Summary of the invention
For overcome on the active computer heterogeneous network that type in the distribution of traffic flows is difficult to distinguish and preferentially distribution be difficult to the defective that guarantees and be difficult to determine automatically the distribution of traffic flows weight, the present invention proposes a kind of service flow distribution system and method for multi-module cooperation communication.
According to an aspect of the present invention, proposed a kind of service flow distribution system of multi-module cooperation communication, having comprised:
Sensing module is used for the link circuit condition of the network that perception connects, and obtains the network information and Business Stream information;
Judge module is used for determining Business Stream priority and filtering out available network according to the described network information and the Business Stream information that receive;
Object module is used for user's request is converted into the decision objective that can weigh according to relative approach degree;
Decision-making module, according to receiving from the usable network information of described judge module with from the decision objective information of described object module, use multi-parameter weight self-generating algorithm and TOPSIS multiple attributive decision making method, obtain the approach degree of described available network and theoretical optimal network, determine to satisfy the optimal network of each Business Stream.
Wherein, described sensing module is according to the network online measurement technology, and the current link circuit condition of each network that perception connected comprises available bandwidth, signal strength signal intensity and time delay, shake, the packet loss of network; And according to the type of service cognition technology, perception needs serviced traffic flow types.
Wherein, described judge module is determined Business Stream priority and formed priority query according to following principle: the priority of real time business is higher than non-real-time service, and interactive service priority is higher than the non-interactive type business, and service switchover priority is higher than new access service.
Wherein, described judge module is chosen the available network collection that satisfies this Business Stream requirement according to the network qualifications of highest priority traffic stream.
Wherein, in the described object module, described relative approach degree be restricted to be used to weigh available network and theoretical optimal network press close to degree and with the variable of the poorest network of theory away from degree.
Wherein, described decision-making module uses existing multi-parameter weight self-generating algorithm according to described usable network information, generates the weight of network parameters correspondence; Described decision-making module is according to described usable network information and described network parameters respective weights, uses the TOPSIS multiple attributive decision making method, calculates the relative approach degree of each available network and desirable optimal network.
Wherein, described TOPSIS multiple attributive decision making method comprises: by the theoretical optimal network and the poorest network of theory of structure evaluating network, and with near optimal network with away from the degree of poor network, calculate available network and come available network is sorted, select optimum available network with the relative approach degree of desirable optimal network.
According to a further aspect in the invention, proposed a kind of distribution of traffic flows method of multi-module cooperation communication, having comprised:
The link circuit condition of step 10), network that perception connected obtains the network information and Business Stream information;
Step 20), according to the described network information and the Business Stream information that receive, determine Business Stream priority and filter out available network;
Step 30), user's request is converted into the decision objective that can weigh according to relative approach degree;
Step 40), according to the described usable network information and the described decision objective information that receive, use multi-parameter weight self-generating algorithm and TOPSIS multiple attributive decision making method, obtain the approach degree of described available network and theoretical optimal network, determine to satisfy the optimal network of each Business Stream.
Wherein, step 10) also comprises: according to the network online measurement technology, the current link circuit condition of each network that perception connected comprises available bandwidth, signal strength signal intensity and time delay, shake, the packet loss of network; And according to the type of service cognition technology, perception needs serviced traffic flow types.
Wherein, step 20) comprise that also determine Business Stream priority and form priority query according to following principle: the priority of real time business is higher than non-real-time service, and interactive service priority is higher than the non-interactive type business, and service switchover priority is higher than new access service.
Wherein, step 20) also comprise:, choose the available network collection that satisfies described Business Stream requirement according to the network qualifications of highest priority traffic stream.
Wherein, step 30) in, described relative approach degree be restricted to be used to weigh available network and theoretical optimal network press close to degree and with the variable of the poorest network of theory away from degree.
Wherein, step 40) also comprise: according to described usable network information, use existing multi-parameter weight self-generating algorithm, generate the weight of network parameters correspondence; According to described usable network information and described network parameters respective weights, use the TOPSIS multiple attributive decision making method, calculate the relative approach degree of each available network and theoretical optimal network.
Wherein, step 40) in, described TOPSIS multiple attributive decision making method comprises: by the theoretical optimal network and the poorest network of theory of structure evaluating network, and with near optimal network with away from the degree of poor network, calculate available network and come available network is sorted, select optimum available network with the relative approach degree of desirable optimal network.
According to the present invention,,, guarantee that when selecting only distributing network important Business Stream has the priority of using network for Business Stream for each bar Business Stream is given different service priority based on the feature of different kinds of business stream; The present invention introduces multi-parameter weight self-generating algorithm, is used for definite automatically weight that influences every network parameter correspondence of distribution of traffic flows decision-making, alleviates burden for users.
Embodiment
Be described in detail below in conjunction with the service flow distribution system and the method for the drawings and specific embodiments a kind of multi-module cooperation communication provided by the invention.
Fig. 1 as shown in the figure, provides the composition topology example of next generation mobile networks for multi-module cooperation communication typical scene figure in the existing heterogeneous network, and next generation mobile networks is a heterogeneous network that plurality of access modes merges mutually.Various wireless access technologys differ from one another, multiple access network coexistence, and for example: high bandwidth can be provided in the focus covering but coverage is little; Big but the limited bandwidth of the coverage that honeycomb covers; Even same type overlay network, the difference of realization technology or operator also can have influence on user's access and select.Therefore be difficult to replace other all access technology with single access technology, various wireless access ways are with long-term co-existence.The target of next generation mobile networks is exactly well to merge these different Radio Access Networks, allows their have complementary advantages, thereby provides the most perfect experience for the mobile subscriber.
In addition, Fig. 1 also illustrates the typical scene of multimode terminal collaboration communication in the heterogeneous network.When having dissimilar Business Streams to transmit on the multimode terminal, can each bar distribution of traffic flows be transmitted to different network interfaces according to the feature of different kinds of business stream, the factors such as current link circuit condition of heterogeneous networks, on the basis that ensures the traffic stream service quality, obtain minimum communications cost and maximum resource utilization; In addition, when no longer available or service quality obviously reduced when certain access network, multimode terminal can in time be made a response, and loaded service stream on this network is seamlessly switched on other networks that satisfy traffic flow requirements transmit.As shown in Figure 1, this multimode terminal is equipped with multiple interfaces, can directly link to each other with cellular network, focus overlay network, and supposing can 3 kinds of sessions of concurrent running---speech, video and file transfer on the multimode terminal.3 kinds of sessions have different service features and QoS requirement, and desirable transmission plan is: voice service, because relatively more responsive, should select the good cellular network transmission of service quality to time delay, shake; Video traffic, data volume is big, delay requirement is general, preferably select focus overlay network transmission, but the user move and cause focus overlay network time spent no longer, should transmit just seamlessly switching on the honeycomb overlay network that satisfies the demands as much as possible at serviced Business Stream; The file transfer business, data volume is big, to the time delay no requirement (NR), should select the focus overlay network to transmit as far as possible, and the zone that does not have focus to cover can be transmitted under the prerequisite that does not influence the transmission of other types Business Stream.
Fig. 2 illustrates according to an embodiment of the invention the multi-module cooperation communication service flow distribution system based on priority and weight self-generating.According to existing Multiple Attribute Decision Making Theory, in conjunction with service priority notion and weight self-generating algorithm, propose a kind of multi-module cooperation communication service flow distribution system according to embodiments of the invention, comprise sensing module, judge module, decision-making module and object module based on priority and weight self-generating.
Sensing module, on the one hand according to existing network online measurement technology, the current link circuit condition of each network that perception connected, comprise two time-varying parameters (available bandwidth, signal strength signal intensity) and three service quality measurement indexs (time delay, shake, packet loss) of network, form the current collection of network Y={y that exists
1, y
2, y
3... y
m, m ∈ N, each network y
iAll use one group of parameter to characterize y
i={ y
I, 1, y
I, 2, y
I, 3... y
I, k, k ∈ N, wherein y
I, jJ parameter representing i network.On the other hand, according to existing type of service cognition technology, perception needs serviced traffic flow types, comprises that this Business Stream to the demand of bandwidth, signal strength signal intensity with to the tolerance limit of time delay, shake, packet loss, forms the Business Stream set X={x of concurrent running on the multi-module mobile terminal
1, x
2, x
3... x
n, n ∈ N, final sensing module passes to judge module with the network information and the Business Stream information that perception obtains.
Judge module, on the one hand according to the Business Stream feature, for each the bar Business Stream that receives is determined a priority, and Business Stream set is adjusted according to priority, form a priority query, in the formation each is represented Business Stream to be distributed such as, and formation is according to the priority descending, the Business Stream X of head of the queue
iBe the object of decision-making distribution next time.In general, the Business Stream Determination of priority is followed following rules: the priority of real time business is higher than non-real-time service, and interactive service priority is higher than the non-interactive type business, and service switchover priority is higher than new access service.On the other hand, judge module is according to head of the queue Business Stream X
iQualifications at aspects such as bandwidth, time delay, shake and packet loss screens collection of network, is met m available network of distribution of traffic flows, constitutes available network set Y={y
1, y
2, y
3... y
m.Available network must satisfy the demand of Business Stream on bandwidth and signal strength signal intensity, and is dropping on the features such as time delay, shake and packet loss within the limit that Business Stream can tolerate.Final judge module passes to decision-making module with judged result (usable network information).
The effect of object module is exactly that demand with the user changes into the target that can weigh comparison by certain mathematical method.In the middle of the heterogeneous network environment in future, the user wishes to obtain convenient, fast, seamless access most and experiences.Promptly when having dissimilar Business Streams to transmit on the multimode terminal, can each bar distribution of traffic flows be transmitted to different network interfaces according to the feature of different kinds of business stream, the factors such as current link circuit condition of heterogeneous networks; In addition, when no longer available or service quality obviously reduced when certain access network, multimode terminal can in time be made a response, and loaded service stream on this network is seamlessly switched on other networks that satisfy traffic flow requirements transmit.The present invention defines relative approach degree N
jFor weigh available network and theoretical optimal network press close to degree and with the variable of the poorest network of theory away from degree, the big more explanation scheme of its value is excellent more, thereby the optimal network selection problem of Business Stream is exactly the N to all available networks
jSort and choose N
jThe network that the maximum network of value transmits as Business Stream.
Approach degree is by constructing the positive ideal solution C of a problem relatively
+(
) and negative ideal solution C
-(
), calculate each available network i and theoretical optimal solution C
+The poorest C that separates
-Distance
With
Finally by formula
Calculate.Wherein, c
iA certain parameter of expression network, as: bandwidth, time delay, shake etc.The span of approach degree is [0,1] relatively.Decision-making module receives from the usable network information that satisfies this distribution of traffic flows of judge module with from the decision objective of object module, uses existing multiple attribute decision making (MADM) analytical method, realizes the distribution decision-making of Business Stream.When estimating a plurality of available networks good and bad, must determine to influence the weight of every network parameter of distribution of traffic flows decision-making, Chang Yong method was by user's specified weight in the past, for reducing burden for users, the present invention has introduced multi-parameter weight self-generating algorithm, in the present embodiment, select for use lagrange's method of multipliers to find the solution the weight of each parameter.Therefore, decision-making module uses existing multi-parameter weight self-generating algorithm at first according to usable network information, generates the weights W={ w of network parameters correspondence
1, w
2, w
3... w
n, suppose that at this each network all described by n parameter.At last, decision-making module is according to usable network information and network parameters respective weights, use TOPSIS (Technique for Order Performance by Similarity to Ideal Solution) multiple attributive decision making method, by calculating the relative approach degree N of each available network and desirable optimal network
j, N
jThe big more network of selection that shows of value approach desirable optimal network more, thereby find the optimal network that satisfies this distribution of traffic flows, and this distribution of traffic flows is transmitted in selected optimal network.The TOPSIS method is classical multiple attributive decision making method, it is widely used in numerous areas such as management, industrial production, its basic thought is the positive ideal solution and the negative ideal solution of structure evaluation problem, and with close positive ideal solution with away from the degree of bearing ideal solution, come each scheme is sorted to the relative approach degree of ideal scheme by calculating each alternative, thereby select optimal case.Article " Inter-companycomparison using modified TOPSIS with objective weights " is asked for an interview in the detailed introduction of this method.(H.Deng,C.H.Yeh?and?R.J?Willis,“Computers?&?Operations?Research”Volume?27,Number10,September?2000,pp.963-973)
Fig. 3 illustrates the multi-module cooperation communication distribution of traffic flows method based on priority and weight self-generating according to another embodiment of the invention, comprising:
S1: for better differentiated service, guarantee that important service has the priority of using network, introduced the notion of priority based on the multi-module cooperation communication distribution of traffic flows algorithm of priority and weight self-generating, each Business Stream is endowed a priority, in general, the definite of service priority can abide by following rules: the priority of real time business is higher than non-real-time service, the priority of interactive service is higher than the non-interactive type business, and the priority of service switchover is higher than the business of new access.On each terminal, algorithm of the present invention is being safeguarded a priority query, in the formation each is represented Business Stream to be distributed such as, and formation is according to the descending of Business Stream priority, and the Business Stream that comes the formation head has the priority of using network than the Business Stream of afterbody.
S2:, measure the current link circuit condition of each network according to existing network measure method.Want two time-varying parameters of Measurement Network on the one hand: network remaining bandwidth and current demand signal intensity; We adopt the definition of International Telecommunications Union to QoS on the other hand, weigh with time-delay, shake and three indexs of packet loss, therefore want the current time-delay of Measurement Network, shake and packet loss.
S3: conventional algorithm of the present invention at first takes out each bar Business Stream successively from the formation head of each terminal and distributes decision-making.
S4: the distribution of traffic flows decision process is divided into two stages, at first according to the qualifications of current business flow at aspects such as bandwidth, time delay, shake and packet loss, set is screened to current network, the available network that obtains must satisfy the bandwidth demand of Business Stream, and features such as time delay, shake and packet loss drop within the limit that Business Stream can tolerate.The feasible solution space of available network being formed this distribution of traffic flows.
S5: judge whether the feasible solution space is empty.If the feasible solution space is empty, enter the S9 adjustment algorithm; Otherwise, enter the second stage that S6 distributes decision process.
S6: each network is all described by a plurality of parameters, can compare mutually in order to make between the dissimilar parameters, need to carry out normalization according to the characteristics of parameter earlier it is carried out nondimensional standardization processing, thereby obtain normalized feasible solution space.In general, according to the value characteristics of different parameters it is divided into cost shape parameter, benefit shape parameter, threshold type parameter and fuzzy shape parameter, dissimilar parameter correspondences different processing rules and normalization formula.
● the cost shape parameter
The value of this class parameter is the smaller the better, typically as communication cost, transmission power consumption etc., uses following formula to carry out normalization for this class parameter:
k=1...m
● the benefit shape parameter
The value of this class parameter is the bigger the better, and is typical in network coverage, uses following formula to carry out normalization for this class parameter:
k=1...m
● the threshold type parameter
Professional value to this class parameter has mandatory, promptly must be more than or equal to (perhaps smaller or equal to) some threshold value t, and typical in bandwidth demand, use following formula to carry out normalization for this class parameter:
● fuzzy shape parameter
This class parameter is inconvenient to use accurate numerical value to describe usually, but can be " fine ", " generally ", " poor " etc. with the language variable description, typical in internet security, reliability, stability etc., for this class parameter, can be by the instrument of fuzzy mathematics, adopt methods such as gravity model appoach, maximin method with its de-fuzzy, be converted into concrete numerical value.
S7: when estimating a plurality of alternative network qualities, must at first determine to influence the weight of the network parameters of distribution of traffic flows decision-making.Method commonly used at present is by the user-specified parameters weight, and this is a job that is difficult to finish to the domestic consumer that lacks background knowledge, has also increased user's burden simultaneously.The present invention introduces multi-parameter weight self-generating algorithm, can effectively address the above problem, and is the weight of normalized feasible solution space automatic generating network parameters correspondence.
S8: the feasible solution space of satisfying this distribution of traffic flows and the S7 automatic network parameters weight that generate definite according to S6, the present invention introduces TOPSIS (Technique for Order Performance bySimilarity to Ideal Solution) multiple attributive decision making method, for this Business Stream is found out optimal solution in the feasible solution space, and this distribution of traffic flows transmitted in selected optimal network, turn back to the distribution decision-making that S2 carries out next Business Stream simultaneously.The basic thought of TOPSIS is: by the positive ideal solution and the negative ideal solution of structure evaluation problem, and with close positive ideal solution with away from the degree of bearing ideal solution, calculate each alternative and come each scheme is sorted, thereby select optimal case to the relative approach degree of ideal scheme.Step S3-S8 is a conventional algorithm flow process of the present invention, realizes distribution of traffic flows decision-making generally speaking.
S9: the feasible solution space is empty, can't realize distribution of traffic flows, thereby enter into adjustment algorithm of the present invention.Adjustment algorithm still according to professional priority, is supposed the business that stops a low priority on this terminal, recomputates the available bandwidth of network, is about to the bandwidth that low priority traffice takies and joins in the middle of the network availability bandwidth.Be not the transmission that really stops low priority traffice this moment, and the bandwidth of just utilizing low priority traffice to take recomputates the available bandwidth of each network.
S10: the collection of network to available bandwidth recomputates, according to the qualifications of current business flow at aspects such as bandwidth, time delay, shake and packet loss, calculate the feasible solution space of satisfying this distribution of traffic flows once more.
Whether S11: rejudging the feasible solution space is empty.If the feasible solution space still is empty, enter S12; Otherwise, enter into S13.
S12: current can't the distribution and fail for this Business Stream finds suitable transmission network.This Business Stream is removed, turned back to the distribution decision-making that S2 carries out next Business Stream simultaneously.
S13: feasible solution space non-NULL, nondimensional standardization processing is carried out in the feasible solution space of new composition, principle is with step S6.
S14: use the weight self-generating algorithm of step S7, be the weight of normalized feasible solution space automatic generating network parameters correspondence.
S15: by the TOPSIS multiple attributive decision making method of step S8, in the feasible solution space, find out optimal solution, and distribution of traffic flows is transmitted in this optimal network.
S16: stop the transmission of previous chosen low priority traffice stream, it is put back into etc. in the middle of the Business Stream priority query to be distributed, turn back to the distribution decision-making that S2 carries out next Business Stream simultaneously.Step S9-S16 is under the situation that Business Stream can't be realized distributing for the first time, the adjustment algorithm that the present invention carries out.
The present invention makes following situation become possibility: in following heterogeneous network, when having dissimilar Business Streams to transmit on the multi-module mobile terminal, can each bar distribution of traffic flows be transmitted to different network interfaces according to factors such as Link States before the net of the feature of different kinds of business stream, heterogeneous networks, on the basis that guarantees the traffic stream service quality, obtain minimum communications cost and maximum resource utilization; In addition, when no longer available or service quality obviously reduced when certain access network, multi-module mobile terminal can in time be made a response, and loaded service stream on this network is seamlessly switched on other networks that satisfy the distribution of traffic flows demand transmit.
It should be noted that at last, above embodiment is only in order to describe technical scheme of the present invention rather than the present technique method is limited, the present invention can extend to other modification, variation, application and embodiment on using, and therefore thinks that all such modifications, variation, application, embodiment are in spirit of the present invention and teachings.