q_{n,m}\end{array}\right)$

Wherein, m, n are integer, q_n,mfor the element in matrix Q, with C (v_i) expression dummy node v_imigration cost, every state information of each the line display dummy node in matrix Q, a kind of state information that affects described migration cost is shown in each list in matrix Q.

Preferably, the described computation model generating based on QoS according to described state information foundation specifically also comprises:

Described matrix Q is normalized, the described state information of different dimensions is standardized as to nondimensional normalizing parameter, form a unified criterion;

Described normalizing parameter is divided into groups, every group comprise multiple shown in normalizing parameter, and by group operation, draw the migration cost information of each dummy node after grouping.

Preferably, described described matrix Q is normalized, the described state information of different dimensions is standardized as to nondimensional normalizing parameter, form a unified criterion and specifically comprise:

The first matrix N={n₁, n₂..., n_mrepresent wherein n_j=0 or 1,1≤j≤m;

The second matrix C={c₁, c₂..., c_mrepresent c_ja constant, 1≤j≤m;

Each element in matrix Q is normalized with following formula:

${q^{\&prime;}}_{i,j}=\left\{\begin{array}{c}\frac{q_{i,j}}{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}},\mathrm{if}\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}\&NotEqual;0,\mathrm{and},\frac{q_{i,j}}{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}}<c_j,\mathrm{amd},n_j=1\\ c_j,\mathrm{if}\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}=0,\mathrm{and},n_j=1,\mathrm{or},\frac{q_{i,j}}{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}}\text{\&GreaterEqual;}{c}_j\end{array}\right.$

${q^{\&prime;}}_{i,j}=\left\{\begin{array}{c}\frac{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}}{q_{i,j}},\mathrm{if},q_{i,j}\&NotEqual;0,\mathrm{and},n_j=0,\mathrm{and},\frac{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}}{q_{i,j}}<c_j\\ c_j,\mathrm{if},q_{i,j}=0,\mathrm{and},n_j=0,\mathrm{or},\frac{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}}{q_{i,j}}\&GreaterEqual;c_j\end{array}\right.$

Wherein,

for the mean value of j kind cost standard in matrix Q, obtain matrix Q':

$Q^{\&prime;}=\left(\begin{array}{cccc}{{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{1,1}}}^{\text{\&prime;}}& {{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{1,2}}}^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{1,m}}^{\&prime;}\\ {{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{2,1}}}^{\&prime;}& {{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{2,2}}}^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{2,m}}^{\&prime;}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \text{\&CenterDot;}& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ {q_{n,1}}^{\&prime;}& {q_{n,2}}^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {q_{n,m}}^{\&CenterDot;}\end{array}\right)$

Wherein, q_i,jfor the element in matrix Q, q_n,m' be the element in matrix Q'.

Preferably, described method also comprises: in the time not having new request to arrive or user asks not change, do not carry out the migration of dummy node.

The beneficial effect that the present invention brings is:

1, adopt the cost of the method calculation services migration based on QoS computation model, avoid because influencing factor quantification is too simple, inflexible, and the problem that is unfavorable for migration dynamic adjustment on opportunity causing;

2, the impact of effectively comprehensive many factors on migration cost, allowing increases self-defined input parameter, can dynamically adjust according to virtual network environment and user's demand;

What 3, consideration was relevant with dummy node self property affects parameter, has increased credit and the price of dummy node, more comprehensively, the service quality of the evaluation and test Virtual Service node that gears to actual circumstances;

4, select the node of migration Least-cost to move, be conducive to the expense that minimum migration is brought, improve user's service experience quality, realize green energy conservation, the efficient virtual resource migration and scheduling utilizing.

Accompanying drawing explanation

Fig. 1 is network virtualization schematic diagram in prior art of the present invention;

Fig. 2 is that in prior art of the present invention, Flowvisor realizes virtualized schematic diagram;

Fig. 3 is the schematic diagram that in prior art of the present invention, network virtualization hierarchical service supplies a model;

Fig. 4 is that in prior art of the present invention, user moves schematic diagram;

Fig. 5 is the flow chart of Virtual Service moving method in one embodiment of the invention of the present invention;

Fig. 6 is the schematic diagram that in one embodiment of the invention, local monitor monitors for the each dummy node in network;

Fig. 7 is the flow chart that in one embodiment of the invention, the services migrating cost to each Virtual Service node is calculated.

Embodiment

Below by drawings and Examples, technical scheme of the present invention is described in further detail.

Fig. 5 is the flow chart of Virtual Service moving method in one embodiment of the invention.

As shown in Figure 5, first, instep 501, select migration evaluating according to COS and user's request.Dummy node in monitor monitoring network, in the time that new request arrival or user request changes, obtains the state information of dummy node about evaluating.

Concrete, the cost of a dummy node/virtual server is subject to the joint effect of multiple factors.For problem is simplified, in one embodiment of the invention, only consider to only have above physical network the situation of a virtual network, and in virtual network, only have the situation of a dummy node/server.

Table 1 has provided main several factors of impact migration cost.

Table 1

Affect parameter	Meaning of parameters
		Cdelay	The time delay of service
Caload	The disposable load of server
		Cbandwidth	Bandwidth on migration path
Csize	The size of service itself
		Cinterrupt	The cost of service disruption and recovery
Creputation	ISP's credit
		Cprice	Strike price

[0070]wherein, in order to select the better dummy node of service quality as service provider is provided, the present invention has increased by 2 kinds of attributes relevant with service provider's self property, i.e. C_reputationand C_price.

C_reputationfor ISP's credit, it is the main evaluation index of dummy node trustworthiness.It depends on terminal use's historical service experience.To same ISP, different terminal uses has different evaluations.The credit of dummy node is defined as the mean value of multiple terminal use's Feedback Evaluations,

wherein n is the total degree that node is evaluated, R_ifor the value of feedback of terminal use to node credit, R_ifor belonging to the integer of [0,5].

C_pricefor strike price, in the time that dummy node provides service, user need to pay required expense.Under square one, requestor is more prone to select the node that price is lower that service is provided.

Fig. 6 is the schematic diagram that in one embodiment of the invention, local monitor monitors for the each dummy node in network.

As shown in Figure 6, monitor monitors the each dummy node in network, gathers the state information on dummy node.Wherein, state information comprises the state information shown in the tables 1 such as time of delay, link bandwidth and node disposable load.

Concrete, in an embodiment of the present invention by the set V={v of all Virtual Service provider compositions₁, v₂..., v_nrepresent, at t_kmoment, request δ_karrive.The Virtual Service supplier s of previous stage_k-1represent.The target of the present embodiment is to judge at t_kmoment request δ_kwhen arrival, whether need Virtual Service provider to move to new dummy node, and move to which node.In the time that needs move, we select the node that moves Least-cost to move.By C (v_i) show as node v_imigration cost.Carry out the cost of evaluation node by m standard, can obtain following matrix Q:

$Q=\left(\begin{array}{cccc}{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{\mathrm{1,1}}& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{1,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{1,\mathrm{<figure-callout\; label="m\; q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">m</figure-callout>}}& \\ q_{}{}_{\mathrm{2,1}}& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{2,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{2,m}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \text{\&CenterDot;}& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ q_{n,1}& q_{\text{n,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& q_{n,m}\end{array}\right)---\left(5\right)$

Matrix Q is the matrix of the capable m row of n, and n and m are integer.Wherein, the i line display Virtual Service node v in matrix Q_iindices, a kind of factor that affects cost is shown in every list.

As shown in Figure 5, instep 502 secondarily, calculate the services migrating cost of each dummy node according to state information.The concrete steps of computational methods are as follows:

Fig. 7 is the flow chart in one embodiment of the invention, the service cost of each Virtual Service node being calculated.

As shown in Figure 7, first instep 701, matrix Q is normalized, the state information of different dimensions is standardized as to nondimensional normalizing parameter, form a unified criterion.

Concrete, the object that matrix Q is normalized is: (1) allows migration cost method of measurement irrelevant with unit, standard; (2) for each Virtual Service supplier provides a kind of unified index of evaluating migration cost; (3) threshold value is set.

First, two matrixes of definition.First matrix N={ n₁, n₂..., n_m, get 1≤j≤m, wherein n_j=0 or 1.If along with q_ijincrease, it is large that the cost that migration produces becomes, and makes n_j=1(is as parameters C_sizelarger, migration cost is larger).N on the contrary_j=0(is more prone to C as when migration_aloadlarger node).Second Matrix C={ c₁, c₂..., c_m, c_jbe a constant, be set to every kind of normalized maximum of cost evaluation index.Each element in Q is normalized by formula (6) (7).

${q^{\&prime;}}_{i,j}=\left\{\begin{array}{c}\frac{q_{i,j}}{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}},\mathrm{if}\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}\&NotEqual;0,\mathrm{and},\frac{q_{i,j}}{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}}<c_j,\mathrm{amd},n_j=1\\ c_j,\mathrm{if}\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}=0,\mathrm{and},n_j=1,\mathrm{or},\frac{q_{i,j}}{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}}\text{\&GreaterEqual;}{c}_j\end{array}\right.---\left(6\right)$

${q^{\&prime;}}_{i,j}=\left\{\begin{array}{c}\frac{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}}{q_{i,j}},\mathrm{if},q_{i,j}\&NotEqual;0,\mathrm{and},n_j=0,\mathrm{and},\frac{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}}{q_{i,j}}<c_j\\ c_j,\mathrm{if},q_{i,j}=0,\mathrm{and},n_j=0,\mathrm{or},\frac{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{q}_{i,j}}{q_{i,j}}\&GreaterEqual;c_j\end{array}\right.---\left(7\right)$

In above formula,the mean value of j kind cost standard in matrix Q.Q is carried out to computing by above formula, obtain new matrix Q', as shown in Equation (8):

$Q^{\&prime;}=\left(\begin{array}{cccc}{{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{1,1}}}^{\text{\&prime;}}& {{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{1,2}}}^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{1,m}}^{\&prime;}\\ {{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{2,1}}}^{\&prime;}& {{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{2,2}}}^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{2,m}}^{\&prime;}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \text{\&CenterDot;}& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ {q_{n,1}}^{\&prime;}& {q_{n,2}}^{\&prime;}& \&CenterDot;\&CenterDot;\&CenterDot;& {q_{n,m}}^{\&CenterDot;}\end{array}\right)---\left(8\right)$

Instep 702 secondarily, normalizing parameter is divided into groups, every group comprise multiple shown in normalizing parameter, and by group operation, draw the migration cost information of each dummy node after grouping.

Concrete, the cost parameter of dummy node is divided into groups, every group can comprise many kinds of parameters, and by group operation.In an embodiment of the present invention, can be divided into access cost Cost_acc, migration cost Cost_mig, server cost Cost_prithree parts, as the cost of bandwidth, service size itself, service disruption and recovery all belongs to migration cost Cost_mig, service delay and disposable load etc. belongs to a part of Cost that accesses cost_mig, the price of server and credit belong to server cost Cost_pri.In the present embodiment, introduce matrix D, matrix D is for defining the relation between cost index and cost grouping.

$Q=\left(\begin{array}{cccc}{\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{\mathrm{1,1}}& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{1,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{1,\mathrm{<figure-callout\; label="m\; q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">m</figure-callout>}}& \\ q_{}{}_{\mathrm{2,1}}& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout></figure-callout>}}_{\mathrm{2,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; label="q"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">q</figure-callout>}}_{2,m}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \text{\&CenterDot;}& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ q_{n,1}& q_{\text{n,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& q_{n,m}\end{array}\right)---\left(9\right)$

Wherein, a kind of cost index factor of every line display in matrix D, a kind of cost grouping is shown in every list, wherein l is total packet count.In matrix D, if i cost index belongs to j grouping, d in Q'_i,j=1, otherwise d_i,j=0.

Matrix G is the cost information of each Virtual Service node after grouping.

$G=\left(\begin{array}{cccc}{\mathrm{<figure-callout\; label="g"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">g</figure-callout>}}_{\mathrm{1,1}}& {\mathrm{<figure-callout\; label="g"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="g"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">g</figure-callout></figure-callout>}}_{\mathrm{1,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; label="g"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">g</figure-callout>}}_{1,\mathrm{<figure-callout\; label="l\; g"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">l</figure-callout>}}& \\ g_{}{}_{\mathrm{2,1}}& {\mathrm{<figure-callout\; label="g"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}"><figure-callout\; label="g"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">g</figure-callout></figure-callout>}}_{\mathrm{2,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {\mathrm{<figure-callout\; label="g"\; filenames="HDA0000478635170000011.png"\; state="\{\{state\}\}">g</figure-callout>}}_{2,l}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \text{\&CenterDot;}& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ g_{n,1}& g_{\text{n,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& g_{n,l}\end{array}\right)---\left(10\right)$

Wherein, in G, every row represents an Internet Service Provider, and a kind of value of cost grouping is shown in every list.

Matrix G can calculate by following formula:

G= Q'*D （11）

To matrix, G is normalized, and first defines two new matrixes: matrix T and matrix F.Matrix T={ t₁, t₂..., t_l, wherein constant t_jrepresent every group of normalized threshold value.According to formula (12), G is normalized, the result obtaining is G '.

${g_{i,j}}^{\&prime;}=\left\{\begin{array}{c}\frac{g_{i,j}}{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{g}_{i,j}},\mathrm{if}\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{g}_{i,j}\&NotEqual;0,\mathrm{and},\frac{g_{i,j}}{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{g}_{i,j}}<t_j\\ t_j,\mathrm{if}\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{g}_{i,j}=0,\mathrm{or},\frac{g_{i,j}}{\frac{1}{n}{\mathrm{\&Sigma;}}_{i=1}^n{g}_{i,j}}\text{\&GreaterEqual;}{t}_j\end{array}\right.---\left(12\right)$

$G^{\&prime;}=\left(\begin{array}{cccc}{g^{\&prime;}}_{\mathrm{1,1}}& {g^{\&prime;}}_{\mathrm{1,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {g^{\text{\&prime;}}}_{1,l}\\ {g^{\&prime;}}_{\mathrm{2,1}}& {g^{\&prime;}}_{\mathrm{2,2}}& \&CenterDot;\&CenterDot;\&CenterDot;& {g^{\&prime;}}_{2,l}\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \text{\&CenterDot;}& \&CenterDot;\\ \&CenterDot;& \&CenterDot;& \&CenterDot;& \&CenterDot;\\ {g^{\&prime;}}_{n,1}& {g^{\&prime;}}_{n,2}& \&CenterDot;\&CenterDot;\&CenterDot;& {g^{\&prime;}}_{n,l}\end{array}\right)---\left(13\right)$

As shown in Figure 7, in the end instep 703, according to the migration cost of all dummy nodes of grouping weight COMPREHENSIVE CALCULATING.Definition matrix F={ f₁, f₂..., f_l, wherein f_jexpression cost is evaluated j and is grouped in the weight in overall merit, can be used for representing the preference of user to j grouping evaluation criterion.Provide node s to Virtual Service_icost calculate can obtain by formula (14):

$\mathrm{Cos}\left(s_i\right){\mathrm{\&Sigma;}}_{j=1}^l(g_{i,j}*f_j)---\left(14\right)$

Instep 503 subsequently, the dummy node that obtains Virtual Service migration Least-cost by migration cost computational methods from dummy node is as service node.

In an embodiment of the present invention, preferably have the virtual network of 6 nodes, node set is V={v₁, v₂..., v₆.Initial service node S₀for node v₁, at t₁moment, request sequence δ₁arrive.Preferably 5 kinds of migration cost evaluatings: the time delay C of service_delay, server disposable load C_aload, migration path bandwidth C_bandwidth, price C_pricewith credit C_reputation.Obtain the matrix Q of each Node evaluation parameter by local monitor:

$Q=\left(\begin{array}{ccccc}10& 10& 2& 40& 4\\ 8& 4& 4& 80& 1\\ 7& 4& 4& 100& 0.5\\ 5& 2& 6& 80& 2\\ 3& 4& 6& 100& 4\\ 6& 8& 4& 100& 2\end{array}\right)$

Definition threshold value C=(5,5,5,5,5), obtain N=(1 according to 5 kinds of evaluation indexes, 0,0,1,0).By above-mentioned formula (6) and (7), calculate normalized matrix Q':

$G^{\&prime;}\left(\begin{array}{ccccc}1.5382& 0.5333& 2.1667& 0.48& 0.5625\\ 1.2308& 1.3333& 1.0833& 0.96& 2.25\\ 1.0769& 1.3333& 1.0833& 1.2& 4.5\\ 0.7692& 2.6667& 0.7222& 0.96& 1.125\\ 0.4615& 1.3333& 0.7222& 1.2& 0.5625\\ 0.9231& 0.6667& 1.0833& 1.2& 1.125\end{array}\right)$

In embodiment, 5 kinds of evaluation indexes are divided into 3 classes, access cost Costacc, migration cost Costmig, server cost Costpri.Matrix D is:

$D=\left(\begin{array}{ccc}1& 0& 0\\ 1& 0& 0\\ 0& 1& 0\\ 0& 0& 1\\ 0& 0& 1\end{array}\right)$

Calculating matrix G by formula (11) is:

$G=\left(\begin{array}{ccc}2.0718& 2.1667& 1.0425\\ 2.5641& 1.0833& 3.21\\ 2.4102& 1.0833& 5.7\\ 3.4359& 0.7222& 2.085\\ 1.7948& 0.7222& 1.7625\\ 1.5898& 1.0833& 2.325\end{array}\right)$

In embodiment, define threshold matrix T=(3,3,3), be normalized and obtain G' and be by formula (12):

$G^{\&prime;}=\left(\begin{array}{ccc}0.8965& 1.8948& 0.3879\\ 1.1095& 0.9474& 1.1944\\ 1.0429& 0.9474& 2.1209\\ 1.4867& 0.6316& 0.7758\\ 0.7766& 0.6316& 0.6558\\ 0.6879& 0.9474& 0.8651\end{array}\right)$

Define the weight of 3 class evaluation criterions in gross migration evaluation and be respectively 0.4,0.4,0.2.The migration cost that calculates 6 service nodes by above-mentioned formula (14) is respectively 1.1941,1.0616,1.2203,1.0025,0.6944 and 0.8271.

Concrete, be respectively 1.1941,1.0616,1.2203,1.0025,0.6944 and 0.8271 from the migration cost of above-mentioned 6 service nodes that calculate, wherein 0.6944 migration cost value minimum, selects 0.694 corresponding dummy node v₅for service providing node.

Get back to Fig. 5, inlast step 504, select service node to move Virtual Service, move to the dummy node of migration Least-cost from the service node of previous moment.

Concrete, in 6 above-mentioned service nodes, from the service node v of previous moment₁move to v₅, i.e. the corresponding Virtual Service node of minimum transition cost 0.6944.In sum, the invention provides and dynamic QoS computation model open, fair based on one, proposed a kind of Virtual Service moving method.Wherein, QoS reaches an agreement the demand of service requester and ISP's needs according to available Internet resources.QoS comprises the possibility that service responds to a request in preset time, the situation quality of service execution task, speed speed and reliability of service and the fail safe etc. of service operation.Therefore, service impacting qos parameter has throughput, time of delay, time of implementation, reliability, expense, fail safe, credit etc.Unit difference between various parameters, numerical value difference is very large, there is no comparativity between index, and directly big or small result is inaccurate.Need to be normalized and COMPREHENSIVE CALCULATING parameter information, thereby allow service requester to select the service that service quality is higher in multiple services that meet the demands.

Virtual Service moving method provided by the invention calculates thought by QoS and applies to services migrating cost and calculate up, in the process of selection object migration node, not only consider that two kinds of the time delay of service and the bandwidth of link affect parameter, the multiple parameters that affect services migrating cost such as server disposable load, price are more considered, they are carried out to COMPREHENSIVE CALCULATING, select the node of migration Least-cost as the service provider of request sequence.The effectively comprehensively impact of various factors on migration, judges and carries out migration, utilizes services migrating to realize network resource management and improves user's service experience quality.

Professional should further recognize, unit and the algorithm steps of each example of describing in conjunction with embodiment disclosed herein, can realize with electronic hardware, computer software or the combination of the two, for the interchangeability of hardware and software is clearly described, composition and the step of each example described according to function in the above description in general manner.These functions are carried out with hardware or software mode actually, depend on application-specific and the design constraint of technical scheme.Professional and technical personnel can realize described function with distinct methods to each specifically should being used for, but this realization should not thought and exceeds scope of the present invention.

The software module that the method for describing in conjunction with embodiment disclosed herein or the step of algorithm can use hardware, processor to carry out, or the combination of the two is implemented.Software module can be placed in the storage medium of any other form known in random asccess memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technical field.

Above-described embodiment; object of the present invention, technical scheme and beneficial effect are further described; institute is understood that; the foregoing is only the specific embodiment of the present invention; the protection range being not intended to limit the present invention; within the spirit and principles in the present invention all, any modification of making, be equal to replacement, improvement etc., within all should being included in protection scope of the present invention.

Claims (7)

Translated fromChinese

1.一种虚拟服务迁移方法，其特征在于，所述方法包括：1. A virtual service migration method, characterized in that the method comprises:根据服务类型和用户需求选择迁移评价参数，监视器监视网络中的虚拟节点，当新请求到达或用户请求发生改变时，得到所述虚拟节点关于评价参数的状态信息;Select the migration evaluation parameters according to the service type and user needs, monitor the virtual nodes in the network, and obtain the status information about the evaluation parameters of the virtual nodes when a new request arrives or a user request changes;根据所述状态信息计算每个虚拟节点的服务迁移代价；Calculate the service migration cost of each virtual node according to the state information;通过所述计算每个虚拟节点的服务迁移代价从虚拟节点中获取虚拟服务迁移代价最小的虚拟节点作为服务节点；Obtaining the virtual node with the smallest virtual service migration cost from the virtual nodes as a service node through the calculation of the service migration cost of each virtual node;选择所述服务节点对所述虚拟服务进行迁移。Selecting the service node to migrate the virtual service.2.如权利要求1所述的虚拟服务迁移方法，其特征在于，所述状态信息包括服务的时延、服务器的可用负载、迁移路径上的带宽、服务本身的大小、服务中断及恢复的代价、服务提供者信用和执行价格。2. The virtual service migration method according to claim 1, wherein the status information includes service delay, available load of the server, bandwidth on the migration path, size of the service itself, service interruption and recovery cost , service provider credit and execution price.3.如权利要求1所述的虚拟服务迁移方法，其特征在于，所述根据所述状态信息计算每个虚拟节点的服务迁移代价具体包括:3. The virtual service migration method according to claim 1, wherein said calculating the service migration cost of each virtual node according to said state information specifically comprises:根据所述状态信息建立生成基于QoS的计算模型。A calculation model based on QoS is established and generated according to the state information.4.如权利要求3所述的虚拟服务迁移方法，其特征在于，所述根据所述状态信息建立生成基于QoS的计算模型具体包括：4. The virtual service migration method according to claim 3, wherein said establishing and generating a QoS-based computing model according to said status information specifically comprises:根据所述状态信息，得到矩阵Q：According to the state information, the matrix Q is obtained:$\mathrm{<span><span>Q</span>Q</span>}<span><span>=</span>=</span>\left(\begin{array}{cccc}{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>1,1</span>1,1</span>}}& {\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>1,2</span>1,2</span>}}& <span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span>& {\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>1</span>1</span>}<span><span>,</span>,</span>\mathrm{<span><span>m</span>m</span>}}\\ {\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>2,1</span>2,1</span>}}& {\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>2,2</span>2,2</span>}}& <span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span>& {\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>2</span>2</span>}<span><span>,</span>,</span>\mathrm{<span><span>m</span>m</span>}}\\ <span><span>\&CenterDot;</span>\&Center\; Dot;</span>& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>\\ <span><span>\&CenterDot;</span>\&CenterDot;</span>& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>& \text{<span><span>\&CenterDot;</span>\&Center Dot;</span>}& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>\\ <span><span>\&CenterDot;</span>\&Center\; Dot;</span>& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>\\ {\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>n</span>no</span>}<span><span>,</span>,</span>\mathrm{<span><span>1</span>1</span>}}& {\mathrm{<span><span>q</span>q</span>}}_{\text{<span><span>n,2</span>n,2</span>}}& <span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span>& {\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>n</span>no</span>}<span><span>,</span>,</span>\mathrm{<span><span>m</span>m</span>}}\end{array}\right)$其中，m、n为整数，q_n,m为矩阵Q中的元素，用C(v_i)表示虚拟节点v_i的迁移代价，矩阵Q中的每一行表示虚拟节点的各项状态信息，矩阵Q中的每一列表示影响所述迁移代价的一种状态信息。Among them, m and n are integers, q_{n and m} are elements in the matrix Q, C(v_i ) represents the migration cost of the virtual node v_i , each row in the matrix Q represents the state information of the virtual node, and the matrix Each column in Q represents a state information that affects the migration cost.5.如权利要求4所述的虚拟服务迁移方法，其特征在于，所述根据所述状态信息建立生成基于QoS的计算模型具体还包括：5. The virtual service migration method according to claim 4, wherein said establishing and generating a QoS-based computing model according to said state information specifically further comprises:对所述矩阵Q进行归一化，将不同量纲的所述状态信息标准化为无量纲的标准化参数，形成一个统一的衡量标准；normalizing the matrix Q, and standardizing the state information of different dimensions into dimensionless standardized parameters to form a unified measurement standard;将所述标准化参数进行分组，每组包括多种所示标准化参数，并按组操作，得出分组后各个虚拟节点的迁移代价信息。The standardized parameters are grouped, and each group includes a variety of normalized parameters shown, and are operated on a group basis to obtain migration cost information of each virtual node after grouping.6.如权利要求5所述的虚拟服务迁移方法，其特征在于，所述对所述矩阵Q进行归一化，将不同量纲的所述状态信息标准化为无量纲的标准化参数，形成一个统一的衡量标准具体包括：6. The virtual service migration method according to claim 5, wherein the matrix Q is normalized, and the state information of different dimensions is standardized into dimensionless standardized parameters to form a unified The measurement criteria specifically include:第一矩阵用N={n₁,n₂,...,n_m}表示，其中n_j=0或1，1≤j≤m；The first matrix is represented by N={n₁ ,n₂ ,...,n_m }, where n_j =0 or 1, 1≤j≤m;第二矩阵用C={c₁,c₂,...,c_m}表示，c_j是一个常数，1≤j≤m；The second matrix is represented by C={c₁ ,c₂ ,...,c_m }, c_j is a constant, 1≤j≤m;对矩阵Q中的每个元素用以下公式进行归一化：Normalize each element in the matrix Q with the following formula:${{\mathrm{<span><span>q</span>q</span>}}^{<span><span>\&prime;</span>\&prime;</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}<span><span>=</span>=</span>\left\{\begin{array}{c}\frac{{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}{\frac{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}^{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}<span><span>,</span>,</span>\mathrm{<span><span>if</span>if</span>}\frac{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}^{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}<span><span>\&NotEqual;</span>\&NotEqual;</span>\mathrm{<span><span>0</span>0</span>}<span><span>,</span>,</span>\mathrm{<span><span>and</span>and</span>}<span><span>,</span>,</span>\frac{{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}{\frac{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}^{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}<span><span><</span><</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>,</span>,</span>\mathrm{<span><span>amd</span>amd</span>}<span><span>,</span>,</span>{\mathrm{<span><span>n</span>no</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}\\ {\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>,</span>,</span>\mathrm{<span><span>if</span>if</span>}\frac{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}^{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}<span><span>=</span>=</span>\mathrm{<span><span>0</span>0</span>}<span><span>,</span>,</span>\mathrm{<span><span>and</span>and</span>}<span><span>,</span>,</span>{\mathrm{<span><span>n</span>no</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}<span><span>,</span>,</span>\mathrm{<span><span>or</span>or</span>}<span><span>,</span>,</span>\frac{{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}{\frac{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}^{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}\text{<span><span>\&GreaterEqual;</span>\&Greater Equal;</span>}{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>j</span>j</span>}}\end{array}\right.$${{\mathrm{<span><span>q</span>q</span>}}^{<span><span>\&prime;</span>\&prime;</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}<span><span>=</span>=</span>\left\{\begin{array}{c}\frac{\frac{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}^{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}{{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}<span><span>,</span>,</span>\mathrm{<span><span>if</span>if</span>}<span><span>,</span>,</span>{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}<span><span>\&NotEqual;</span>\&NotEqual;</span>\mathrm{<span><span>0</span>0</span>}<span><span>,</span>,</span>\mathrm{<span><span>and</span>and</span>}<span><span>,</span>,</span>{\mathrm{<span><span>n</span>no</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>=</span>=</span>\mathrm{<span><span>0</span>0</span>}<span><span>,</span>,</span>\mathrm{<span><span>and</span>and</span>}<span><span>,</span>,</span>\frac{\frac{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}^{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}{{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}<span><span><</span><</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>j</span>j</span>}}\\ {\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>,</span>,</span>\mathrm{<span><span>if</span>if</span>}<span><span>,</span>,</span>{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}<span><span>=</span>=</span>\mathrm{<span><span>0</span>0</span>}<span><span>,</span>,</span>\mathrm{<span><span>and</span>and</span>}<span><span>,</span>,</span>{\mathrm{<span><span>n</span>no</span>}}_{\mathrm{<span><span>j</span>j</span>}}<span><span>=</span>=</span>\mathrm{<span><span>0</span>0</span>}<span><span>,</span>,</span>\mathrm{<span><span>or</span>or</span>}<span><span>,</span>,</span>\frac{\frac{\mathrm{<span><span>1</span>1</span>}}{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>\&Sigma;</span>\&Sigma;</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>=</span>=</span>\mathrm{<span><span>1</span>1</span>}}^{\mathrm{<span><span>n</span>no</span>}}{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}{{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>i</span>i</span>}<span><span>,</span>,</span>\mathrm{<span><span>j</span>j</span>}}}<span><span>\&GreaterEqual;</span>\&Greater\; Equal;</span>{\mathrm{<span><span>c</span>c</span>}}_{\mathrm{<span><span>j</span>j</span>}}\end{array}\right.$其中，

为矩阵Q中第j种代价标准的平均值，得到矩阵Q'：in,

is the average value of the jth cost standard in the matrix Q, and the matrix Q' is obtained:${\mathrm{<span><span>Q</span>Q</span>}}^{<span><span>\&prime;</span>\&prime;</span>}<span><span>=</span>=</span>\left(\begin{array}{cccc}{{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>1,1</span>1,1</span>}}}^{\text{<span><span>\&prime;</span>\&prime;</span>}}& {{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>1,2</span>1,2</span>}}}^{<span><span>\&prime;</span>\&prime;</span>}& <span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span>& {{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>1</span>1</span>}<span><span>,</span>,</span>\mathrm{<span><span>m</span>m</span>}}}^{<span><span>\&prime;</span>\&prime;</span>}\\ {{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>2,1</span>2,1</span>}}}^{<span><span>\&prime;</span>\&prime;</span>}& {{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>2,2</span>2,2</span>}}}^{<span><span>\&prime;</span>\&prime;</span>}& <span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span>& {{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>2</span>2</span>}<span><span>,</span>,</span>\mathrm{<span><span>m</span>m</span>}}}^{<span><span>\&prime;</span>\&prime;</span>}\\ <span><span>\&CenterDot;</span>\&Center\; Dot;</span>& <span><span>\&CenterDot;</span>\&CenterDot;</span>& <span><span>\&CenterDot;</span>\&CenterDot;</span>& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>\\ <span><span>\&CenterDot;</span>\&CenterDot;</span>& <span><span>\&CenterDot;</span>\&CenterDot;</span>& \text{<span><span>\&CenterDot;</span>\&CenterDot;</span>}& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>\\ <span><span>\&CenterDot;</span>\&Center\; Dot;</span>& <span><span>\&CenterDot;</span>\&CenterDot;</span>& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>& <span><span>\&CenterDot;</span>\&Center\; Dot;</span>\\ {{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>n</span>no</span>}<span><span>,</span>,</span>\mathrm{<span><span>1</span>1</span>}}}^{<span><span>\&prime;</span>\&prime;</span>}& {{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>n</span>no</span>}<span><span>,</span>,</span>\mathrm{<span><span>2</span>2</span>}}}^{<span><span>\&prime;</span>\&prime;</span>}& <span><span>\&CenterDot;</span>\&CenterDot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span><span><span>\&CenterDot;</span>\&Center\; Dot;</span>& {{\mathrm{<span><span>q</span>q</span>}}_{\mathrm{<span><span>n</span>no</span>}<span><span>,</span>,</span>\mathrm{<span><span>m</span>m</span>}}}^{<span><span>\&CenterDot;</span>\&CenterDot;</span>}\end{array}\right)$其中，q_i,j为矩阵Q中的元素，q_n,m'为矩阵Q'中的元素。Wherein, q_{i, j} are elements in the matrix Q, and q_{n, m} ' are elements in the matrix Q'.7.如权利要求1所述的虚拟服务迁移方法，其特征在于，所述方法还包括：当没有新请求到达或用户请求发送未发生改变时不进行虚拟节点的迁移。7 . The virtual service migration method according to claim 1 , further comprising: not migrating the virtual node when no new request arrives or the transmission of the user request does not change. 8 .

Images