CN104038487B - A kind of networking component clustering method based on Hausdorff distances - Google Patents
A kind of networking component clustering method based on Hausdorff distances Download PDFInfo
- Publication number
- CN104038487B CN104038487B CN201410245986.9A CN201410245986A CN104038487B CN 104038487 B CN104038487 B CN 104038487B CN 201410245986 A CN201410245986 A CN 201410245986A CN 104038487 B CN104038487 B CN 104038487B
- Authority
- CN
- China
- Prior art keywords
- component
- cluster
- networking
- networking component
- topology
- 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.)
- Active
Links
Classifications
-
- 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
- Y02D30/50—Reducing energy consumption in communication networks in wire-line communication networks, e.g. low power modes or reduced link rate
Landscapes
- Data Exchanges In Wide-Area Networks (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The present invention provides a kind of networking component clustering methods based on Hausdorff distances, general networking component model is first defined, includes component mark NID (Node ID), component behavior description NBD (Node Behavior Description)；Then using multidimensional Hausdorff distances as tool, Unify legislation is carried out to component with Description Matrix, based on the otherness given threshold H ' between the cluster and cluster divided by function0, based on cluster inner assembly self-similarity given threshold H0, based on cluster inner assembly scheduling flexibility ratio setting Topology Potential centrad P0, by threshold determination three times, complete the sub-clustering of networking component.For the demand for services on Proper Match upper strata and the Internet resources of lower floor, suitable network group and its internal corresponding networking component are selected, realizes the demands such as efficient utilization, the green energy conservation of Internet resources.
Description
Technical field
A kind of networking component clustering method based on Hausdorff distances, belongs to network virtualization field.
Background technology
Existing internet is using the design philosophy of " hourglass model ", the feature with " triple bindings ", i.e.,：" the money of service
" control and the data binding " and " identity and position binding " of source and binding positions ", network.This network system and mechanism are phases
To " static state " and " rigid ", evolution on this basis is difficult to break through the limitation of original design thought with development, can not be from root
Meet the communication requirements such as information network " high speed ", " efficient ", " magnanimity ", " ubiquitous " on this, it is difficult to solve network scalability, move
The problems such as dynamic property, safety, it is more difficult to realize the efficiently utilization, energy saving etc. of Internet resources.
Future network has many components, whole including server, router, memory, interchanger, base station, AP, computer
End, various mobile terminals, various sensing nodes；These components are can to perceive, addressable, can route, can control.It is based on
Networking component with identity function is clustered cluster by the networking component clustering method of Hausdorff distances.For Proper Match
The demand for services on upper strata and the Internet resources of lower floor are completed science decision by appropriate network game playing algorithm, are selected for service
Suitable network group and its internal corresponding networking component, the wisdom for completing service obtain, and provide good user service body
It tests, realizes the demands such as efficient utilization, the green energy conservation of Internet resources.
Invention content
Technical problem：The purpose of the present invention is be based on Internet resources virtualization visual angle, with reference to scheduling of resource flexibility ratio ensure,
General networking component Definition Model is established, completing component using multidimensional Hausdorff distances clusters, applied to future network
In resources effective utilization and green energy conservation.
Technical solution：Method proposes a kind of general networking components to define and the cluster based on Hausdorff distances
Method.It is efficiently utilized for Internet resources, the demands such as green energy conservation, general definition is carried out to networking component first, uses group
Component is described in part Description Matrix；Then using the multidimensional Hausdorff distances of two point set similarity measures of characterization as tool,
The weight of multidimensional Hausdorff distances is set according to the function of cluster, is set based on the otherness between the cluster and cluster divided by function
Determine threshold value H'0, based on cluster inner assembly self-similarity given threshold H0, based on cluster inner assembly scheduling flexibility ratio setting Topology Potential center
Spend P0, by threshold determination three times, complete the sub-clustering of networking component.
The general definition that the present invention is perceived under future network based on module information, utilizes multidimensional Hausdorff distances
The method for measuring component similarity completes sub-clustering, it includes following steps：
Step 1. defines networking component：
1) static defining：In networking component layer, define networking component for it is a kind of complete the acquisitions of data, generation, storage,
The network equipment of one or more functions such as forwarding, reception, calculating；It proposes to carry out one group of networks of label using component mark NID
Part is defined as follows：
In above formula, NtypeRepresent the type (transmission assembly, storage assembly etc.) of networking component, NdeviceRepresent networking component
Equipment of itself information, ω () proxy component mark generating function；
2) dynamic definition：
Information Perception based on networking component, the characteristic information of collection network component；And on this basis, component rows are proposed
For describe NBD concept, function, topology or performance characteristics of networking component etc. are further described, with to component behavior into
Row characterization.The behavior description NBD of networking component is defined as follows：
In above formula, T, P, F correspond to topological behavior, performance behavior and behaviour respectively.For NBD, topology information packet
Include module positionComponent subordinate relationComponent syntopleAnd degree of communicationDeng；Performance information includes bandwidth
EnergyDelay performancePacket loss performanceAnd stabilityDeng；Functional information includes component typeAssembly functionOperatorAnd security levelDeng；
Clustering method of the step 2. based on Hausdorff distances：
1) calculates average Topology Potential centradCalculate the average Topology Potential centrad for the cluster that New Parent is waited for add in
Wherein, n represents the number of cluster inner assembly, P (ni) represent the Topology Potential of component i, c (nj) it is node njHandle
Ability changes with the variation of the state of node.The distance between d (i, j) component i and component j, with the length table of shortest path
Show.bwa(j) the available bandwidth summation external for component j is used for representing here the sphere of action of component j.In network topology not
In the case that shortest path is certain between change, component, the available CPU abilities of component j are bigger, then the component is to component niEffect
Bigger, vice versa.The Topology Potential of component i is the summation that all components act on it；
2) the Hausdorff distances H of computation modules and clusterAi：The multidimensional Hausdorff distances of computation module i and cluster A：
It is the weight depending on component cluster property, such as：If cluster is memory,Weight compared with
Greatly；Cluster is transponder, thenWeight is larger；Points for attention：Calculate specific Hausdorff apart from when, should be a certain attribute
Binary system bit, be converted to the decimal system and calculated；
3) calculates the Hausdorff distances H (A', B) of new cluster A' and adjacent cluster B：It calculates between new cluster A' and adjacent cluster B
Distance：
H (A', B)=max (HA'B,HBA')
4) calculates the average Topology Potential centrad of new cluster A'Calculate the average Topology Potential for the cluster that New Parent is waited for add in
Centrad：
Wherein, n ' is that the quantity of component in original cluster adds 1；
5) joint decisions：IfCluster A is then added in, forms new cluster
A'；Otherwise component i oneself clusters.Wherein, P0It represents based on cluster inner assembly scheduling flexibility ratio setting Topology Potential centrad, H0It represents
Based on cluster inner assembly self-similarity given threshold, H'0It represents to set threshold based on the otherness between the cluster and cluster divided by function
Value.
Wherein：
The average Topology Potential centrad is：The average Topology Potential centrad of cluster embodies the tight ness rating in cluster between component
Relationship and component topology location and action intensity, therefore the flexibility of cluster inner assembly scheduling can be weighed.The structure of cluster is not
Can be excessive, it otherwise can not meet the flexible dispatching of resource, and resource is excessively disperseed, be not easy to manage concentratedly；Here cluster is set
Flexibility ratio with average Topology Potential centrad P0It represents, meets flexibility ratio more than P0。
The Hausdorff distances of the component and cluster are：The similarity of a component and cluster is weighed, component can be used
Description Matrix by certain numerical value conversion, carries out similarity calculation, according to the function of cluster, multidimensional with component each in cluster
Every one-dimensional weighted of Hausdorff distances；Here given threshold H0。
The cluster and the Hausdorff distances of cluster are：There is different functional characters in view of different clusters, for example, having
The different clusters such as storage, forwarding, processing, have certain otherness between cluster and cluster, component will add in cluster, should not destroy original
The property of cluster, the difference similarity between cluster and cluster can also be weighed with Hausdorff；Here given threshold H'0。
Advantageous effect
The present invention is based on Internet resources to virtualize visual angle, using multidimensional Hausdorff distance metric component similarities, with reference to
Scheduling of resource flexibility ratio ensures, establishes wisdom networking component model and behavior aggregate, high applied to the resource in future network
Effect utilizes and green energy conservation.
Description of the drawings
Fig. 1 is the flow chart of the present invention.
Specific embodiment
The present invention is Internet resources virtualization main line, based on module information describes, with two point set similarity measures
Multidimensional Hausdorff distances are tool, with reference to cluster inner assembly dispatching flexibility to ensure, the function difference between comprehensive cluster and cluster
Property, wisdom networking component model and behavior aggregate are established, applied to the resources effective utilization and green energy conservation in future network.
Networking component defines method：
In networking component layer, define networking component for it is a kind of complete the acquisitions of data, generation, storage, forwarding, reception,
The network equipment of one or more functions such as calculating；It proposes to carry out one networking component of label using component mark NID, definition is such as
Under：
In above formula, NtypeRepresent the type (transmission assembly, storage assembly etc.) of networking component, NdeviceRepresent networking component
Equipment of itself information, ω () proxy component mark generating function.
Information Perception based on networking component, the characteristic information of collection network component；And on this basis, component rows are proposed
For describe NBD concept, function, topology or performance characteristics of networking component etc. are further described, with to component behavior into
Row characterization.The behavior description NBD of networking component is defined as follows：
In above formula, T, P, F correspond to topological behavior, performance behavior and behaviour respectively.For NBD, topology information packet
Include module positionComponent subordinate relationComponent syntopleAnd degree of communicationDeng；Performance information includes bandwidth
EnergyDelay performancePacket loss performanceAnd stabilityDeng；Functional information includes component typeAssembly functionOperatorAnd security levelDeng.
Component describes the generation of Main Basiss information Perception, can be divided into two kinds according to describing mode：Data describe and semanteme
Description.Data description refers mainly to be presented the content of behavior description with the form of data, and it is specific to be primarily directed to networking component
The description of index parameter, such as the specific load capacity of network node, actual average packet loss.The source of these data description
It is divided into two kinds, a kind of fixed performance index for network node, another kind is the actual measurement Parameter analysis of network node, surveys class
Data description needs to obtain from a large amount of network measurement data, and it is brief to be that a large amount of truthful datas obtain after statistical analysis
Description, need embodiment behavior description simplifies characteristic.
Semantic description refers mainly to be presented the content of behavior description with semantic form.Such as network entity stability is divided into
Entity description is from excellent to 9 poor ranks by the factors such as high, normal, basic three ranks, integration node time delay, shake, packet loss, load
Etc..Specific research point in semantic description includes：Rank, the rank of semantic description of semantic description differentiate.Semantic description
Rank refers in a kind of description space of behavioral indicator, the description number of levels of permission.The rank differentiation of semantic description refers to
The description rank that one behavior of judgement best suits, rank, which differentiates, mainly includes 3 kinds of forms：Threshold value form, contrast version and integration
Form.
For certain specifies behaviors, need to determine the description space { x of its behavior description firstk, the behavior describes total
Rank is n.For the behavior description x of wherein certain level-onek, its specific degree of membership is set as uk, and use uk/xkRepresent at different levels
Other behavior description, then the final description b of the behavior is as follows：
Component Description Matrix
Position：(float x,float y)；
Subordinate relation：IP address network number 24bit；
Component proximity relations：Hello packets are sent, neighbours' component is replied after receiving to be confirmed, int；
Memory capacity：The memory capacity (M) of component；
Bandwidth performance：Message transmission rate, float (M/s)；
Delay performance：The processing time of component forwarding packet, float (s)；
Packet loss performance：Due to transfer capability limit, caused by packet loss probability, float；
Stability：The performance indicator that the combined factors such as time delay, shake, packet loss, the load of component determine, given threshold,
It is divided into excellent poor 3 rank 2bit；
Component type：Router, server, interchanger etc., 3bit；
Assembly function：It is divided into storage, forwarding, processing 3,3bit；
Operator：3bit；
Security level：Fire wall firewall points are WEP, WPA-PSK, WPA2-PSK, the encryption levels such as WPA2, WPA,
2bit；
Hausdorff is apart from clustering method：
Hausdorff multidimensional distances, establish the distance between component cluster；Component obtains component cluster by cluster, with component
Cluster builds group for unit.
Three-dimensional Hausdorff distances between two components：
hij=a | li-lj|+β|ri-rj|+γ|ci-cj| (5)
Wherein α, β, γ are the distance coefficients depending on component cluster property.Such as：If cluster is memory, γ weights
It is larger；Cluster is transponder, then β weights are larger.
Hausdorff distances between cluster and component：
It is the weight depending on component cluster property, such as：If cluster is memory,Weight compared with
Greatly；Cluster is transponder, thenWeight is larger.
Points for attention：Calculate specific Hausdorff apart from when, binary system bit of a certain attribute should be converted to
The decimal system is calculated.
Hausdorff distances between cluster A and cluster B：
H (A, B)=max (HAB,HBA)
Topology Potential centrad：
Each component cluster regarded to the physical system for including n component as, the interaction between component is by between component
Distance, hop count influence, and the sphere of action of component then applicable components external bandwidth represent.Each component considers three categories
Property, i.e., the length of the shortest path of other assemblies, component arrive it in the available CPU processing capacities of each component, this component to cluster
The sum of available bandwidth of his component.The Topology Potential P calculation formula of component are as follows：
Wherein, n represents the number of cluster inner assembly, P (ni) represent the Topology Potential of component i, c (nj) it is node njHandle
Ability changes with the variation of the state of node.The distance between d (i, j) component i and component j, with the length table of shortest path
Show.bwa(j) the available bandwidth summation external for component j is used for representing here the sphere of action of component j.In network topology not
In the case that shortest path is certain between change, component, the available CPU abilities of component j are bigger, then the component is to component niEffect
Bigger, vice versa.The Topology Potential of component i is the summation that all components act on it.
Claims (1)
1. a kind of networking component clustering method based on Hausdorff distances, it includes following steps：
Step 1 defines networking component：
1) static definings：In networking component layer, define networking component and be a kind of acquisition for completing data, generation, storage, turn
Send out, receive, calculating the network equipment of one or more functions；It proposes to carry out one networking component of label using component mark NID, it is fixed
Justice is as follows：
In above formula, NtypeRepresent the type of networking component, NdeviceRepresent networking component equipment of itself information, ω () proxy component
Identify generating function；
2) dynamic definitions：
Information Perception based on networking component, the characteristic information of collection network component；And on this basis, propose that component behavior is retouched
The concept of NBD is stated, the function, topology or performance characteristics of networking component are further described, to be characterized to component behavior；
The behavior description NBD of networking component is defined as follows：
In above formula, T, P, F correspond to topological behavior, performance behavior and behaviour respectively；For NBD, topology information includes group
Part positionComponent subordinate relationComponent syntopleAnd degree of communicationPerformance information includes bandwidth performance
Delay performancePacket loss performanceAnd stabilityFunctional information includes component typeAssembly functionOperation
QuotientAnd security level
Step 2：
1) calculates average Topology Potential centradCalculate the average Topology Potential centrad for the cluster that New Parent is waited for add in：
Wherein, n represents the number of cluster inner assembly, P (ni) represent the Topology Potential of component i, c (nj) it is node njCan processing capacity,
Change with the variation of the state of node；D (i, j) is the distance between component i and component j, is represented with the length of shortest path；
bwa(j) the available bandwidth summation external for component j is used for representing here the sphere of action of component j；Network topology is constant, group
In the case that shortest path is certain between part, the available CPU abilities of component j are bigger, then the component is bigger to the effect of component i,
Vice versa；The Topology Potential of component i is the summation that all components act on it；
2) the Hausdorff distances H of computation modules and clusterAi：The multidimensional Hausdorff distances of computation module i and cluster A：
It is the weight depending on component cluster property；
3) calculates the Hausdorff distances H (A', B) of new cluster A' and adjacent cluster B：Calculate between new cluster A' and adjacent cluster B away from
From：
H (A', B)=max (HA'B,HBA')
4) calculates the average Topology Potential centrad of new cluster A'Calculate the average Topology Potential center for the cluster that New Parent is waited for add in
Degree：
Wherein, n ' is that the quantity of component in original cluster adds 1；
5) joint decisions：IfHAi<H0、HA'B<H′0、Cluster A is then added in, forms new cluster A'；Otherwise component i
Oneself cluster；Wherein, P0It represents based on cluster inner assembly scheduling flexibility ratio setting Topology Potential centrad, H0It represents based on cluster inner assembly
Self-similarity given threshold, H '0It represents based on the otherness given threshold between the cluster and cluster divided by function.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410245986.9A CN104038487B (en) | 2014-06-05 | 2014-06-05 | A kind of networking component clustering method based on Hausdorff distances |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201410245986.9A CN104038487B (en) | 2014-06-05 | 2014-06-05 | A kind of networking component clustering method based on Hausdorff distances |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104038487A CN104038487A (en) | 2014-09-10 |
| CN104038487B true CN104038487B (en) | 2018-06-08 |
Family
ID=51469076
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201410245986.9A Active CN104038487B (en) | 2014-06-05 | 2014-06-05 | A kind of networking component clustering method based on Hausdorff distances |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104038487B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002080458A1 (en) * | 2001-03-30 | 2002-10-10 | Nokia Corporation | Method for configuring a network by defining clusters |
| CN101287269A (en) * | 2008-04-22 | 2008-10-15 | 中国移动通信集团设计院有限公司 | Mobile communication network optimizing method, device and system |
| CN102594924A (en) * | 2012-03-29 | 2012-07-18 | 北京交通大学 | Internet architecture and internet service method and system thereof |
-
2014
- 2014-06-05 CN CN201410245986.9A patent/CN104038487B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2002080458A1 (en) * | 2001-03-30 | 2002-10-10 | Nokia Corporation | Method for configuring a network by defining clusters |
| CN101287269A (en) * | 2008-04-22 | 2008-10-15 | 中国移动通信集团设计院有限公司 | Mobile communication network optimizing method, device and system |
| CN102594924A (en) * | 2012-03-29 | 2012-07-18 | 北京交通大学 | Internet architecture and internet service method and system thereof |
Non-Patent Citations (2)
| Title |
|---|
| 可重构网络中基于中心度与拓扑势排序的资源分配算法;王子厚等;《通信学报》;20120831;第33卷(第8期);10-20 * |
| 智慧网络组件协同机制研究;郜帅等;《电子学报》;20130731;第41卷(第7期);1261-1267 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104038487A (en) | 2014-09-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zhang et al. | A kind of effective data aggregating method based on compressive sensing for wireless sensor network | |
| Zhou et al. | Social big-data-based content dissemination in internet of vehicles | |
| CN107566194B (en) | Method for realizing cross-domain virtual network mapping | |
| US9426233B2 (en) | Multi-objective server placement determination | |
| US8713151B2 (en) | Configuration method and system of complex network and configuration and management module of server resources | |
| CN109446385B (en) | Method for establishing network resource equipment map and using method of equipment map | |
| CN114040272B (en) | Path determination method, device and storage medium | |
| CN113992259B (en) | Method for constructing time slot resource expansion graph | |
| Feng et al. | Confident information coverage hole prediction and repairing for healthcare big data collection in large-scale hybrid wireless sensor networks | |
| Marche et al. | Navigability in social networks of objects: The importance of friendship type and nodes' distance | |
| CN110298687A (en) | A kind of region attraction appraisal procedure and equipment | |
| Atzori et al. | A SIoT-aware approach to the resource management issue in mobile crowdsensing | |
| Cao et al. | A mobility-supported routing mechanism in industrial IoT networks | |
| Raptis | When wireless crowd charging meets online social networks: A vision for socially motivated energy sharing | |
| Wu et al. | Data transmission scheme based on node model training and time division multiple access with IoT in opportunistic social networks | |
| Khan et al. | Data aggregation mechanisms in the internet of things: a study, qualitative and quantitative analysis | |
| Zhang et al. | Cellular traffic offloading via link prediction in opportunistic networks | |
| Rana et al. | Energy-efficient routing for wireless sensor network using genetic algorithm and particle swarm optimisation techniques | |
| CN111935748B (en) | Virtual network resource allocation method with high reliability and load balance | |
| Khan et al. | βDSC2DAM: beta-dominating set centered Cluster-Based Data Aggregation mechanism for the Internet of Things | |
| CN104038487B (en) | A kind of networking component clustering method based on Hausdorff distances | |
| Wibowo | Classification And Clustering Performances on The Protocol of The Wireless Sensor Network-Chimpanzee Leader Election Optimization | |
| Tao et al. | Efficient task allocation for mobile crowd sensing based on evolutionary computing | |
| Allani et al. | Towards better data management/gathering in connected environments | |
| CN108596778A (en) | A kind of community division method based on space of interest |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |