CN103532759A - Cloud service-oriented admission control method for converged flows - Google Patents

Abstract

The invention discloses a cloud service-oriented admission control method for converged flows, and belongs to the field of network management and network QoS (quality of service) assurance. The admission control method comprises the following steps of initiating a request of accessing cloud service by a user side, shaping through a shaper to obtain a plurality of shaped business flows, and converging the shaped business flows to produce the converged flows through multiplexing of a multiplexer; computing the effective bandwidth EB and/or the equivalent capacity EC of the converged flows by adopting a network calculus method; admitting the user request when the effective bandwidth EB is not bigger than C, and otherwise, rejecting the request; admitting the user request when the equivalent capacity EC is not bigger than C to allow a user to enter a network domain, and otherwise, rejecting the request. Compared with a traditional admission control method for uniflows, the admission control method for the converged flows disclosed by the invention is capable of admitting more business flows when the node output rate is fixed, i.e., the network throughput is increased.

Description

The acceptance controlling method of the aggregated flow of cloud service-oriented

Technical field

The present invention relates to network management technology field, relate in particular to network QoS guarantee and the Admission control of cloud service-oriented.

Background technology

Development along with cloud computing, the explosive growth of cloud service value volume and range of product, user wants to accept the service in cloud, must be through traditional network, this is faced with the legacy network of limited resources (as bandwidth, buffer length etc.) may to occur at any time congested and causes the threat that cloud service quality (Quality of Service, QoS) declines.Add that current user is more and more higher to quality of service requirement, the service of network tradition " doing one's best " can not meet user's QoS demand.And receive to control as a kind of effective preventative flow control means, be avoid network congestion, (QoS) important technology that ensures of realizing network service quality.Implement to receive the network of controlling, require user by the traffic transmission feature of oneself and the service quality of requirement, to tell network when request access network, the resource situation that network based user's traffic characteristics and network are existing, whether decision receives user's connection request.

In recent years, along with the rise of cloud computing, the acceptance controlling method that the QoS of cloud service-oriented ensures has become the focus of research.Traditional acceptance controlling method is the application traffic model based on precognition or simplification usually; as the arrival of all application streams being all considered as to Poisson process and supposing their obedience independent same distribution; in fact; major applications stream; as the Business Stream of video, cloud service etc. all has long-range correlation, therefore traditional acceptance controlling method inapplicable.Meanwhile, because various application on network flow to the randomness reaching, the traffic characteristic after multiple stream converges is also difficult to describe, and traditional method is difficult to construct its discharge model.Network calculus is a kind of new quantitative tool of mathematical analysis of effective network performance, compare with conventional statistics theory, its great advantage is that it can provide certainty marginal analysis to network performance, utilize Network calculus can calculate the discharge model after the polymerization of isomery stream, also can analyze the feature of long-range correlation stream.Main research at present has Chen Yanping to adopt the statistical property of the method computing service stream of on-line measurement to estimate that its burst level is to select suitable traffic models, be different from traditional acceptance controlling method and only take a kind of traffic models, traffic models based on selected, Transformation Theorem by effective bandwidth and effective envelope, obtains how fractal and effective envelope self-similar traffic.Then, adopt statistics network calculation to obtain qos parameter and receive control decision, and by selecting receiving to reduce to receive the impact of control on QoS itself opportunity, experimental result shows that the admission control algorithm based on the method has obtained higher receiving ability, but she does not consider the feature of aggregated flow, and the acceptance controlling method to aggregated flow.LE Boudec professor J-Y has proposed to utilize the data volume of Business Stream to describe the method for estimating the resource requirement of stream with parametric description, provided (the effective bandwidth of the effective bandwidth based on delay constraint that flows multiplexing rear aggregated flow, EB) equivalent capacity (the equivalent capacity and based on buffer length, EC) definition, and propose to utilize them to carry out receiving control.But for polytype cloud service, he does not provide the EB of aggregated flow and the model of EC, and the performance of analyzing their feature and the control of the receiving based on EB and EC.

Summary of the invention

For above deficiency of the prior art, the object of the present invention is to provide a kind of throughput that improves network, guarantee the acceptance controlling method of aggregated flow of the cloud service-oriented of cloud service QoS.Technical scheme of the present invention is as follows:

A kind of acceptance controlling method of aggregated flow of cloud service-oriented, wherein the network model of cloud service comprises reshaper, network domains and cloud service territory, described reshaper is connected with user side, network domains comprises entrance router node and egress router node, reshaper is connected with the ingress router node of network domains, the egress router node of network domains is connected with cloud service territory, and the receiving of the aggregated flow of cloud service-oriented is controlled and comprised the following steps:

101, when user side is initiated the request of access cloud service, the Business Stream of this request is obtained to shape traffic streams by reshaper shaping, some shape traffic streams converge at the ingress router node of network domains, and the multiplexing generation aggregated flow of process multiplexer, wherein drop multiplexer apparatus is placed in ingress router node, and ingress router node is also referred to as remultiplexer node herein;

102, Adoption Network the method fluxions to the effective bandwidth EB of aggregated flow in step 101 or/and equivalent capacity EC calculate;

During 103, as the given delay constraint D of user side and as effective bandwidth EB≤C, the request that user initiates to access cloud service is accepted, and allows to enter network domains, otherwise refusal request, wherein C is the constant output speed of ingress router node;

During as the given buffer length of network terminal constraint B and as equivalent capacity EC≤C, the request that user initiates to access cloud service is accepted, and allows to enter network domains, and wherein C is the constant output speed of ingress router node, otherwise refusal is asked;

When the given delay constraint D of user side and the given buffer length of network terminal constraint B, when effective bandwidth EB≤C and equivalent capacity EC≤C, the request that user initiates to access cloud service is accepted, and allows to enter network domains, otherwise refusal request.

Further, obtain data volume time diagram in step 102, the effective bandwidth EB value of aggregated flow is for to make tangent slope at time time=-D place to the arrival curve of polymerization;

The equivalent capacity EC value of aggregated flow is for to make tangent slope at data data=B place to the arrival curve of polymerization.

Further, in step 101, adopt leaky bucket reshaper to carry out shaping to Business Stream.

Further, the arrival curve of polymerization

, and arrival curve α _imeet traffic specification T-SPEC (p _i, M _i, r _i, b _i), p wherein _i, M _i, r _i, b _irepresent respectively the peak rate of Business Stream i, maximum bag is long, Mean Speed and burst tolerance, n _ithe quantity that represents the Business Stream of i type.

Further, the acceptance region that in step 103, user initiates to access the set of the quantity that the request of cloud service is accepted is (n ₁, n ₂..., n _i), set the EB of aggregated flow and the unified expression formula F (α, X) after EC mapping, F (α, X)≤C conversion can be obtained to following formula:

$\left\{\begin{array}{c}{n}_i\≤g_1(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\\ n_i\≤g_2(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\\ n_i\≤g_3(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\\ .\\ .\\ .\\ n_i\≤G_{I+1}(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\\ n_i\≤G_{I+2}(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\end{array}\right.$

Acceptance region (n ₁, n ₂..., n _i) be exactly the disaggregation of this inequality group, and n ₁, n ₂..., n _i∈ N.

Advantage of the present invention and beneficial effect are as follows:

The present invention considers the diversity of cloud service, utilizes Network calculus to calculate EB and the EC of aggregated flow, solves the inenarrable problem of traffic characteristic after the polymerization of isomery stream.By calculating the arrival curve of the polymerization of polytype Business Stream, according to the arrival curve of polymerization and given constraint, estimate the bandwidth resources that aggregated flow is required, the namely EB of aggregated flow or EC, admission control cell is carried out to receive according to the bandwidth resources situation of the EB of aggregated flow or EC and node and is controlled, and reaches the object of avoiding network congestion, realizing cloud service QoS guarantee with this.Isomery stream is in buffering area of multiplexing nodes sharing, based on this kind of mode, than the fixedly mode of buffering area of distributing for each single current, save more bandwidth resources, namely, in the situation that node output speed is identical, the receiving based on aggregated flow is controlled and can be received more stream, improves network throughput.

Accompanying drawing explanation

Fig. 1 is network model figure of the present invention;

Fig. 2 is the multiplexing illustraton of model of the node of Business Stream of the present invention;

Fig. 3 is that illustraton of model is controlled in the receiving of aggregated flow of the present invention;

Fig. 4 is EB and the EC computational methods figure of aggregated flow of the present invention, is also the data volume time diagram of aggregated flow;

Wherein Fig. 4 A is the data volume time diagram of single type service stream, the data volume time diagram that Fig. 4 B is aggregated flow;

Fig. 5 is the acceptance controlling method flow chart of the aggregated flow of cloud service-oriented of the present invention.

Embodiment

The invention will be further elaborated below in conjunction with accompanying drawing, to provide the embodiment of an indefiniteness.But should be appreciated that, these descriptions are exemplary, and do not really want to limit the scope of the invention.In addition, in the following description, omitted the description to known features and technology, to avoid unnecessarily obscuring concept of the present invention.

Be illustrated in figure 1 the network model figure of the preferred embodiment of the present invention.Comprising: reshaper, network domains, cloud service territory.In network domains, comprise again entrance router node and egress router node, isomery stream Multiplexing module of the present invention is deployed in the ingress router node of network domains with receiving control module.

Figure 2 shows that the multiplexing model of node of Business Stream of the present invention (isomery stream), comprising shaping unit, remultiplexer node part, this remultiplexer node is wherein the ingress router node in Fig. 1 network domains namely.

With reference to Fig. 1, user initiates to access the request of cloud service, and the Business Stream of request (isomery stream) is first by reshaper shaping, and in this preferred embodiment, reshaper adopts the burst of the level and smooth stream of leaky bucket reshaper, avoids burst to make a very bad impression to network.Stream after exporting by reshaper is subject to arrival curve α _iconstraint, the ingress router node of the mixed flow of these isomeries in Fig. 1 network domains carries out multiplexing and receives and control, and the stream being accepted accesses cloud service by network domains.Note, stream be herein Business Stream;

With reference to Fig. 2, the stream of service request

by the stream after reshaper i shaping, meet arrival curve and be constrained to α _i, wherein

representing has n in i type of Business Stream _ibar stream, the stream after shaping is multiplexing and receive control by ingress router node, i=1 wherein, 2 ..., I.

In step 102, the acceptance controlling method to aggregated flow, and the process that solves acceptance region is described in detail.

1. calculate the arrival curve of polymerization

Isomery for polytype cloud service request flows the arrival curve α after passing through reshaper _imeet traffic specification T-SPEC (p _i, M _i, r _i, b _i), p wherein _i, M _i, r _i, b _ithe peak rate of i is flowed in representative respectively, and maximum bag is long, Mean Speed and the tolerance that happens suddenly.As shown in Figure 4, the namely arrival curve summation to single current of the arrival curve α of polymerization.Be:

$\mathrm{\α}=\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{n}_i{\mathrm{\α}}_i---\left(3\right)$

2. calculate EB and the EC of aggregated flow

The computational methods of the effective bandwidth of aggregated flow (EB) and equivalent capacity (EC) as shown in Figure 4, effective bandwidth for aggregated flow, namely meeting given delay constraint D, preferred the present embodiment D that gives an example is under the situation of 100ms, the needed bandwidth of aggregated flow; The equivalent capacity of aggregated flow refers to and is constrained to B meeting given buffer length, and in the situation of not overflowing of buffering area, the needed bandwidth of aggregated flow.With reference to Fig. 4, it is also the data volume time diagram of aggregated flow; Wherein Fig. 4 A is the data volume time diagram of single type service stream, the data volume time diagram that Fig. 4 B is aggregated flow;

The equivalent bandwidth of aggregated flow is exactly to the arrival curve of polymerization, to make tangent slope at time=-D place.

For there being I type, and i (i=1,2 ..., the quantity of stream I) type is n _istream, by the arrival curve α of the stream after reshaper _ifor T-SPEC (p _i, M _i, r _i, b _i), suppose that the delay constraint of each type stream is all D, if hypothesis is false, consider that acceptance controlling method of the present invention guarantees for every kind of cloud service provides definite QoS, gets D=min (D _i).The value of the EB of aggregated flow is following formula:

Wherein, ${\mathrm{\τ}}_1=\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{n}_i{M}_i/\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{n}_i{p}_i,{\mathrm{\τ}}_2=(n_1{b}_1+\underset{i=2}{\overset{I}{\mathrm{\Σ}}}{n}_i{M}_i)/(n_1{r}_1+\underset{i=2}{\overset{I}{\mathrm{\Σ}}}{n}_i{p}_i),$

${\mathrm{\τ}}_3=(\underset{i=1}{\overset{2}{\mathrm{\Σ}}}{n}_i{b}_i+\underset{i=3}{\overset{I}{\mathrm{\Σ}}}{n}_i{M}_i)/(\underset{i=1}{\overset{2}{\mathrm{\Σ}}}{n}_i{r}_i+\underset{i=3}{\overset{I}{\mathrm{\Σ}}}{n}_i{p}_i),$

${\mathrm{\τ}}_I=(\underset{i=1}{\overset{I-1}{\mathrm{\Σ}}}{n}_i{b}_i+n_I{M}_I)/(\underset{i=1}{\overset{I-1}{\mathrm{\Σ}}}{n}_i{r}_i+n_I{p}_I),{\mathrm{\τ}}_{I+1}=\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{n}_i{b}_i/\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{n}_i{r}_i$

The EB that can draw aggregated flow from (4) formula is a piecewise function, and along with the value of delay constraint D is different, the expression formula of EB is different.In formula

The EB of aggregated flow meets following formula:

$e_D\left(\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{\mathrm{\α}}_i\right)\≤\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{e}_D\left(\mathrm{\α}\right)---\left(5\right)$

It is the effective bandwidth sum that the effective bandwidth of aggregated flow is less than or equal to each single current, namely foregoing while considering by aggregated flow than considering required bandwidth still less by single current, the output speed C of ingress router node in Fig. 1 so, when in preferred the present embodiment, C is 1.5M/s, the acceptance controlling method based on aggregated flow can be received more stream.

With reference to the method shown in Fig. 4, the equivalent capacity of aggregated flow is exactly to the arrival curve of polymerization, to make tangent slope at data volume data=B place.Preferred the present embodiment B that gives an example is 100M, and the expression formula of the EC of aggregated flow is following formula.

Wherein, $B_1=\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{n}_i{M}_i,B_2=n_1{b}_1+\underset{i=2}{\overset{I}{\mathrm{\Σ}}}{n}_i{M}_i,B_3=\underset{i=1}{\overset{2}{\mathrm{\Σ}}}{n}_i{b}_i+\underset{i=3}{\overset{I}{\mathrm{\Σ}}}{n}_i{M}_i,$

$B_{I-1}=\underset{i=1}{\overset{I-2}{\mathrm{\Σ}}}{n}_i{b}_i+\underset{i=I-1}{\overset{I}{\mathrm{\Σ}}}{n}_i{M}_i,B_I=\underset{i=1}{\overset{I-1}{\mathrm{\Σ}}}{n}_i{b}_i+n_I{M}_I$

Similar to equivalent bandwidth, equivalent capacity also has following character.

$f_B\left(\underset{i=1}{\overset{I}{\mathrm{\Σ}}}\mathrm{\α}\right)\≤\underset{i=1}{\overset{I}{\mathrm{\Σ}}}{f}_{B_i}\left({\mathrm{\α}}_i\right)---\left(7\right)$

Wherein, it is the equivalent capacity sum that the equivalent capacity of aggregated flow is not more than each single current.The acceptance controlling method that is aggregated flow has higher throughput than the acceptance controlling method of single current.

3. the acceptance controlling method of aggregated flow

In the ingress router node of receiving control module deployment diagram 1 as shown in Figure 3.Receiving and controlling control unit (ACU) is the chief component of receiving control module, and admission control cell (ACU) mainly consists of with receiving decision-making module (ADM) data flow estimation module (TEM), resource estimation module (REM).Data volume estimation module utilizes Network calculus to estimate the bandwidth requirement of aggregated flow, i.e. definite EB and the EC of aggregated flow.

Its concrete control procedure of receiving is as follows:

Aggregated flow for cloud service request, data volume estimation module (TEM) is estimated its needed bandwidth according to the parametric description of Business Stream (parameters of foregoing flow stipulations) and the requirement (being delay constraint D or buffer length constraint B) of user or network, the bandwidth of estimating is the EB of aggregated flow or the value of EC namely, and its method of estimation is exactly the EB of the aggregated flow of calculating described in detail or the method for EC in step 1,2.Resource estimation module (REM) is estimated the resource situation of multiplexing node.ADM is according to the input of TEM and REM, and can the stream of judgement request be accepted.Concrete judgement is according to being: if the required bandwidth of aggregated flow is no more than the output speed of node, the stream of request is accepted, and flowing to into network of being accepted transmitted; Otherwise the stream of request is rejected.

If user is given delay constraint D, TEM calculates the effective bandwidth of aggregated flow, and the decision-making technique of the control of the receiving based on effective bandwidth (EBAC) is (1) formula; If given meshwork buffering section length is constrained to B, TEM calculates the equivalent capacity of aggregated flow, and the decision-making technique of the control of the receiving based on equivalent capacity (ECAC) is (2) formula; If user is given delay constraint D, and given meshwork buffering section length is constrained to B, and TEM calculates effective bandwidth and the equivalent capacity of aggregated flow, and to control the decision-making technique of (E2AC) be (1) and (2) formula to the receiving based on equivalent capacity and equivalent capacity.

4. solve acceptance region

Before narrated concrete enforcement and received the method for controlling, next will provide the process that specifically solves acceptance region, the node namely solving based on acceptance controlling method noted earlier can be received the quantity flowing.The object that we solve is in order to analyze the performance of acceptance controlling method proposed by the invention.

The process that solves acceptance region is as follows:

Because EBAC, ECAC are similar with the mode that E2AC method solves acceptance region, so being mapped as general expression formula by the effective bandwidth EB of aggregated flow and equivalent capacity EC, we solve acceptance region.Delay constraint D and buffering area constraint B are mapped as to a constraint X that unification is given, and the effective bandwidth of aggregated flow (EB) is mapped as following expression formula with equivalent capacity EC:

e _D(α)→F(α,X)←f _B(α) (8)

The effective bandwidth (EB) of aggregated flow and the expression formula of equivalent capacity (EC) with reference to (4) formula with (6) formula, we know that EB, the EC expression formula of aggregated flow is a piecewise function, in order to analyze unified solution procedure, now each section of expression formula of piecewise function is mapped as to following formula:

f _i(n ₁,n ₂,…,n _I,X) (9)

In conjunction with (8) formula, (9) formula, the EB of aggregated flow, EC expression formula (4) and (6) can be mapped as a unified expression formula, as follows:

$F(\mathrm{\&alpha;},X)=\left\{\begin{array}{c}{f}_1(n_1,n_2,...,n_I,X),0<X\&le;X_1\\ f_2(n_1,n_2,...,n_I,X),X_1<X\&le;X_2\\ f_3(n_1,n_2,...,n_I,X),X_2<X\&le;X_3\\ .\\ .\\ .\\ f_{I+1}(n_1,n_2,...,n_I,X),X_{I-1}<X\&le;X_{I+1}\\ f_{I+2}(n_1,n_2,...,n_I,X),X>X_{I+1}\end{array}\right.---\left(10\right)$

Known according to (4) formula and (6) formula, the EB of aggregated flow and the value of EC and the receive quantity n flowing _irelevant, namely when given delay constraint D and buffering area constraint B, can not determine in advance D or B(or X) belong to which segment limit in the domain of definition, therefore, can not determine the concrete expression formula of EB and the EC of aggregated flow, EBAC, ECAC and E2AC method can not be determined the concrete formula of receiving decision-making in advance so, also just can not calculate the quantity that acceptance controlling method can be received stream according to pre-determined concrete receiving decision-making expression formula.But known with the receiving controlling decision method of (2) formula according to (1) formula, F (α, X)≤C, that is:

$\left\{\begin{array}{c}{f}_1(n_1,n_2,...,n_I,X)\&le;C\\ f_2(n_1,n_2,...,n_I,X)\&le;C\\ f_3(n_1,n_2,...,n_I,X)\&le;C\\ .\\ .\\ .\\ f_{I+1}(n_1,n_2,...,n_I,X)\&le;C\\ f_{I+2}(n_1,n_2,...,n_I,X)\&le;C\end{array}\right.---\left(11\right)$

The conversion of (11) formula can be obtained to following formula:

$\left\{\begin{array}{c}{n}_i\&le;g_1(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\\ n_i\&le;g_2(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\\ n_i\&le;g_3(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\\ .\\ .\\ .\\ n_i\&le;g_{I+1}(n_1,n_2,...n_{i-1},n_{i+1},{...n}_I,X,C)\\ n_i\&le;g_{I+2}(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\end{array}\right.---\left(12\right)$

So far, the value of acceptance region is the disaggregation of inequality group formula (12) namely, and n ₁, n ₂..., n _i∈ N.Collection (n in acceptance region so ₁, n ₂..., n _i) be exactly the quantity that in the stream of I type, the stream of each type can be accepted.So, can pass through acceptance region, can analyze the performance of the acceptance controlling method under different parameters X.Because the receiving decision-making expression formula of E2AC method is simultaneously based on (1) and (2) formula, therefore for the method for solving of the acceptance region of E2AC method, difference is only that the inequality in the inequality group of (11) formula is the formula that simultaneously meets (1) and the receiving decision-making technique of (2) formula.If I=2, acceptance region can represent with plane right-angle coordinate, a point (n in acceptance region ₁, n ₂) quantity that can be accepted corresponding to the stream of every type; If I=3, acceptance region can represent by space coordinates, a point (n in acceptance region ₁, n ₂, n ₃) quantity that can be accepted corresponding to the stream of every type; If I>=3, can calculate the acceptance region of the acceptance controlling method of aggregated flow, but are difficult to represent intuitively acceptance region by the mode of how much.

Referring to Fig. 5, the receiving of the aggregated flow of cloud service-oriented is controlled and is comprised the following steps:

101, when user side is initiated the request of access cloud service, the Business Stream of this request is obtained to shape traffic streams by reshaper shaping, some shape traffic streams converge at the ingress router node of network domains, and the multiplexing generation aggregated flow of process multiplexer, wherein drop multiplexer apparatus is placed in ingress router node;

102, Adoption Network the method fluxions to the effective bandwidth EB of aggregated flow in step 101 or/and equivalent capacity EC calculate;

During 103, as the given delay constraint D of user side and as effective bandwidth EB≤C, the request that user initiates to access cloud service is accepted, and allows to enter network domains, otherwise refusal request, wherein C is the constant output speed of ingress router node;

During as the given buffer length of network terminal constraint B and as equivalent capacity EC≤C, the request that user initiates to access cloud service is accepted, and allows to enter network domains, and wherein C is the constant output speed of ingress router node, otherwise refusal is asked; When the given delay constraint D of user side and the given buffer length of network terminal constraint B, when effective bandwidth EB≤C and equivalent capacity EC≤C, the request that user initiates to access cloud service is accepted, and allows to enter network domains, otherwise refusal request.

The acceptance controlling method of the aggregated flow of the cloud service-oriented that the present invention proposes makes legacy network be faced with congested and causes the problem that cloud service QoS declines for the service diversity in cloud environment, and conventional method flows the inenarrable problem of feature flowing after polymerization to isomery, utilize Network calculus to analyze the feature and the bandwidth of estimating that aggregated flow is required of aggregated flow, be the effective bandwidth based on delay constraint or the equivalent capacity based on buffer length constraint of aggregated flow, acceptance controlling method is exactly whether over the output speed of node, to do receiving decision-making according to the bandwidth of estimating aggregated flow, with this, guarantee the QoS demand of cloud service.Share the buffering area of flowing when node is multiplexing by isomery, method based on aggregated flow required bandwidth when pressing single current and consider is few, therefore the acceptance controlling method based on aggregated flow has higher receiving ability than the acceptance controlling method based on single current, has improved the throughput of network.Meanwhile, in order to analyze the acceptance controlling method performance of aggregated flow, provided the concrete method for solving of acceptance region.

These embodiment are interpreted as only for the present invention is described, is not used in and limits the scope of the invention above.After having read the content of record of the present invention, technical staff can make various changes or modifications the present invention, and these equivalences change and modify the acceptance controlling method claim limited range of the aggregated flow that falls into equally cloud service-oriented of the present invention.

Claims (5)

1. the acceptance controlling method of the aggregated flow of a cloud service-oriented, wherein the network model of cloud service comprises reshaper, network domains and cloud service territory, described reshaper is connected with user side, network domains comprises entrance router node and egress router node, reshaper is connected with the ingress router node of network domains, the egress router node of network domains is connected with cloud service territory, it is characterized in that, the receiving of the aggregated flow of cloud service-oriented is controlled and comprised the following steps:

101, when user side is initiated the request of access cloud service, the Business Stream of this request is obtained to shape traffic streams by reshaper shaping, some shape traffic streams converge at the ingress router node of network domains, and the multiplexing generation aggregated flow of process multiplexer, wherein drop multiplexer apparatus is placed in ingress router node;

102, Adoption Network the method fluxions to the effective bandwidth EB of aggregated flow in step 101 or/and equivalent capacity EC calculate;

During 103, as the given delay constraint D of user side and as effective bandwidth EB≤C, the request that user initiates to access cloud service is accepted, and allows to enter network domains, otherwise refusal request, wherein C is the constant output speed of ingress router node;

During as the given buffer length of network terminal constraint B and as equivalent capacity EC≤C, the request that user initiates to access cloud service is accepted, and allows to enter network domains, and wherein C is the constant output speed of ingress router node, otherwise refusal is asked;

When the given delay constraint D of user side and the given buffer length of network terminal constraint B, when effective bandwidth EB≤C and equivalent capacity EC≤C, the request that user initiates to access cloud service is accepted, and allows to enter network domains, otherwise refusal request.

2. the acceptance controlling method of the aggregated flow of cloud service-oriented according to claim 1, it is characterized in that: in step 102, obtain data volume time diagram, the effective bandwidth EB value of aggregated flow is for to make tangent slope at time time=-D place to the arrival curve of polymerization;

The equivalent capacity EC value of aggregated flow is for to make tangent slope at data data=B place to the arrival curve of polymerization.

3. the acceptance controlling method of the aggregated flow of cloud service-oriented according to claim 1, is characterized in that: in step 101, adopt leaky bucket reshaper to carry out shaping to Business Stream.

4. the acceptance controlling method of the aggregated flow of cloud service-oriented according to claim 2, is characterized in that: the arrival curve of polymerization

, and arrival curve α _imeet traffic specification T-SPEC (p _i, M _i, r _i, b _i), p wherein _i, M _i, r _i, b _irepresent respectively the peak rate of Business Stream i, maximum bag is long, Mean Speed and burst tolerance, n _ithe quantity that represents the Business Stream of i type.

5. the acceptance controlling method of the aggregated flow of cloud service-oriented according to claim 1, is characterized in that: in step 103 user to initiate to access the acceptance region of the set of the quantity that the request of cloud service is accepted be (n ₁, n ₂..., n _i), set the EB of aggregated flow and the unified expression formula F (α, X) after EC mapping, F (α, X)≤C conversion can be obtained to following formula:

$\left\{\begin{array}{c}{n}_i\&le;g_1(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\\ n_i\&le;g_2(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\\ n_i\&le;g_3(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\\ .\\ .\\ .\\ n_i\&le;G_{I+1}(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\\ n_i\&le;G_{I+2}(n_1,n_2,...n_{i-1},n_{i+1},...,n_I,X,C)\end{array}\right.$

Acceptance region (n ₁, n ₂..., n _i) be exactly the disaggregation of this inequality group, and n ₁, n ₂..., n _i∈ N.

Images