CN103906135A - P2P node selection method and system used in cellular network - Google Patents

Abstract

The invention discloses a P2P node selection method and system used in a cellular network. The method comprises the step 101) that the number n of service nodes needed by data request nodes and the downlink bandwidth of the data request nodes are determined; the step 102) that a cell with the lightest flow load is selected from all the cells and is adopted as a service cell of the data request nodes; the step 103) that a node with the strong service capacity and with an uplink bandwidth close to an average bandwidth is selected from the service cell as the service node; the step 104) that n is set to be equal to n-1, judgment is carried out, and if n is larger than one, the step 102) is executed, and if n is equal to one, the next step is executed; the step 105) that a cell with lightest flow is selected form all the cells as the service cell, a node with the strong service capacity and a proper uplink bandwidth is selected from the cell, and the node with the proper uplink bandwidth enables the sum of the uplink bandwidth of the selected node and the uplink bandwidths of all the selected nodes to be larger than the data request node downlink bandwidth, and the service capacity of the node with the proper uplink bandwidth is the strongest.

Description

A kind of P2P node selecting method and system for cellular network

Technical field

The present invention relates to computer network P2P technology and mobile cellular network, be specifically related to a kind of P2P node selecting method and system for cellular network.

Background technology

Along with the fast development of radio communication and mobile terminal technology, some P2P technology have started to be applied to the mobile terminals such as smart mobile phone, PAD and notebook.Predictably, P2P shared file system will have a wide range of applications in mobile cellular network (comprising 3G and LTE).

One of key technology of P2P system is exactly that node is selected (Peer Selection), has node selection portion from existing resource and divides suitable node to provide data transport service for resource request node.Existing node selecting method mainly contains two kinds: a kind of is random system of selection take BitTorrent as representative, and another kind is " flow localized " node selecting method take P4P as representative.In random system of selection, Tracker server has the random part composition node listing (peer list) of selecting node from multiple resources, and the resource request node node from list obtains content resource.In " flow localized " node selecting method, the support that Tracker server can be served by third party selects the node that belongs to a network domains with resource request node together to form node listing as far as possible.Random system of selection realizes simple, but has ignored the topological property of bottom-layer network, easily produces a large amount of cross-domain flow rate; " flow localized " node selecting method can reduce cross-domain flow rate, optimizes flow distribution, alleviates the pressure of the net loaded P2P flow of operator backbone.

The shortcoming of prior art: node selects that node service performance (Peer Performance) is had to very large impact.For example, in P2P shared file system, file transfer rate is often used as the index of weighing node service performance, and selects the node with higher uploading bandwidth can obtain file transfer rate faster.But, in mobile cellular network, the normally downlink bandwidth of receiving terminal but not the uplink bandwidth of transmitting terminal of the bottleneck of file transfer rate.This is because the uplink bandwidth summation of multiple service nodes (being transmitting terminal) can be greater than the downlink bandwidth of requesting node (being receiving terminal) conventionally.Such as, suppose that a requesting node has 5 neighbours (being service node) simultaneously for it transmits file, although these 5 neighbours' uplink bandwidth sum exceedes 1Mbps, but because the downlink bandwidth of receiving terminal only has 500Kbps, therefore actual file transfer rate will be limited in 500Kbps.Therefore,, for node selecting method in the P2P shared file system based on mobile cellular network, there is no need always to select to have the node of higher upstream bandwidth.But existing node selecting method is not considered the downlink bandwidth restriction of receiving terminal in mobile network.

In the P2P shared file system based on mobile cellular network, carry out node selection, except the restriction of downlink bandwidth, also have two problems to mention.First, due to the bottleneck effect of each cell base station (such as, total wireless bandwidth restriction), the thought of famous " flow localized " is also inapplicable: from single subdistrict, select too much node can make the physical link bandwidth of distributing to each node be less than the available link bandwidth of each node, thereby reduce node service performance.Moreover, also can cause the load imbalance of minizone.Secondly,, due to the complex characteristics of mobile environment, the node in mobile cellular network is carried out to service ability assessment will be than much complicated in fixed network.The service ability of node refers to the comprehensive of the indexs such as computing capability, memory capacity, transmission rate and the stability of this node.In fixed network, the service ability of a node can be assessed according to its uplink bandwidth and time delay conventionally.But under mobile environment, some other factor, such as electric weight, residence time and translational speed etc., also can affect the service ability of node.Therefore, under mobile environment, node service ability is assessed and must be considered multiple factors.

Summary of the invention

The object of the invention is to, for overcoming the problems referred to above, the invention provides a kind of P2P node selecting method and system for cellular network.

To achieve these goals, the invention provides a kind of P2P node selecting method for cellular network, the P2P request of data node that the method is restricted in cellular network according to receiving terminal downlink bandwidth is selected service providing node, and described method comprises:

Step 101) the required number n of service node and the downlink bandwidth of request of data node of specified data requesting node, wherein n is the parameter that request of data node is specified when to P2P system transmission request of data, its value is not less than 1;

Step 102) from all communities, select a flow load Qing community, as the Serving cell of request of data node;

Step 103) from Serving cell, select a service ability is strong and upstream bandwidth approaches average bandwidth node as service node;

Step 104) n=n-1 is set, and judge, if n>1 returns to step 102), if n=1 enters next step, if n=0 completes service node and selects;

Step 105) from all communities, select a flow Qing community as Serving cell, in Bing Conggai community, select the strong and node with suitable upstream bandwidth of a service ability as last service providing node, the node of described suitable upstream bandwidth is: the upstream bandwidth sum that makes this node and above-mentioned all nodes of choosing is to be greater than request of data node downlink bandwidth and the strongest node of service ability;

Wherein, described average bandwidth is the average of downlink bandwidth and the required requesting node total number of request of data node; The parameter of described measurement node service ability comprises: upstream bandwidth, time delay and packet loss.

Said method also comprises: for the step that records each cell flow load and each cell flow load is upgraded and sorted, be specially:

Step 201) flow load for the ratio of the total bandwidth resources of the wireless bandwidth resource that used on base station, place, recording cell and base station as community;

Step 202) after a node is selected, upgrade the flow load of this community, node place, specific formula for calculation is:

$L_{\mathrm{est}}=\min \{L_{\mathrm{cur}}+\frac{B_p^a}{B_{\mathrm{cell}}},1\}$

Step 203) flow load of all communities is sorted;

Wherein, B _p ^afor the available upstream bandwidth of maximum of all nodes in community, B _cellrepresent the wireless bandwidth of base station, place, community, L _currepresent the cell flow load before upgrading.

Above-mentioned steps 103) further comprise:

Step 103-1) the bandwidth difference DELTA b value of each node in calculation services community, described bandwidth difference DELTA b is the difference between actual uplink bandwidth and the B ' of each node in Serving cell;

Wherein, B ' represents that last round of node selection is later to B _refthe correction of value, is the selected node upstream bandwidth target of actual use, B _refbe a fixed target of selected node upstream bandwidth, computational methods are that the downlink bandwidth of request of data node is divided by n; Step 103-2) in Serving cell | minimum K the node of Δ b| value, select a strongest node of service ability, as service node;

Described K value is:

Wherein, S is the total number of community interior nodes.

The service ability of above-mentioned node adopts Fuzzy Cognitive Map to assess, and described Fuzzy Cognitive Map is specifically by these 7 factors of the upstream bandwidth of node, time delay, packet loss, SINR, electric weight, residence time and translational speed assessing node service ability.

Above-mentioned steps 105) further comprise:

Step 105-1) from all communities, select a flow Qing community as Serving cell;

Step 105-2) the Δ b value of each node in calculation services community, concrete grammar is:

The L of the each node in calculation plot _estand B _p ^evalue, then calculates B '=B to the each node in community _ref ^-Δ b, then calculates Δ b=B _p ^e-B '; Enter step 105-3);

Wherein, B ' represents that last round of node selection is later to B _refthe correction of value is the selected node upstream bandwidth target of actual use; B _refbe a fixed target of selected node upstream bandwidth, computational methods are that the downlink bandwidth of request of data node is divided by n; Δ b is the difference between actual uplink bandwidth and the B ' of node in Serving cell, and initial value is 0;

Step 105-3) in community, K node of Δ b>0 and Δ b minimum, select a service node that the node that service ability is the strongest is chosen as request of data node;

Wherein, the computing formula of described K value is:

Wherein, S is the total number of community interior nodes.

Specifically comprise following steps for the once complete P2P node selecting method for cellular network:

Step 401) select the step of Yi Ge community, this step is specially: sorted according to flow load value in all communities; The community of selecting a flow load minimum, enters step 402);

Step 402) the Δ b value of each node in calculation plot, be specially: the Lest of the each node in calculation plot and Bpe value; Each node in community is calculated to B '=Bref-Δ b, then calculate Δ b=Bpe-B '; Check the value of n, if n>1 enters step 403), if n=1 enters step 404);

Step 403) according to service ability, K and | Δ b| selects a node from selected community, is specially: in community | K node of Δ b| minimum, select a strongest node of service ability; Record Bpe and the Lest value of selected node, record the value of Δ b, make n=n-1, enter step 401);

Step 404) from selected community, select a node according to service ability, K and Δ b, be specially: in community, K node of Δ b>0 and Δ b minimum, select a strongest node of service ability; Record the B of selected node _p ^eand L _estbe worth, record the value of Δ b, end node selection course;

When step 404) in the interstitial content k of Δ b>0 while being less than K, will use k replacement K; Step 401 in the time that the value of k is 0) in use traffic load second or San Di community are replaced to the community that flow load is minimum, then continue execution step 402);

Wherein,

L _estfor after a node is selected, the flow load value obtaining is recalculated in this community, node place, and its computing formula is:

$L_{\mathrm{est}}=\min \{L_{\mathrm{cur}}+\frac{B_p^a}{B_{\mathrm{cell}}},1\}$

L _estthe initial flow load of initial value Wei Gai community while not having node selected;

B _p ^afor the available upstream bandwidth of maximum of both candidate nodes;

B _p ^ebe the actual uplink bandwidth of a node after selected, its computing formula is:

$B_p^e=\left\{\begin{array}{cc}{B}_p^a,& L_{\mathrm{est}}<1\\ (1-L_{\mathrm{cur}})*B_{\mathrm{cell}},& L_{\mathrm{est}}=1\end{array}\right.$

N is the number of the selected node of requesting node requirement;

B _refupstream bandwidth average reference value while selection for node, concrete, B _ref=B _r/ n;

L _currepresent the cell flow load before recalculating;

B _cellrepresent the wireless bandwidth of this cell base station.

The invention provides a kind of P2P node selective system for cellular network based on said method, the P2P request of data node that this system is restricted in cellular network according to receiving terminal downlink bandwidth is selected service providing node, and described system comprises:

Control module, for arranging the initial service number of nodes n of request of data node

Service node number upgrades driver module, the initial service number of nodes n setting for real-time update control module, and upgrading principle is: in the time having selected a service node for request of data node, n value is updated to " n-1 ", in the time of n=1, drive Section Point to select module, in the time of n>1, drive first node to select module;

Cell load information upgrades statistical module, upgrades the flow load information of each community for real-time statistics;

Node service ability acquisition module, for obtaining the service ability of each node of community;

Module is selected in Serving cell, and the result selection flow load Qing community of upgrading statistical module counts according to cell load information in the time carrying out certain node is as Serving cell;

First node is selected module, for Serving cell being selected a service ability is strong and upstream bandwidth approaches average bandwidth node as service node; With

Section Point is selected module, be used for the node of a service ability being selected to by force in Serving cell and having suitable upstream bandwidth, the node of described suitable upstream bandwidth is: the upstream bandwidth sum that makes this node and above-mentioned all nodes of choosing is to be greater than request of data node downlink bandwidth and the strongest node of service ability.

Above-mentioned first node selects module further to comprise:

Node service ability statistical module, for adding up the service ability of node;

Bandwidth difference DELTA b acquisition module, for the difference between the actual uplink bandwidth of the each node in Serving cell and B ', as bandwidth difference, described B ' represents that last round of node selection is later to B _refthe correction of value, is the selected node upstream bandwidth target of actual use, described B _refbe a fixed target of selected node upstream bandwidth, computational methods are that the downlink bandwidth of request of data node is divided by n;

The first both candidate nodes collection chooser module, for will | Δ b| value is carried out size and is sorted, and selection | minimum K the node of Δ b| value is as both candidate nodes collection; The computing formula of described K is:

Wherein, S is the total number of community interior nodes;

The first chooser module, for concentrating and select a node that service ability is the strongest as service node from both candidate nodes.

Above-mentioned Section Point selects module further to comprise:

Process submodule, for the Δ b value of the each node in calculation services community, concrete grammar is:

The L of the each node in calculation plot _estand B _p ^evalue, then calculates B '=B to the each node in community _ref-Δ b, then calculates Δ b=B _p ^e-B ';

Sequence submodule, for sorting the Δ b value of acquisition;

The second both candidate nodes collection chooser module, for selecting K node of Δ b>0 and Δ b minimum as both candidate nodes collection according to the result of sequence submodule;

Service node chooser module, for concentrating and select a node that service ability is the strongest as service node from both candidate nodes;

Compensation judgment is processed submodule, for in the time meeting the interstitial content of Δ b>0 and be less than K, the actual node number that use meets Δ b>0 replaces K, if or the interstitial content that meets Δ b>0 is, will carry out node selection from flow load second or San Di community at 0 o'clock.

Said system also comprises: for the module that records each cell flow load and each cell flow load is upgraded and sorted, this module further comprises:

Cell flow load statistics submodule, the flow load for the ratio of the total bandwidth resources of the wireless bandwidth resource that used on base station, place, recording cell and base station as community;

Upgrade submodule, for after a node is selected, upgrade the flow load of this community, node place, specific formula for calculation is:

$L_{\mathrm{est}}=\min \{L_{\mathrm{cur}}+\frac{B_p^a}{B_{\mathrm{cell}}},1\}$

Sequence submodule, sorts to the flow load of all communities;

Wherein, B _p ^afor the available upstream bandwidth of maximum of all nodes in community, B _cellrepresent the wireless bandwidth of base station, place, community, L _currepresent the cell flow load before upgrading.

Compared with prior art, the present invention is that P2P shared file system based on mobile cellular network has proposed a kind of node selecting method of considering the restriction of receiving terminal downlink bandwidth.Inventor thinks, since file transfer rate can be limited to receiving terminal downlink bandwidth, the load equilibrium of minizone just should become the performance index of node selecting method so.In P2P shared file system, requesting node can be asked one and ascertain the number the node listing of (being represented as Numwant in the Tracker of BitTorrent agreement) conventionally.The technical problem that the application's patent will solve comprises: the load balancing that 1) realizes minizone under the prerequisite that meets the requirement of selected node number; 2) according to the multiple complicated factors under mobile environment, node service ability is carried out to comprehensive assessment.For this reason, the present invention proposes the node selecting method of a kind of by name DBaT (Downlink Bandwidth as Target), the method can obtain excellent cell load equalization performance obtaining good file transfer rate simultaneously.A kind of Fuzzy Cognitive Map that can be used for the node service ability in mobile cellular network to carry out comprehensive assessment has also been proposed in specific embodiment of the present invention.

Accompanying drawing explanation

Fig. 1-a is the flow chart of a kind of P2P node selecting method of considering the restriction of receiving terminal downlink bandwidth provided by the invention;

Fig. 1-b is the function curve diagram of formula provided by the invention (3);

Fig. 2 is the Fuzzy Cognitive Map based on mobile cellular networking P2P system of the present invention;

Fig. 3 is the emulation topology schematic diagram that the embodiment of the present invention 2 provides;

Fig. 4 is the simulation result figure under the 3G network scene that provides of the embodiment of the present invention 2;

Fig. 5 is the simulation result figure under the LTE network scenarios that provides of the embodiment of the present invention 2;

Embodiment

Below in conjunction with accompanying drawing, content of the present invention is described in detail.

As shown in Fig. 1-a, a kind of P2P node selecting method flow process of considering the restriction of receiving terminal downlink bandwidth that the present invention proposes mainly comprises:

1) each before selecting last node is taken turns in node selection, all first from all communities, select a flow load Qing community at every turn, then in Cong Gai community, select the strong and upstream bandwidth of a service ability to approach the node of selected node average bandwidth;

2) selecting last of last node to take turns during node selects, first from all communities, select a flow load Qing community, then in Cong Gai community, select a service ability strong and there is the node of suitable upstream bandwidth, make the upstream bandwidth sum of all selected nodes just in time be greater than the restriction of receiving terminal downlink bandwidth.

Selected node average bandwidth B in above-mentioned flow process _refby receiving terminal downlink bandwidth B _rdetermine i.e. B with requesting node number n _ref=B _r/ n.As previously mentioned, in mobile network, due to the existence of downlink bandwidth restriction, there is no need the node of always selecting upstream bandwidth larger.Therefore, the target of DBaT method using receiving terminal downlink bandwidth as multiple selected node uplink bandwidth sums, and adopt the preferential thought in community to strive for the load balancing of minizone.Moreover, under traditional fixed network environment, file transfer rate can be accelerated along with the increase of service node number conventionally.But under mobile cellular network environment, owing to being limited by the downlink bandwidth of receiving terminal, file transfer rate might not be accelerated along with the increase of service node number.Therefore, the design object of DBaT method is exactly to make the upstream bandwidth sum of selected node just in time exceed the restriction of receiving terminal downlink bandwidth.

One of standard that DBaT method is selected the service ability of node as node.As previously mentioned, owing to there are multiple factor impacts, under mobile environment, carry out node service ability assessment more complicated.Here only describe the technical scheme of DBaT method, in specific embodiment, will introduce a kind of method of assessing node service ability in mobile cellular network.

The symbol that need to use while being listed in description technical solution of the present invention below and their meaning:

● B _r: the downlink bandwidth upper limit of requesting node.

● L _est: after a node is selected, the flow load value obtaining, L are recalculated in this community, node place _estthe initial flow load of initial value Wei Gai community while not having node selected.

● B _p ^a: the available upstream bandwidth of maximum of node.

● B _p ^e: the actual uplink bandwidth after node is selected.

● n: the number of the selected node (service node) that requesting node requires.

● B _ref: upstream bandwidth average reference value when node is selected, concrete, B _ref=B _r/ n.

● B ': upstream bandwidth accumulation reference value when node is selected.

● Δ b: be the actual uplink bandwidth of each node in Serving cell and the difference of B ', initial value is 0.

● S: the total number of community interior nodes.

● K: the concentrated interstitial content of both candidate nodes in community.

Definition for " cell flow load " is: the ratio of the total bandwidth resources in the wireless bandwidth resource having used on base station and base station.Such as, suppose that the current bandwidth resources that used of NodeB are 1.2Mbps, total because the bandwidth resources of this base station are 2Mbps, the current flow load in this community is exactly that 0.6. is obvious so, the span of cell flow load is from 0 to 1 closed interval.

For Yi Ge community, in node selecting method Cong Gai community, select a node to carry out transfer of data, the flow load of this community will increase because selected node produces new flow so, until flow load value reaches the upper limit 1.So it is as follows to recalculate the formula of cell flow load:

$L_{\mathrm{est}}=\min \{L_{\mathrm{cur}}+\frac{B_p^a}{B_{\mathrm{cell}}},1\}---\left(1\right)$

Wherein L _currepresent the cell flow load before recalculating, B _cellrepresent the wireless bandwidth of this cell base station.B like this _p ^ejust can calculate by formula below,

$B_p^e=\left\{\begin{array}{cc}{B}_p^a,& L_{\mathrm{est}}<1\\ (1-L_{\mathrm{cur}})*B_{\mathrm{cell}},& L_{\mathrm{est}}=1\end{array}\right.---\left(2\right)$

Wherein L _curand B _cellidentical with meaning in formula (1).

Here it is worth mentioning that, is not linear relationship between the link bandwidth of node on the flow load Yu Gai community on actual environment Zhong Yige community.But, between them, there is linear relationship for need to supposing of simplifying in above-mentioned formula.Be understood that, such hypothesis can't affect the effect of DBaT method.

Take turns in node selection B at each _refcan regard a fixed target of selected node upstream bandwidth as, because the last round of node of value representation of B ' is selected later B _refthe correction of value, therefore B ' can regard the selected node upstream bandwidth target of actual use as, Δ b is used for recording the difference between actual uplink bandwidth and the B ' of both candidate nodes.Therefore, the basic thought of DBaT method may be summarized to be: in the end each of taking turns is before taken turns in node selection, selects a upstream bandwidth to approach the node of B '; In the end one take turns in node selection, select a node according to the value of Δ b, make the actual uplink bandwidth sum of selected node higher than B _r.

In order to guarantee that selected node has good service ability, take turns node at each and all use a both candidate nodes collection to draw a circle to approve K node to select in selecting.That is, take turns during node selects in each of DBaT method, the quantity of real candidate node is K but not S.The value of K is discussed below.First, the value of K should be the function of S; Secondly,, along with the increase of S, the value of K should remain the sub-fraction that accounts for S.Therefore (3) formula that, the present invention proposes is determined the value of K.In specific embodiment, (3) formula demonstrates good effect.(3) function curve of formula, as shown in Fig. 1-b, therefrom can be found out the value variation tendency of K.

The concrete execution flow process of DBaT method is as follows:

Step 1. is selected Yi Ge community:

1-1 sorts all communities according to flow load value;

1-2 selects the community of a flow load minimum, enters step 2.

The Δ b value of each node in step 2. calculation plot:

The L of the each node in 2-1 calculation plot _estand B _p ^evalue;

Each node in 2-2Dui community calculates B '=B _ref-Δ b, then calculates Δ b=B _p ^e-B ';

2-3 checks the value of n, if n>1 enters step 3, if n=1 enters step 4.

Step 3. according to service ability, K and | Δ b| selects a node from selected community:

In 3-1Cong community | in K node of Δ b| minimum, select a strongest node of service ability;

3-2 records the B of selected node _p ^eand L _estbe worth, record the value of Δ b, make n=n – 1, enter step 1.

Step 4. is selected a node according to service ability, K and Δ b from selected community:

In 4-1Cong community, in K node of Δ b>0 and Δ b minimum, select a strongest node of service ability;

4-2 records Bpe and the L of selected node _estbe worth, record the value of Δ b, end node selection course.

In step 4, likely there will be meet Δ b>0 interstitial content (be designated as the situation that k) is less than K, in this case method will use k replace K.The situation that the value that even, also may occur k is 0.In particular cases this, method replaces by use traffic load second (or the 3rd, etc.) Di community the community that flow load is minimum in step 1, then continues execution step 2.

Specific embodiment one

In specific embodiment one, the present invention proposes the Fuzzy Cognitive Map of node service ability assessment in a kind of DBaT of can be used for method.Fuzzy Cognitive Map can give expression to the direct relation that affects between multiple concepts intuitively, also can obtain the remote-effects relation between these concepts by some matrix computations.In prior art, Fuzzy Cognitive Map theory is used to carry out node failure risk assessment in the P2P system based on wireless Ad hoc network.This thought is equally applicable to the assessment of node service ability in the P2P shared file system based on mobile cellular network.The present invention considers the factor of following 7 major effect mobile cellular network environment lower node service ability:

● node upstream bandwidth: the upstream bandwidth of node has directly and larger impact its service ability.

● time delay: the time delay between service node and resource request node has a certain impact to its service ability.

● packet loss: emulation experiment confirmation, the link packet drop rate between service node and resource request node has larger impact to node service ability.

● SINR (the Signal to Interference and Noise Ratio of node wireless link, Signal to Interference plus Noise Ratio): this factor is the direct indicator of radio link quality, and radio link quality will directly affect the packet loss of wireless channel, also can affect thus the service ability of node.

● node electric weight: because the comparision of quantity of electricity of mobile terminal is limited, therefore this factor can directly have influence on " node failure risk ", and " node failure risk " has considerable influence to its service ability.Moreover, residence time conventionally also can be more longer in system for the node that electric weight is higher.

● node residence time: consistent with existing research, the present invention supposes that the node its " failure risk " of growing residence time is also lower.

● node motion speed: under mobile environment, translational speed means lower radio link quality conventionally faster, i.e. higher packet loss.Moreover, translational speed faster node also has higher " community switching probability ", and " community switching probability " will affect the service ability of node to a certain extent.

According to above analysis, the present invention provides a kind of Fuzzy Cognitive Map that can be used for the P2P system based on mobile cellular networking, as shown in Figure 2.

Can show that according to Fig. 2 adjacency matrix E is as shown in (4) formula, influence degree in this matrix " A little ", " Some " and " Very Much " set 1,2,3} quantizes:

According to the matrix operation rule of (4) formula and Fuzzy Cognitive Map, can obtain following formula:

(5) matrix T in formula has provided each influencing factor to object concept C ₀the quantification summation (comprise directly and indirectly) of influence degree.Such as, " upstream bandwidth " is 3 to " service ability " total quantization influence, and " time delay " is-2 to " service ability " total quantization influence.It should be noted that a total C in Fig. 2 ₁-C ₉9 affect C ₀object concept.But, in real network environment, object concept " node failure probability " (C ₇) and " community switching probability " (C ₉) can not directly be measured; Meanwhile, they are also partly subject to the impact of the factor such as " electric weight ", " residence time " and " translational speed ".Therefore, object concept " node failure probability " and " community switching probability " will can not be used as assessing the principal element of node service ability.

Hence one can see that, and these 7 factors of upstream bandwidth, time delay, packet loss, SINR, electric weight, residence time and translational speed to the influence degree quantized value of node service ability are:

$W=\left(\begin{array}{c}\end{array}\begin{array}{ccccccc}{C}_1& C_2& C_3& C_4& C_5& C_6& {\mathrm{<figure-callout\; id="8"\; label="C"\; filenames="HDA00002676988200023.png,HDA00002676988200031.png"\; state="\{\{state\}\}">C</figure-callout>}}_8\\ 3& -2& 6& 18& 6& -9& -3\end{array}\right)---\left(6\right)$

Formula above provides weight rule when above-mentioned 7 combined factors are considered.The occurrence of the present invention's hypothesis these 7 factors in P2P shared file system can measured and collection.After measured value is done to suitable normalization, then take advantage of power to be added according to (6) formula to the measured value of seven factors, get final product to obtain the comprehensive assessment value of egress service ability.

It is worth mentioning that, the Fuzzy Cognitive Map that the present invention proposes can be revised according to different environment and condition.In addition, except using Fuzzy Cognitive Map, the assessment of node service ability also can realize by other method.In a word, the design of DBaT method has good extensibility---and in DBaT method, can apply diverse ways node service ability is assessed.

Specific embodiment two

The present embodiment is verified the performance of DBaT method by the emulation platform based on OMNet++.Be used for the network topological diagram of emulation as shown in Figure 3, total 20Ge community in this topology.In emulation, we have adopted the P2P system of a similar BitTorrent framework.

In simulation process, in order to simulate one relatively really and to compare the initial condition of normalization, the flow load initial value on each community will random generation in [0.25,0.75] scope.The number of requesting node is 10, and the number of the both candidate nodes of each requesting node is 100, therefore always has 1000 both candidate nodes and is randomly dispersed in 20Ge community.

The partial parameters of both candidate nodes arranges by the following method.Time delay between each both candidate nodes and requesting node is [0,500ms] interior random generation of scope, wireless link SINR value is random generation in [0,100dB] scope, electric weight grade is [0,5] random generation in scope, the value of residence time determines jointly by a random number and electric weight grade, its span is [0,5h], the packet loss of wireless link determines jointly by SINR value, node motion speed and a random number, and its span is [0.001,0.01].

In the process of transfer of data, each service node has a node failure (peer churn) probability, and this probable value is by jointly determining electric weight grade and the residence time of this node.After node failure occurs, the node of inefficacy will leave system, and resource request node obtains remaining data by requiring tracker to return to a new service node address.In addition, each service node You Yige community switching probability (moving to the probability of adjacent cell), this probable value is determined jointly by node motion speed and a random number.When a service node moves to behind neighbor cell, the parameter values such as its time delay, SINR and packet loss will regenerate.

The present embodiment is verified the method performance of DBaT by a file-sharing process that is similar to BitTorrent.First, 10 requesting nodes add system and to the request of Tracker Transmit message simultaneously; Next, Tracker carries out identical node selecting method to 10 requesting nodes, respectively return service node listing; Finally, 10 requesting nodes obtain a complete file content from the service node of oneself respectively.In file-sharing process, record the standard deviation (SD, standard deviation) of the flow load on (average file transfer time) the He20Ge community of average file transfer time of 10 requesting nodes.

The present embodiment compares the performance of DBaT method and other two kinds of methods.A HSA(Highest Service Ability by name), another RS(Random Selection by name).HSA always selects the service ability method for optimizing of strong node, and RS is exactly traditional random system of selection.The present embodiment has been distinguished emulation 3G network and two kinds of scenes of LTE network.Emulation under every kind of scene arranges as shown in table 1.In addition, in emulation each time, the restriction of the downlink bandwidth of 10 requesting nodes is all set to identical.

Table 1 simulating scenes parameter arranges

Fig. 4 has provided the simulation result under 3G network scene, and the service node number that wherein each requesting node is asked changes to 8 from 4.Can do following observation and analysis for the simulation result in Fig. 4.First, DBaT method obtains on can 20Ge community than well a lot of load-balancing performance of other two kinds of methods.Secondly, the average file transfer speed that RS method obtains is the slowest, and unstable properties.The reason that causes this phenomenon is that the stochastic behaviour of RS method can be selected the not good service node of some abilities.Finally, different from Fig. 4, along with the variation of required interstitial content, the obtained average file transfer time of DBaT method always approaches (even more excellent) very much with HSA method.This phenomenon can be made description below.Because the downlink bandwidth of each requesting node is relatively high, the flow load on each community is also relatively heavier.Identical with the DBaT method in a upper trifle, in the time that the flow load on community is heavier, DBaT method can still less be subject to than other two kinds of methods the impact of base station bottleneck effect.

Fig. 5 has provided the simulation result under LTE network scenarios, and the service node number that wherein each requesting node is asked changes to 8 from 4.Result and Fig. 4 in Fig. 5 have similarity: first, DBaT method has obtained optimum load-balancing performance; Secondly, the file transfer speed that RS method obtains is the slowest, and unstable properties; Finally, no matter required interstitial content is how many, and the obtained file transfer speed of DBaT method always approaches (even more excellent) very much with HSA method.Moreover, can find out by Fig. 4 and Fig. 5, no matter how required interstitial content increases, and the average file transfer time does not almost change.This has clearly proved that the restriction of receiving terminal downlink bandwidth is really to exist on the impact of file transfer speed again.

In a word, DBaT method provided by the invention can obtain than random system of selection and " flow localized " node selecting method and obviously more excellent load-balancing performance of method for optimizing; In addition, DBaT method of the present invention can obtain the file transfer speed suitable with method for optimizing, especially in the time that the flow load on community is heavier.

It should be noted that, embodiment of the present invention of more than introducing provide a kind of a kind of method that in cellular network, P2P node selecting method of considering receiving terminal downlink bandwidth restriction can move, and the explanation of this embodiment is just for helping to understand method of the present invention and core concept thereof and unrestricted.One of ordinary skill in the art should be appreciated that any modification to technical solution of the present invention or is equal to and substitute the spirit and scope that do not depart from technical solution of the present invention, and it all should be encompassed within the scope of claim of the present invention.

Claims (10)

1. for a P2P node selecting method for cellular network, the P2P request of data node that the method is restricted in cellular network according to receiving terminal downlink bandwidth is selected service providing node, and described method comprises:

Step 101) the required number n of service node and the downlink bandwidth of request of data node of specified data requesting node, wherein n is the parameter that request of data node is specified when to P2P system transmission request of data, its value is not less than 1;

Step 102) from all communities, select a flow load Qing community, as the Serving cell of request of data node;

Step 103) from Serving cell, select a service ability is strong and upstream bandwidth approaches average bandwidth node as service node;

Step 104) n=n-1 is set, and judge, if n>1 returns to step 102), if n=1 enters next step, if n=0 completes service node and selects;

Step 105) from all communities, select a flow Qing community as Serving cell, in Bing Conggai community, select the strong and node with suitable upstream bandwidth of a service ability as last service providing node, the node of described suitable upstream bandwidth is: the upstream bandwidth sum that makes this node and above-mentioned all nodes of choosing is to be greater than request of data node downlink bandwidth and the strongest node of service ability;

Wherein, described average bandwidth is the average of downlink bandwidth and the required requesting node total number of request of data node; The parameter of described measurement node service ability comprises: upstream bandwidth, time delay and packet loss.

2. the P2P node selecting method for cellular network according to claim 1, is characterized in that, described method also comprises: for the step that records each cell flow load and each cell flow load is upgraded and sorted, be specially:

Step 201) flow load for the ratio of the total bandwidth resources of the wireless bandwidth resource that used on base station, place, recording cell and base station as community;

Step 202) after a node is selected, upgrade the flow load of this community, node place, specific formula for calculation is:

$L_{\mathrm{est}}=\min \{L_{\mathrm{cur}}+\frac{B_p^a}{B_{\mathrm{cell}}},1\}$

Step 203) flow load of all communities is sorted;

Wherein, B _p ^afor the available upstream bandwidth of maximum of all nodes in community, B _cellrepresent the wireless bandwidth of base station, place, community, L _currepresent the cell flow load before upgrading.

3. the P2P node selecting method for cellular network according to claim 1, is characterized in that described step 103) further comprise:

Step 103-1) the bandwidth difference DELTA b value of each node in calculation services community, described bandwidth difference DELTA b is the difference between actual uplink bandwidth and the B ' of each node in Serving cell;

Wherein, B ' represents that last round of node selection is later to B _refthe correction of value, is the selected node upstream bandwidth target of actual use, B _refbe a fixed target of selected node upstream bandwidth, computational methods are that the downlink bandwidth of request of data node is divided by n; Step 103-2) in Serving cell | minimum K the node of Δ b| value, select a strongest node of service ability, as service node;

Described K value is:

Wherein, S is the total number of community interior nodes.

4. the P2P node selecting method for cellular network according to claim 1, it is characterized in that, the service ability of described node adopts Fuzzy Cognitive Map to assess, and described Fuzzy Cognitive Map is specifically by these 7 factors of the upstream bandwidth of node, time delay, packet loss, SINR, electric weight, residence time and translational speed assessing node service ability.

5. the P2P node selecting method for cellular network according to claim 1, is characterized in that described step 105) further comprise:

Step 105-1) from all communities, select a flow Qing community as Serving cell;

Step 105-2) the Δ b value of each node in calculation services community, concrete grammar is:

The L of the each node in calculation plot _estand B _p ^evalue, then calculates B '=B to the each node in community _ref ^-Δ b, then calculates Δ b=B _p ^e-B '; Enter step 105-3);

Wherein, B ' represents that last round of node selection is later to B _refthe correction of value is the selected node upstream bandwidth target of actual use; B _refbe a fixed target of selected node upstream bandwidth, computational methods are that the downlink bandwidth of request of data node is divided by n; Δ b is the difference between actual uplink bandwidth and the B ' of node in Serving cell, and initial value is 0;

Step 105-3) in community, K node of Δ b>0 and Δ b minimum, select a service node that the node that service ability is the strongest is chosen as request of data node;

Wherein, the computing formula of described K value is:

Wherein, S is the total number of community interior nodes.

6. the P2P node selecting method for cellular network according to claim 5, is characterized in that, specifically comprises following steps for the once complete P2P node selecting method for cellular network:

Step 401) select the step of Yi Ge community, this step is specially: sorted according to flow load value in all communities; The community of selecting a flow load minimum, enters step 402);

Step 402) the Δ b value of each node in calculation plot, be specially: the Lest of the each node in calculation plot and Bpe value; Each node in community is calculated to B '=Bref-Δ b, then calculate Δ b=Bpe-B '; Check the value of n, if n>1 enters step 403), if n=1 enters step 404);

Step 403) according to service ability, K and | Δ b| selects a node from selected community, is specially: in community | K node of Δ b| minimum, select a strongest node of service ability; Record Bpe and the Lest value of selected node, record the value of Δ b, make n=n – 1, enter step 401);

Step 404) from selected community, select a node according to service ability, K and Δ b, be specially: in community, K node of Δ b>0 and Δ b minimum, select a strongest node of service ability; Record the B of selected node _p ^eand L _estbe worth, record the value of Δ b, end node selection course;

When step 404) in the interstitial content k of Δ b>0 while being less than K, will use k replacement K; Step 401 in the time that the value of k is 0) in use traffic load second or San Di community are replaced to the community that flow load is minimum, then continue execution step 402);

Wherein,

L _estfor after a node is selected, the flow load value obtaining is recalculated in this community, node place, and its computing formula is:

$L_{\mathrm{est}}=\min \{L_{\mathrm{cur}}+\frac{B_p^a}{B_{\mathrm{cell}}},1\}$

L _estthe initial flow load of initial value Wei Gai community while not having node selected;

B _p ^afor the available upstream bandwidth of maximum of both candidate nodes;

B _p ^ebe the actual uplink bandwidth of a node after selected, its computing formula is:

$B_p^e=\left\{\begin{array}{cc}{B}_p^a,& L_{\mathrm{est}}<1\\ (1-L_{\mathrm{cur}})*B_{\mathrm{cell}},& L_{\mathrm{est}}=1\end{array}\right.$

N is the number of the selected node of requesting node requirement;

B _refupstream bandwidth average reference value while selection for node, concrete, B _ref=B _r/ n;

L _currepresent the cell flow load before recalculating;

B _cellrepresent the wireless bandwidth of this cell base station.

7. for a P2P node selective system for cellular network, the P2P request of data node that this system is restricted in cellular network according to receiving terminal downlink bandwidth is selected service providing node, and described system comprises:

Control module, for arranging the initial service number of nodes n of request of data node;

Service node number upgrades driver module, the initial service number of nodes n setting for real-time update control module, and upgrading principle is: in the time having selected a service node for request of data node, n value is updated to " n-1 ", in the time of n=1, drive Section Point to select module, in the time of n>1, drive first node to select module;

Cell load information upgrades statistical module, upgrades the flow load information of each community for real-time statistics;

Node service ability acquisition module, for obtaining the service ability of each node of community;

Module is selected in Serving cell, and the result selection flow load Qing community of upgrading statistical module counts according to cell load information in the time carrying out certain node is as Serving cell;

First node is selected module, for Serving cell being selected a service ability is strong and upstream bandwidth approaches average bandwidth node as service node; With

Section Point is selected module, be used for the node of a service ability being selected to by force in Serving cell and having suitable upstream bandwidth, the node of described suitable upstream bandwidth is: the upstream bandwidth sum that makes this node and above-mentioned all nodes of choosing is to be greater than request of data node downlink bandwidth and the strongest node of service ability.

8. the P2P node selective system for cellular network according to claim 7, is characterized in that, described first node selects module further to comprise:

Node service ability statistical module, for adding up the service ability of node;

Bandwidth difference DELTA b acquisition module, for the difference between the actual uplink bandwidth of the each node in Serving cell and B ', as bandwidth difference, described B ' represents that last round of node selection is later to B _refthe correction of value, is the selected node upstream bandwidth target of actual use, described B _refbe a fixed target of selected node upstream bandwidth, computational methods are that the downlink bandwidth of request of data node is divided by n;

The first both candidate nodes collection chooser module, for will | Δ b| value is carried out size and is sorted, and selection | minimum K the node of Δ b| value is as both candidate nodes collection; The computing formula of described K is:

Wherein, S is the total number of community interior nodes;

The first chooser module, for concentrating and select a node that service ability is the strongest as service node from both candidate nodes.

9. the P2P node selective system for cellular network according to claim 7, is characterized in that, described Section Point selects module further to comprise:

Process submodule, for the Δ b value of the each node in calculation services community, concrete grammar is:

The L of the each node in calculation plot _estand B _p ^evalue, then calculates B '=B to the each node in community _ref ^-Δ b, then calculates Δ b=B _p ^e-B ';

Sequence submodule, for sorting the Δ b value of acquisition;

The second both candidate nodes collection chooser module, for selecting K node of Δ b>0 and Δ b minimum as both candidate nodes collection according to the result of sequence submodule;

Service node chooser module, for concentrating and select a node that service ability is the strongest as service node from both candidate nodes;

Compensation judgment is processed submodule, for in the time meeting the interstitial content of Δ b>0 and be less than K, the actual node number that use meets Δ b>0 replaces K, if or the interstitial content that meets Δ b>0 is, will carry out node selection from flow load second or San Di community at 0 o'clock.

10. the P2P node selective system for cellular network according to claim 7, it is characterized in that, described system also comprises: for the module that records each cell flow load and each cell flow load is upgraded and sorted, this module further comprises:

Cell flow load statistics submodule, the flow load for the ratio of the total bandwidth resources of the wireless bandwidth resource that used on base station, place, recording cell and base station as community;

Upgrade submodule, for after a node is selected, upgrade the flow load of this community, node place, specific formula for calculation is:

$L_{\mathrm{est}}=\min \{L_{\mathrm{cur}}+\frac{B_p^a}{B_{\mathrm{cell}}},1\}$

Sequence submodule, sorts to the flow load of all communities;

Wherein, B _p ^afor the available upstream bandwidth of maximum of all nodes in community, B _cellrepresent the wireless bandwidth of base station, place, community, L _currepresent the cell flow load before upgrading.

Images