CN114513426A - CCN community division method based on node similarity and influence - Google Patents
CCN community division method based on node similarity and influence Download PDFInfo
- Publication number
- CN114513426A CN114513426A CN202210198386.6A CN202210198386A CN114513426A CN 114513426 A CN114513426 A CN 114513426A CN 202210198386 A CN202210198386 A CN 202210198386A CN 114513426 A CN114513426 A CN 114513426A
- Authority
- CN
- China
- Prior art keywords
- node
- community
- value
- ccn
- similarity
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 47
- 239000013598 vector Substances 0.000 claims abstract description 48
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 37
- 238000004891 communication Methods 0.000 claims description 23
- 238000004364 calculation method Methods 0.000 claims description 11
- 230000005540 biological transmission Effects 0.000 claims description 9
- 239000011159 matrix material Substances 0.000 claims description 9
- 230000002860 competitive effect Effects 0.000 claims description 5
- 238000005265 energy consumption Methods 0.000 claims description 5
- 230000000694 effects Effects 0.000 claims description 3
- 230000003993 interaction Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 claims description 2
- 230000008569 process Effects 0.000 abstract description 8
- 238000004088 simulation Methods 0.000 description 4
- 235000008694 Humulus lupulus Nutrition 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000000638 solvent extraction Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000010354 integration Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000006855 networking Effects 0.000 description 1
- 238000005192 partition Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/12—Discovery or management of network topologies
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0803—Configuration setting
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/14—Network analysis or design
- H04L41/142—Network analysis or design using statistical or mathematical methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Algebra (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Analysis (AREA)
- Mathematical Optimization (AREA)
- Mathematical Physics (AREA)
- Probability & Statistics with Applications (AREA)
- Pure & Applied Mathematics (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
The invention provides a CCN community division method based on node similarity and influence, which comprises the following steps: firstly, calculating the centrality of a feature vector corresponding to each node in a community to obtain a standardized value of the centrality of the feature vector; improving a label propagation algorithm, and determining an updating rule of a label propagation value; dividing the CCN network into a plurality of non-overlapping communities based on the characteristic vector centrality after standardization and an improved label propagation algorithm; finally, an SDN controller is deployed in each divided community to help manage the communities. The invention solves the problems of low efficiency and redundancy in the content retrieval process of the existing content center network and the technical problems of poor stability, lack of consideration of the similarity between nodes and the importance of the nodes in the existing CCN community division method, can accelerate the speed of content retrieval and route distribution by introducing the SDN controller and community division, and improves the performance of the CCN route.
Description
Technical Field
The invention relates to the technical field of network community division, in particular to a CCN community division method based on node similarity and influence.
Background
The traditional network architecture based on TCP/IP can not well support the increasing network flow, and the user often does not know the content provider when requesting the content, thereby causing the problem that the destination address can not be embedded into the interest packet for forwarding. The content-centric network (CCN) is converted from a provider-driven end-to-end communication mode to an interest-driven content retrieval communication mode, which makes the CCN unable to route according to the source and destination IP addresses by using a hop-by-hop return method like TCP/IP, so the design of an efficient CCN routing scheme will be a problem that needs to be solved urgently by the CCN network architecture. In recent years, CCN routing mechanisms have been extensively studied, and in order to solve the problems of scalability and large-scale deployment of CCNs, there are two effective schemes. One is integration with Software Defined Networking (SDN), since SDN is a unique model of future internet, enabling separation of control plane from data plane; another solution is to perform community division, but when performing community division, the size and content of the community are difficult to determine, which brings new challenges to community division.
Scholars propose a plurality of community discovery algorithms, Girvan et al propose edge betweenness-based split GN algorithm, delete the edge with the maximum edge betweenness each time until all edges are removed; blondel et al propose a luvain algorithm with maximized modularity, each point is regarded as an independent community, and the folded community interval and the continuous edge weight in the community are calculated until the community is finally integrated. Raghavan et al propose a Label Propagation Algorithm (LPA) which has linear time complexity compared to the first two algorithms and is suitable for community partitioning in large networks.
The core idea of the LPA algorithm is as follows: first, a label value is distributed to each node in the way to represent the community where the node is located. Finding out the nodes with the maximum label propagation value in the neighbor nodes (when the maximum value is not unique, one node is randomly selected), adding the node into the community, updating the label propagation value of the node, and stopping iteration when the label propagation value of the node is not changed any more; otherwise, repeating the steps.
In the conventional LPA algorithm, the initial tag value is determined to be random, so that the divided communities have the defects of poor stability and low accuracy. Luo et al improve the quality of community division by optimizing an objective function, but have the problem of poor LPA stability; zhang et al changes the tag update sequence by calculating the tag importance, but does not consider the similarity between nodes; song et al changes the tag update sequence by computing node similarity, but does not consider the importance of the nodes.
Disclosure of Invention
Aiming at the technical problems of low efficiency and redundancy in a content retrieval process of the existing content center network and poor stability and lack of consideration of similarity among nodes and importance of the nodes in the existing CCN community division method, the invention provides the CCN community division method based on the node similarity and the influence.
In order to solve the technical problems, the invention adopts the following technical scheme: a CCN community division method based on node similarity and influence comprises the following steps:
the method comprises the following steps: calculating the centrality of the feature vector corresponding to each node in the community, and standardizing the centrality of the feature vector to obtain n standardized values of the centrality of the feature vector;
step two: inputting the obtained value of the centrality of the standardized feature vector into a tag propagation algorithm to serve as an initial value of a tag propagation value, and determining an updating rule of the tag propagation value based on interest similarity and social influence among nodes in a community;
step three: dividing the CCN network into a plurality of non-overlapping communities based on the characteristic vector centrality after standardization and an improved label propagation algorithm;
step four: and selecting the most competitive node in each divided community to deploy the SDN controller so as to help manage community information and topology and interaction with other communities.
The method for calculating the centrality of the feature vector corresponding to each node in the community comprises the following steps: by A ═ e (e)ij)n×nRepresenting an adjacency matrix corresponding to an undirected graph, where X is (X)1,x2,…,xn) One eigenvector representing the adjacency matrix, for an arbitrary node viThe corresponding characteristic vector value xiComprises the following steps:
in the formula: i is more than or equal to 1 and less than or equal to n, n represents the total number of nodes, xjRepresenting a node viV of a neighbor nodejThe corresponding characteristic vector value, wherein lambda is the characteristic value corresponding to the characteristic vector X of the adjacent matrix A; when the feature vector value x is matchediPerforming multiple iterations until the value reaches steady state, at which time the characteristic vector value xiRepresenting a node viThe corresponding feature vector is centralised.
Centrality x of feature vectors within interval (0,1)iNormalizing to obtain n normalized feature vector centrality values xi':
The method for calculating the interest similarity between the nodes comprises the following steps: suppose node viAnd a neighbor node vjIf the same interest field number is p, the node viAnd a neighbor node vjThe interest similarity between them is:
when node viGenerating a contained interest field fkWhen requesting (2), there are:
the calculation formula of the social influence among the nodes is as follows:
ki=|Γ(i)|,kj=|Γ(j)| (6)
in the formula :dijFor flowing through the node viAnd a neighbor node vjThe sum of the data traffic of (a), Γ (i) and Γ (j) are respectively a node viAnd a neighbor node vjSet of (a), kiAnd k isjAre respectively node viAnd a neighbor node vjDegree of (d), dis (i, j) being node viAnd a neighbor node vjThe euclidean distance between.
The updating rule for determining the label propagation value in the second step is as follows:
wij=αTwij+(1-α)Swij (8)
in the formula :NiIs a node viV of a neighbor nodejSet of (a)jFor neighbor node vjTag activity value of, kjFor neighbor node vjDegree of (d), wijIs a node viWith neighbor node vjThe similarity of interest Tw betweenijAnd social influence SwijAlpha is a corresponding constant coefficient, and reflects interest similarity Tw between nodesijAnd social influence SwijThe occupied weight.
The method for dividing the non-overlapping communities in the third step comprises the following steps: firstly, determining the number of divided communities, and calculating the centrality value x of n standardized feature vectorsiCarrying out descending order arrangement, and selecting nodes corresponding to the first k values as corresponding initial communities to obtain k initial communities; then, other nodes are developed as community members for the k initial communities in a parallel mode based on a hierarchy traversal method.
The calculation formula of the node competitiveness in each community in the fourth step is as follows:
in the formula :DfxiIs a node viData forwarding capability of CoxiFor SDN controller SxDeployed in Community CxNode v ofiThe cost of the communication over the network,to adjust the coefficients.
The data forwarding capability of the node is related to the label propagation value and the network bandwidth after the node is standardized, wherein:
the label propagation value after node normalization is:
the network bandwidth after node standardization is as follows:
in the formula :LkRepresenting a node vkA tag propagation value of bxiRepresenting a node vxAnd node viTransmission bandwidth between bxkRepresenting a node vxAnd node vkA transmission bandwidth therebetween;
the data forwarding capability of the node is:
Dfxi=β·Li′+(1-β)·Bwi′ (12)
where β is the weight occupied by the tag propagation value and the network bandwidth.
The calculation formula of the communication cost of the node is as follows:
Coxi=Nui·Csi+Tfi·Eci (13)
in the formula :NuiRepresenting a node viNumber of requests of, CsiRepresenting a node viFixed time delay consumption, TfiRepresenting flow through node viData flow of (Ec)iRepresenting the energy consumption for processing data per bit;
communication costs are normalized, and values are distributed between intervals (0,1), so that the normalized communication costs are as follows:
in the formula :CoxkRepresenting a node vxAnd node viThe cost of communication between.
According to the invention, a CCN routing mechanism is researched by introducing an SDN controller and community division, and a CCN network is divided into a plurality of communities by combining two mainstream methods of feature vector centrality and label propagation; then, one node is selected to deploy the SDN controller in the community according to the node competitiveness, which is mainly determined by the data forwarding capability of the node and the communication cost of the community, wherein each SDN controller is composed of SIT and SCT and used for recording and topology management of network information. In addition, the network topology is simulated based on the real data set, and the experimental result shows that the method has good performance.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart of the present invention;
FIG. 2 is a pseudo-code algorithm for community partitioning in accordance with the present invention;
fig. 3 is a pseudo code algorithm selected by an SDN controller deployment node according to the present invention;
FIG. 4 is a graph showing changes in the module degree values corresponding to different Lim values when the community divides the community;
FIG. 5 is a graph of average routing hops for four different algorithms, SI-LPA, S-LPA, LPAh, and HPI _ LPA, for different interest request times;
FIG. 6 is a graph of average cache hit rates for four different algorithms, SI-LPA, S-LPA, LPAh, and HPI _ LPA, for different interest request times.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without inventive effort based on the embodiments of the present invention, are within the scope of the present invention.
As shown in fig. 1, the present invention provides a CCN community division method based on node similarity and influence, and a value of feature vector centrality of a node is used as an initial value of a tag propagation value in a tag propagation algorithm; then, a method for updating a label propagation value is provided by comprehensively considering the similarity and the influence between the node and the neighbor node, and a CCN network topological structure is divided into a plurality of non-overlapping communities; and finally, selecting the most competitive node from each divided community to deploy the SDN controller so as to help better manage community functions and perform inter-community routing. Experiments show that the speed of CCN content retrieval and route distribution can be increased by introducing the SDN controller and community division, so that the performance of CCN routing is improved.
The method specifically comprises the following steps:
the method comprises the following steps: firstly, calculating the centricity of the feature vectors corresponding to each node in the community, and using A ═ eij)n×nRepresenting an adjacency matrix corresponding to an undirected graph, where X is (X)1,x2,…,xn) One eigenvector representing the adjacency matrix, for an arbitrary node viThe corresponding characteristic vector value xiComprises the following steps:
in the formula: n represents the total number of nodes, xjRepresenting a node viV of a neighbor nodejThe corresponding eigenvector value, λ, is the eigenvalue corresponding to the eigenvector X of the adjacency matrix a.
Using the above formula, when the characteristic vector value x is comparediPerforming multiple iterations until the value reaches steady state, and the characteristic vector value x at this timeiI.e. representing node viThe corresponding feature vector is centrality. For convenience of calculation, the feature vector centrality x is calculated within the interval (0,1)iNormalizing to obtain n normalized feature vector centrality values xi':
Step two: and (3) improving a label propagation algorithm: the obtained normalized feature vector centrality value xi' input into the tag propagation algorithm as an initial value for the tag propagation value. And then determining an updating rule of the label propagation value based on the interest similarity and the social influence among the nodes in the community, namely further solving the latest label propagation value of the nodes according to the determined label propagation value updating rule.
The interest similarity between the nodes is the Tanimoto coefficient correlation corresponding to the interest weight sets of the same interest fields in the interest sets of the two nodes. The interest similarity calculation method among the nodes comprises the following steps: suppose node viAnd a neighbor node vjIf the same interest field number is p, the node viAnd a neighbor node vjThe interest similarity between them is:
The interest weight represents the historical request times of the node for an interest field, and is used forRepresenting a node viFor the field of interest fkInterest weight of when node viGenerating a contained interest field fkWhen requesting (2), there are:
the social influence among the nodes is embodied by the social distance between two nodes. The social distance between nodes is proportional to the data traffic through the two nodes, inversely proportional to the distance between the two nodes, and related to the degree of the nodes. The calculation formula of the social influence among the nodes is as follows:
ki=|Γ(i)|,kj=|Γ(j)| (6)
in the formula :dijFor flowing through the node viAnd a neighbor node vjThe sum of the data traffic of (a), Γ (i) and Γ (j) are respectively a node viAnd a neighbor node vjSet of (a), kiAnd k isjAre respectively node viAnd a neighbor node vjDegree of (d), dis (i, j) is node viAnd a neighbor node vjThe euclidean distance between.
According to the interest similarity and the social influence among the nodes in the community, the updating rule for determining the propagation value of the label is as follows:
wij=αTwij+(1-α)Swij (8)
in the formula :NiIs a node viV of a neighbor nodejSet of (a)jFor neighbor node vjTag activity value of, kjFor neighbor node vjDegree of (d), wijIs a node viAnd a neighbor node vjThe similarity of interest Tw betweenijAnd social influence SwijAlpha is a corresponding constant coefficient, and reflects interest similarity Tw between nodesijAnd social influence SwijThe occupied weight.
Step three: based on the normalized feature vector centrality and the improved label propagation algorithm, the CCN network is divided into a plurality of non-overlapping communities. Community CxCan not develop the community members without limit, so that some limit conditions are needed when making community decisionsAnd determining the number of the divided communities and the number of nodes contained in each community.
The specific method comprises the following steps: firstly, determining the number of divided communities, and carrying out the step one on the n normalized feature vector centrality values xi' descending order arrangement is carried out, nodes corresponding to the first k values in the nodes are selected as corresponding initial communities to obtain k initial communities, and C is used1,C2,…,CkRepresents; then, other nodes are developed for the k initial communities in a parallel mode based on a hierarchy traversal method to serve as members of the communities. And in the process of carrying out community division level traversal, if newly traversed nodes v to be added in two communitiesiAnd vjBeing a neighbouring node, i.e. node viAnd vjIf the two communities are directly connected, the two communities are merged into one community; if there are eta nodes v to joiniAnd vjIn the case of an adjacent node, the number of finally divided communities is k- η ═ p.
Secondly, the number of members in the community is limited, namely the number of nodes is limited. Firstly, an upper limit Lim of the number of members in a community is given, and in the process of hierarchical traversal, if the number of nodes in the community exceeds the upper limit value, no new node is added into the community, and member development of the community is terminated. That is to say, in the process of developing community members, a node may be contended by different communities, and the final ownership of the node is determined according to the label propagation values of the node in the different communities, because in the process of traversing the hierarchy, the same node may have different label propagation values due to different communities. Suppose node viFrom community C, respectivelyx and CyGo through the hierarchy traversal if the node viFrom community CxThe traversed label propagation value is greater than from community CyThe traversed tag propagates the value, and Community CxIf the number of the members of (1) does not reach the upper limit value, the node viSubordinate to community CxContinue community CxDevelopment of members of (1); else if Community CyDoes not reachLimit, node viSubordinate to community Cy。
The network communities divided based on the feature vector centrality and the improved label propagation algorithm are non-overlapping, and the number of the finally divided communities is known. According to the description, a pseudo code algorithm for community division is shown in fig. 2, namely, a value of the centrality of the normalized feature vector is used as an initial label propagation value, the first k communities are selected as the initial communities, and the members of the communities of the k communities are updated by using the label propagation value.
Step four: and selecting the most competitive node in each divided community to deploy the SDN controller so as to help manage community information and topology and interaction with other communities.
SDN controller deployment is based on per community CxThe method is characterized in that the competitiveness of a middle node is determined, and an SDN controller is allocated to a community C by researching several aspects of data forwarding capacity, communication cost and the like of the nodexThe node with the largest inner competitiveness. For community CxWith CpxiRepresenting nodes v within a communityiThe competitive power of (2) is as follows:
in the formula :DfxiIs a node viData forwarding capability of CoxiFor SDN controller SxDeployed in Community CxNode v ofiThe cost of the communication over the network,is a suitable adjustment factor. It can be known that the competitiveness of a node is proportional to the data forwarding capability of the node and inversely proportional to the communication cost of the node.
The data forwarding capability of the node is related to the label propagation value and the network bandwidth after the node is standardized, wherein:
the label propagation value after node normalization is:
the network bandwidth after node standardization is as follows:
in the formula :LkRepresenting a node vkA tag propagation value of bxiRepresenting a node vxAnd node viTransmission bandwidth between bxkRepresenting a node vxAnd node vkThe transmission bandwidth in between.
The data forwarding capability of the node is:
Dfxi=β•Li′+(1-β)·Bwi′ (12)
where β is the weight occupied by the tag propagation value and the network bandwidth.
The node viThe communication cost can be expressed by two aspects of transmission delay and energy consumption of the router, wherein the transmission delay is related to the transmission delay of the passing node viNumber of requests NuiAnd node viFixed time delay consumption Csi(ii) related; energy consumption and flow through node viData flow rate Tf ofiAnd energy consumption Ec to process per bit dataiIn relation to this, the calculation formula of the communication cost is:
Coxi=Nui•Csi+Tfi·Eci (13)
communication costs are normalized, and values are distributed between intervals (0,1), so that the normalized communication costs are as follows:
in the formula :CoxkRepresenting a node vxAnd node viThe cost of communication between.
From the above description, SDN controller deploys nodes viIs selected fromThe pseudo code algorithm is selected as shown in fig. 3, that is, the node with the highest competitiveness is calculated in the community according to the data forwarding capability and the communication cost among the nodes, and a controller is deployed on the node.
In order to enable the SDN controllers to effectively manage community contents and improve the content retrieval speed, an information index table (SIT) is designed for each SDN controller and used for recording the mapping of the nodes where the contents in the community are located and providing support for routing in the community; and a community topology structure table (SCT) for helping to maintain community topology information and perform inter-community routing.
The information index table (SIT) consists of three fields: content name Prefix (PRE), content name (CAN), Content Holder (CHL). The method has the main function of establishing the mapping of the content on each node in the community on the SDN controller, so that the SDN controller can clearly master the content holding condition of community members and can accurately forward the content to a requester. A node in a community may store multiple different contents (but not store the same content multiple times), and the same content may be held by multiple different nodes.
The topology structure table (SCT) comprises three fields of Adjacent Community (ACS), content name Prefix (PRE) and content transmission time (NCT). Wherein, the Adjacent Community (ACS) refers to which communities are adjacent to the current community; a content name Prefix (PRE) refers to a prefix tag for transmitting content between two communities; the number of transmission times of contents (NCT) refers to the number of times data is transmitted in total between two communities. The main function is to record the total times of transmission of different types of contents among communities, and provide support for subsequent routing query of interest packages among communities.
In the invention, ndnSIM is adopted to carry out data simulation and performance analysis, and a synthetic network topology structure consisting of 100 points and 474 edges is generated, wherein 98 points are router nodes, and 2 points are server nodes. In the initial state, all data content is stored only on the server node, and no content is stored by the router node. The community division comparison algorithms adopted by the experiment are LPAh, HPI-LPL, S-LPA and the like, and the ant colony algorithm is uniformly adopted for CCN data routing. Multiple simulation experiments were performed on a Windows system with Intel (R) core (TM) i7-9700 CPU @3.00GHz CPU, 32 GBRAM. By performing 100 experiments on each division algorithm, the average routing hop count and the cache hit rate of each community division algorithm are calculated under different interest request times.
Setting simulation environment and parameters: a specific numerical value is determined by performing simulation experiments under different parameter settings, so that the whole routing decision can achieve the optimal performance. The determination of the upper limit Lim of the number of members in the community in all the parameters is absolutely critical, and the slight performance of the route is directly influenced. All the parameters involved in the invention and their values are shown in table 1 below:
TABLE 1 parameter value settings
In order to evaluate the quality of the community division result, a concept of modularity Q is introduced to measure the quality of the community division, and the closer the value is to 1, the better the community division quality is proved, but the modularity is difficult to reach 1 in a common situation. The modularity value corresponding to different Lim values when community division is performed is changed as shown in fig. 4. It was observed that the image exhibited a peak, especially when Lim is 20, the modularity was closest to 1. Further, it can be seen from fig. 4 that the change in the modularity value rapidly increases before the peak is reached, and the decrease in the modularity gradually becomes gentle after the peak is reached. This indicates that the strong modularity establishment process is fast, but the destruction is relatively slow.
By observing the average number of route hops at different interest request times, the result is shown in fig. 5. As can be seen from the data in fig. 6: firstly, with the increase of the number of interest requests, the average routing hop count of an SI-LPA algorithm is obviously lower than that of S-LPA, LPAh and HPI _ LPA algorithms, because the SI-LPA algorithm introduces an SDN controller to manage a community, with the increase of the number of requests, the probability of requesting the same content again increases, at the moment, the content can be directly acquired in the community according to the SDN controller, and the routing hop count is reduced to a certain extent; and the SI-LPA algorithm comprehensively uses the interest similarity and social influence of the nodes to divide communities, so that the community structure of each division tends to be stable. And secondly, the routing hop counts of the SI-LPA, S-LPA and HPI-LPA algorithms finally tend to be stable, but the number of the routing hops of the LPAh algorithms is always in a fluctuation state, because the LPAh algorithms optimize the community division quality by using an objective function, but the problem of poor stability is not solved.
By requesting the same content for different nodes in the community, the cache hit rate in the community is observed for different interest request times, and the result is shown in fig. 6. As can be seen from fig. 6, as the number of requests increases, the cache hit rate of each partition policy increases, but the SI-LPA algorithm has the highest cache hit rate, because the management function of the SDN controller provides better support for caching the community content of the SI-LPA algorithm policy, so that the content whose requests are changed in the community can have higher priority, thereby increasing the cache hit rate.
According to the invention, a CCN route is researched by introducing an SDN controller and community division, and a CCN network topology is divided into a plurality of communities by combining two mainstream methods of feature vector centrality and label propagation; then, one node is selected to deploy the SDN controller in the community according to the node competitiveness, which is mainly determined by the data forwarding capability of the node and the communication cost of the community, wherein each SDN controller is composed of SIT and SCT and used for recording and topology management of network information. In addition, the network topology is simulated based on the real data set, and the experimental result shows that the scheme has good performance.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (10)
1. A CCN community division method based on node similarity and influence is characterized by comprising the following steps:
the method comprises the following steps: calculating the centrality of the feature vector corresponding to each node in the community, and standardizing the centrality of the feature vector to obtain n standardized values of the centrality of the feature vector;
step two: inputting the obtained value of the centrality of the standardized feature vector into a tag propagation algorithm to serve as an initial value of a tag propagation value, and determining an updating rule of the tag propagation value based on interest similarity and social influence among nodes in a community;
step three: dividing the CCN network into a plurality of non-overlapping communities based on the characteristic vector centrality after standardization and an improved label propagation algorithm;
step four: and selecting the most competitive node in each divided community to deploy the SDN controller so as to help manage community information and topology and interaction with other communities.
2. The CCN community division method based on node similarity and influence according to claim 1, wherein the calculation method of the centrality of the feature vector corresponding to each node in the community comprises the following steps: by A ═ e (e)ij)n×nRepresenting an adjacency matrix corresponding to an undirected graph, where X is (X)1,x2,…,xn) One eigenvector representing the adjacency matrix, for an arbitrary node viThe corresponding characteristic vector value xiComprises the following steps:
in the formula: i is more than or equal to 1 and less than or equal to n, n represents the total number of nodes, xjRepresenting a node viOf a neighbor node vjThe corresponding characteristic vector value, wherein lambda is the characteristic value corresponding to the characteristic vector X of the adjacent matrix A; when the feature vector value x is matchediPerforming multiple iterations until the value reaches steady state, at which time the characteristic vector value xiRepresenting a node viThe corresponding feature vector is centralised.
4. The CCN community division method based on node similarity and influence as claimed in claim 1 or 3, wherein the interest similarity between nodes is calculated by the following method: suppose node viAnd a neighbor node vjIf the same interest field number is p, the node viAnd a neighbor node vjThe interest similarity between them is:
when node viGenerating a contained interest field fkWhen requesting (2), there are:
5. the CCN community division method based on node similarity and influence according to claim 4, wherein the calculation formula of the social influence among the nodes is as follows:
ki=|Γ(i)|,kj=|Γ(j)| (6)
in the formula :dijFor flowing through the node viAnd a neighbor node vjThe sum of the data traffic of (a), Γ (i) and Γ (j) are respectively a node viAnd a neighbor node vjSet of (a), kiAnd k isjAre respectively node viAnd a neighbor node vjDegree of (d), dis (i, j) being node viAnd a neighbor node vjThe euclidean distance between.
6. The CCN community division method based on node similarity and influence as claimed in claim 5, wherein the updating rule for determining the label propagation value in the second step is:
wij=αTwij+(1-α)Swij (8)
in the formula :NiIs a node viV of a neighbor nodejSet of (a)jFor neighbor node vjTag activity value of, kjFor neighbor node vjDegree of (d), wijIs a node viAnd a neighbor node vjThe similarity of interest Tw betweenijAnd social influence Swijα is a constant coefficient.
7. The CCN community division method based on node similarity and influence according to claim 3 or 6, wherein the division method of non-overlapping communities in the three steps is as follows: firstly, determining the number of divided communities, and calculating the centrality value x of n standardized feature vectorsiCarrying out descending order arrangement, and selecting nodes corresponding to the first k values as corresponding initial communities to obtain k initial communities; then the k initial communities are based on hierarchy in a parallel modeThe traversal method develops other nodes as its community members.
8. The CCN community division method based on node similarity and influence as claimed in claim 7, wherein the calculation formula of node competitiveness in each community in the fourth step is as follows:
9. The CCN community division method based on node similarity and influence according to claim 8, wherein the data forwarding capability of the node is related to the label propagation value and the network bandwidth after node standardization, wherein:
the label propagation value after node normalization is:
the network bandwidth after node standardization is as follows:
in the formula :LkRepresenting a node vkA tag propagation value of bxiRepresenting a node vxAnd node viTransmission bandwidth between bxkRepresenting a node vxAnd node vkThe transmission bandwidth in between;
the data forwarding capability of the node is:
Dfxi=β•Li′+(1-β)•Bwi′ (12)
where β is the weight occupied by the tag propagation value and the network bandwidth.
10. The CCN community division method based on node similarity and influence according to claim 8 or 9, wherein the calculation formula of the communication cost of the node is as follows:
Coxi=Nui·Csi+Tfi·Eci (13)
in the formula :NuiRepresenting a node viNumber of requests of, CsiRepresenting a node viFixed time delay consumption, TfiRepresenting flow through node viData flow of (Ec)iRepresenting the energy consumption for processing data per bit;
communication costs are normalized, and values are distributed between intervals (0,1), so that the normalized communication costs are as follows:
in the formula :CoxkRepresenting a node vxAnd node viThe cost of communication between.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210198386.6A CN114513426B (en) | 2022-03-02 | 2022-03-02 | CCN community division method based on node similarity and influence |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210198386.6A CN114513426B (en) | 2022-03-02 | 2022-03-02 | CCN community division method based on node similarity and influence |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114513426A true CN114513426A (en) | 2022-05-17 |
CN114513426B CN114513426B (en) | 2023-09-15 |
Family
ID=81553821
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210198386.6A Active CN114513426B (en) | 2022-03-02 | 2022-03-02 | CCN community division method based on node similarity and influence |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114513426B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115797871A (en) * | 2022-12-22 | 2023-03-14 | 廊坊师范学院 | Analysis method and system for infant companion social network |
CN116132310A (en) * | 2023-02-17 | 2023-05-16 | 西安电子科技大学广州研究院 | Large-scale software defined network performance prediction method |
Citations (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7023979B1 (en) * | 2002-03-07 | 2006-04-04 | Wai Wu | Telephony control system with intelligent call routing |
JP2014157502A (en) * | 2013-02-15 | 2014-08-28 | Dainippon Printing Co Ltd | Server device, program and communication system |
US20170161279A1 (en) * | 2015-12-08 | 2017-06-08 | International Business Machines Corporation | Content Authoring |
CN106888163A (en) * | 2017-03-31 | 2017-06-23 | 中国科学技术大学苏州研究院 | The method for routing divided based on network domains in software defined network |
CN106886524A (en) * | 2015-12-15 | 2017-06-23 | 天津科技大学 | A kind of community network community division method based on random walk |
CN107194818A (en) * | 2017-04-13 | 2017-09-22 | 天津科技大学 | Label based on pitch point importance propagates community discovery algorithm |
CN107862617A (en) * | 2017-10-20 | 2018-03-30 | 江苏大学 | A kind of microblogging community division method based on user's comprehensive similarity |
CN108073944A (en) * | 2017-10-18 | 2018-05-25 | 南京邮电大学 | A kind of label based on local influence power propagates community discovery method |
CN108595684A (en) * | 2018-05-04 | 2018-09-28 | 中南大学 | A kind of overlapping community discovery method and system based on preferential learning mechanism |
US20180341696A1 (en) * | 2017-05-27 | 2018-11-29 | Hefei University Of Technology | Method and system for detecting overlapping communities based on similarity between nodes in social network |
US20180349779A1 (en) * | 2015-11-25 | 2018-12-06 | Systamedic Inc. | Method and descriptors for comparing object-induced information flows in a plurality of interaction networks |
CN109067588A (en) * | 2018-08-21 | 2018-12-21 | 电子科技大学 | A kind of semi-supervised non-overlap community discovery method based on partial tag information |
CN110909173A (en) * | 2019-11-13 | 2020-03-24 | 河海大学 | Non-overlapping community discovery method based on label propagation |
CN111159577A (en) * | 2019-12-31 | 2020-05-15 | 北京明略软件系统有限公司 | Community division method and device, storage medium and electronic device |
CN113254999A (en) * | 2021-06-04 | 2021-08-13 | 郑州轻工业大学 | User community mining method and system based on differential privacy |
WO2021189729A1 (en) * | 2020-03-27 | 2021-09-30 | 深圳壹账通智能科技有限公司 | Information analysis method, apparatus and device for complex relationship network, and storage medium |
-
2022
- 2022-03-02 CN CN202210198386.6A patent/CN114513426B/en active Active
Patent Citations (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7023979B1 (en) * | 2002-03-07 | 2006-04-04 | Wai Wu | Telephony control system with intelligent call routing |
JP2014157502A (en) * | 2013-02-15 | 2014-08-28 | Dainippon Printing Co Ltd | Server device, program and communication system |
US20180349779A1 (en) * | 2015-11-25 | 2018-12-06 | Systamedic Inc. | Method and descriptors for comparing object-induced information flows in a plurality of interaction networks |
US20170161279A1 (en) * | 2015-12-08 | 2017-06-08 | International Business Machines Corporation | Content Authoring |
CN106886524A (en) * | 2015-12-15 | 2017-06-23 | 天津科技大学 | A kind of community network community division method based on random walk |
CN106888163A (en) * | 2017-03-31 | 2017-06-23 | 中国科学技术大学苏州研究院 | The method for routing divided based on network domains in software defined network |
CN107194818A (en) * | 2017-04-13 | 2017-09-22 | 天津科技大学 | Label based on pitch point importance propagates community discovery algorithm |
US20180341696A1 (en) * | 2017-05-27 | 2018-11-29 | Hefei University Of Technology | Method and system for detecting overlapping communities based on similarity between nodes in social network |
CN108073944A (en) * | 2017-10-18 | 2018-05-25 | 南京邮电大学 | A kind of label based on local influence power propagates community discovery method |
CN107862617A (en) * | 2017-10-20 | 2018-03-30 | 江苏大学 | A kind of microblogging community division method based on user's comprehensive similarity |
CN108595684A (en) * | 2018-05-04 | 2018-09-28 | 中南大学 | A kind of overlapping community discovery method and system based on preferential learning mechanism |
CN109067588A (en) * | 2018-08-21 | 2018-12-21 | 电子科技大学 | A kind of semi-supervised non-overlap community discovery method based on partial tag information |
CN110909173A (en) * | 2019-11-13 | 2020-03-24 | 河海大学 | Non-overlapping community discovery method based on label propagation |
CN111159577A (en) * | 2019-12-31 | 2020-05-15 | 北京明略软件系统有限公司 | Community division method and device, storage medium and electronic device |
WO2021189729A1 (en) * | 2020-03-27 | 2021-09-30 | 深圳壹账通智能科技有限公司 | Information analysis method, apparatus and device for complex relationship network, and storage medium |
CN113254999A (en) * | 2021-06-04 | 2021-08-13 | 郑州轻工业大学 | User community mining method and system based on differential privacy |
Non-Patent Citations (7)
Title |
---|
GUIQIONG XU ECT.: "TNS-LPA: An Improved Label Propagation Algorithm for Community Detection Based on Two-Level Neighbourhood Similarity", 《IEEE ACCESS ( VOLUME: 9)》 * |
JUNHAI LUO , LEI YE: "Label propagation method based on bi-objective optimization for ambiguous community detection in large networks", 《SCIENTIFIC REPORTS 》 * |
YAN XING ECT.: "A Node Influence Based Label Propagation Algorithm for Community Detection in Networks", 《THE SCIENTIFIC WORLD JOURNAL》 * |
宋琛;张贤坤;费松;荚佳;刘栋;: "基于随机游走相似度矩阵的改进标签传播算法", 计算机应用与软件, no. 08 * |
潘曙灿;许青林;: "融合特征向量中心性与标签熵的标签传播算法", 计算机工程与科学, no. 08 * |
蔡增玉,崔梦梦,侯佳林,张建伟: "基于残差神经网络的心脏病预测系统的设计与实现", 《现代电子技术》 * |
郝梓琳;李雷;施化吉;: "基于节点综合相似度的多标签传播社区划分算法", 计算机应用研究, no. 06 * |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115797871A (en) * | 2022-12-22 | 2023-03-14 | 廊坊师范学院 | Analysis method and system for infant companion social network |
CN116132310A (en) * | 2023-02-17 | 2023-05-16 | 西安电子科技大学广州研究院 | Large-scale software defined network performance prediction method |
CN116132310B (en) * | 2023-02-17 | 2023-10-20 | 西安电子科技大学广州研究院 | Large-scale software defined network performance prediction method |
Also Published As
Publication number | Publication date |
---|---|
CN114513426B (en) | 2023-09-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Yeh et al. | VIP: A framework for joint dynamic forwarding and caching in named data networks | |
CN114513426B (en) | CCN community division method based on node similarity and influence | |
CN104717304B (en) | A kind of CDN P2P content optimizations select system | |
WO2015048773A2 (en) | System and method for joint dynamic forwarding and caching in content distribution networks | |
JP6190288B2 (en) | Cache control apparatus, method, and program | |
Hou et al. | Bloom-filter-based request node collaboration caching for named data networking | |
Wu et al. | MBP: A max-benefit probability-based caching strategy in information-centric networking | |
Shan et al. | Proactive caching placement for arbitrary topology with multi-hop forwarding in ICN | |
Zhang et al. | NCPP-based caching and NUR-based resource allocation for information-centric networking | |
Mahdian et al. | MinDelay: Low-latency joint caching and forwarding for multi-hop networks | |
Nguyen et al. | Rethinking virtual link mapping in network virtualization | |
EP3193490A1 (en) | Method and system for distributed optimal caching of content over a network | |
Yao et al. | A SMDP-based forwarding scheme in named data networking | |
Delvadia et al. | CCJRF-ICN: A novel mechanism for coadjuvant caching joint request forwarding in information centric networks | |
Mahananda et al. | Performance of homogeneous and heterogeneous cache policy for named data network | |
Aloulou et al. | Effective controller placement in controller-based named data networks | |
CN115473854A (en) | Intelligent flow control method for multi-mode network | |
Zhou et al. | Clustered k-center: Effective replica placement in peer-to-peer systems | |
Barford et al. | Cache placement methods based on client demand clustering | |
Lin et al. | Proactive multipath routing with a predictive mechanism in software‐defined networks | |
CN108512685B (en) | Information center network flow control method and device | |
Xu et al. | Minimizing bandwidth cost of CCN: a coordinated in-network caching approach | |
Zhang et al. | Pextcuts: A high-performance packet classification algorithm with pext cpu instruction | |
Hsu et al. | An integrated end-to-end QoS anycast routing on DiffServ networks | |
Alahmadi | A New Efficient Cache Replacement Strategy for Named Data Networking |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |