CN101521679A - Self-organizing method based on composite structured peer-to-peer network - Google Patents
Self-organizing method based on composite structured peer-to-peer network Download PDFInfo
- Publication number
- CN101521679A CN101521679A CN200910026197A CN200910026197A CN101521679A CN 101521679 A CN101521679 A CN 101521679A CN 200910026197 A CN200910026197 A CN 200910026197A CN 200910026197 A CN200910026197 A CN 200910026197A CN 101521679 A CN101521679 A CN 101521679A
- Authority
- CN
- China
- Prior art keywords
- dna
- network
- node
- value
- territory
- 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.)
- Pending
Links
Images
Landscapes
- Multi Processors (AREA)
Abstract
The invention relates to a self-organizing method based on a composite structured peer-to-peer network, which comprises the following steps: establishing a two-layer structured logical network model, performing corresponding adjustments on the main machine communication in the network model along with the changes of the network size and the main machine performance, and performing corresponding maintenance on resources in the network so as to achieve the characteristics of self adaptation and self organization to better ensure the network performance . Simultaneously, the network size, the network running conditions, and the like are manageable, thus the network is a manageable P2P and the blank condition of P2P management is changed.
Description
Technical field
The present invention sets up a composite structured peer-to-peer network, and this network has taken into full account the difference between the joint behavior, has very strong self-organization, manageability and dilatancy, and in this network operation maintenance process, realize the self-organizing strategy.
Background technology
The application of peer-to-peer network in the Internet (P2P) has become one of main application of the Internet, and the pattern of P2P also becomes the preferred mode of many novel business, and P2P has produced significant effects to whole Internet service pattern and oversight mechanism.
The P2P technology is widely used in fields such as file-sharing, Internet video, the networking telephone, online game, with the characteristics of distributed resource sharing and parallel transmission, for the user provides more resources, higher available bandwidth and better service quality.The P2P node does not rely on Centroid but relies on network edge node, realizes self-organizing and reciprocity resource discovering of cooperating and shared, therefore has advantages such as self-organizing, extensibility, robustness, fault-tolerance and load balancing.P2P uses the flow that is produced and has up-downgoing flow symmetry, hidden, the medium characteristic of data set of flow.
P2P application kind is many, various informative at present, does not have unified procotol standard, and its architecture and organizational form are also in continuous development.The three generations has been experienced in the development of P2P architecture: the P2P structure that the first generation is centralized: whole network has a concentrated server, is responsible for the Resource TOC of storage P2P network; Each node is published to server with shared resource, and server returns to node with the resource retrieval result, and node provides point-to-point service according to return results and other node communication, typical P2P application software such as Napster.Second generation P2P is a kind of complete acentric distributed network: the node retrieve resources adopts the method for inundation, and the traditional mode of searching of pin, has proposed structured P 2 P network, and modes such as use hash algorithm are finished the distributed structured network function.Third generation P2P is a kind of hybrid-type architecture, possess simultaneously preceding two generation architecture high efficiency and the advantage of fault-tolerance, be the main aspect of present research and development.
The current P2P procotol standard that also neither one is unified, existing hybrid P2P has merged the characteristics of centralized P2P and distributed P 2 P, but the self-organizing characteristic of the P2P network that also weakened, the dynamic adaptable of network is restricted.
Summary of the invention
Technical problem: the objective of the invention is to propose self-organizing method based on composite structured peer-to-peer network.By setting up two-layer structured logic network model,, main-machine communication in the network model is adjusted accordingly along with the variation of network size and host performance, and resources in network safeguarded accordingly, reach self adaptation, the characteristics of self-organizing, thus guaranteed network performance preferably.Simultaneously, network size among the present invention, network operation situation etc. all is manageable, is a kind of manageable P2P, having changed P2P can managerial blank situation.
Technical scheme: the self-organizing method based on composite structured peer-to-peer network of the present invention comprises the steps:
A node adds network, according to network operation situation, is defined as DNA or ordinary node, and node adds network and is divided into:
1) at first extracts this machine performance parameter, calculate this machine value of utility according to U value calculative strategy;
2) game server is according to DNA quantity, and whether the decision of the ratio of total number of nodes increases new upper layer node; If total node number is S, stipulate that a DNA allows to connect N ordinary node at most, then required DNA quantity is the supremum of S/N in the network; The record that online DNA quantity can be by checking server end in the network as can be known; Greater than online DNA quantity, then compare the value of utility of all second layer nodes in this territory as required DNA quantity, find to have peaked main frame, making it is new DNA quantity, upgrades the DNA tabulation;
3) choose the result according to DNA and determine the node identity, and return new DNA tabulation to newly added node;
4) be ordinary node as this machine, then, survey the size of all the DNA institute scopes in this territory, calculate the value of utility of each DNA and the ratio of its scope size then according to the DNA tabulation, select a DNA of ratio maximum to send connection request, disconnection is communicated by letter with other DNA's;
5), then, join the DHT network that other DNA node forms, and open upper strata distributed network storage and processing capacity, lower floor's centralized network agent functionality according to the DNA tabulation if this machine is DNA;
Step 2, network node are dynamically adjusted strategy
The P2P network is the network of a dynamic change, according to the situation of the network operation, nodal function in the network is dynamically adjusted, and realizes the self-organizing characteristic of network;
21) ordinary node is changed to DNA: there have new node to add in network to be fashionable, needs to increase the DNA node, carries out DNA and choose strategy, and the node that is selected will be transformed to new DNA node by ordinary node; Server joins the DNA that newly chooses in the DNA tabulation, is convenient to ordinary node and sends probe message.New DNA receives probe requests thereby, will confirm with server, confirms successfully will open the DNA functional module;
22) standby DNA upgrades to DNA: when a DNA territory interior nodes situation changes, DNA will choose standby DNA again; DNA makes that the ordinary node of online value of utility maximum in the territory is standby DNA, when having node to add in the territory or withdrawing from, the ordinary node of value of utility maximum in the territory will be searched again, making it is new standby DNA, when DNA exits network or takes place when unusual, standby DNA will be automatically upgraded to DNA, send the DNA upgrading message to server, open the DNA functional module; Other ordinary node will directly send connection request to standby DNA in the territory;
23) DNA deteriorates to ordinary node: when network size reduces, ordinary node connects the connection rate of DNA less than 50% in the network, be that ordinary node quantity/DNA quantity is during less than N/2, server can recomputate the DNA quantity that needs in the network, and according to the height of the value of utility of DNA, the not high DNA of notice value of utility withdraws from the DHT network, deteriorate to ordinary node, again obtain the DNA tabulation, survey all DNA, select to connect the highest DNA of average utility value.
Beneficial effect: the present invention has the following advantages:
1, makes network configuration flexibly stable, improve the robustness and the stability of network
In this method node in the network is divided into different territories by respective rule, when node connection situation changed in the network, the node in the network in the same territory will real-time perception arrive, and made corresponding adjustment.The variation in a territory changes to adjust to be finished in the territory, does not have influence substantially for node in other territory, has reduced the traffic in the network and has made network stabilization flexible.
At the limitation of the hybrid P2P of the third generation, can not better solve because problem such as super node self the fragility local nodes that may cause is isolated, the DNA in this method chooses strategy and standby DNA and chooses strategy and solve these problems preferably.This method adopts hybrid-type network configuration, and DNA has the similar function of super node, but is served as by stationary nodes unlike super node, but carries out the DNA function according to the node of the situation choice of dynamical better performances of node in the network.Along with the variation of network size, node, the DNA node can be adjusted accordingly, upgrades to DNA as ordinary node, and DNA deteriorates to ordinary node etc.When the DNA node exits network or lost efficacy, standby DNA will replace DNA to safeguard the resource in the group and the service of P2P network will be provided, and carry out work such as distributed treatment simultaneously in addition.Standby DNA make DNA withdraw from or unusual situation under, still can greatly improve stability of network for node provides service.
Abnormality processing is an important component part of this method, and at withdrawing from unusually of node in the network, system adopts polling mechanism.To in time adjust network when unusual detecting, and to server report abnormal information, the real-time of maintenance server end data.By abnormality processing, make whole stability of network higher.
2, network can be managed, and makes network have extensibility
With respect to first and second and three generations P2P, this method makes network that manageability preferably be arranged, and mainly shows the scale of network, and the connection situation and the nodes sharing fileinfo of node all can be managed and be controlled.The last node of DNA connects shared resource tabulation in tabulation and the group, is responsible for they are safeguarded.
After DNA loads the snmp management function in this method, will realize distributed resource management.Hybrid-type network architecture helps building of distributed management network, and this method provides expansion interface for network management in setting up the process of communication system.The management of resource managed in DNA made things convenient for the territory when carrying out centralized agency.The DNA record also upgrades group shared resource of interior each node and tabulates, and DNA can manage the Resources list in the group, deletes some irrational resource-sharing.When the retrieval request of receiving node, also can manage, limit some information, help the sound development of P2P network.Next step can be transplanted to ripe at present network management module on the DNA node, thereby realizes a distributed supervising the network, changes the present situation that the P2P network lacks management.
This method provides autgmentability interface preferably for network management simultaneously, size as each DNA group can decide according to value of utility, the minimum lower bound of a required value of utility of node of promptly average management, as the quantity of an average DNA management node, and cycle of DNA poll etc. all provide some that interface is set.
Description of drawings
Fig. 1 adds the sequential chart of network for node;
The sequential chart that Fig. 2 exits network for node;
Fig. 3 represents is that node adds network development process and DNA chooses flow chart in the network;
What Fig. 4 represented is that node exits network, the network operation process flow diagram;
What Fig. 5 a represented is that ordinary node upgrades to the DNA flow chart, and Fig. 5 b is that DNA is degenerated to the ordinary node flow chart;
Fig. 6 a is an ordinary node abnormality processing module flow chart, and Fig. 6 b is a DNA abnormality processing module flow chart.
Embodiment
With third generation P2P---hybrid P2P network is a background, proposes composite structured P2P network: replace super node by DNA, constitute upper layer network, the distributed network processing capacity is provided; And provide the on-line file agent functionality for lower level node.DNA identity, DNA quantity can be dynamically to adjust, and satisfy the requirement of self-organization of network more.When network size, when network state changes, node state in the network and communicating to connect are each other adjusted, to adapt to the variation of network thereupon accordingly.Its technical scheme mainly comprises following three aspects:
1, composite structured network model:
The first generation all is one deck, i.e. overlay network on the application layer on the second generation P2P cellular logic.Third generation P2P is a kind of hybrid P2P network, and the connection distributivity of full mesh is not high between the super node, the resource retrieval limited efficiency.The present invention proposes a kind of composite structured P2P, and the structured network characteristic is dissolved into the hybrid network structure, and uses structuring to search algorithm and greatly improve network performance.
In order further to improve the performance and the adaptivity of P2P network, need to consider the otherness between the different host nodes in the P2P network.Mainly comprise: the otherness of performance, difference in geographical location.Composite structured network is a two-tier network model completely, the relative performance of node in one side extraction and the comparing cell, allowing the excellent host node of the adjacent performance of physics move the upper layer network function on the other hand, is that the host node of the second layer improves the service of P2P network.
Current third generation P2P network mainly is that the complete distributed network model with the totally centralized network model of the first generation and the second generation merges, and does not fundamentally consider the otherness of node in the network.Its super node is to be served as by fixed server, thereby has limited the adaptivity of network greatly, because fixing super node quantity is dynamically to adjust according to network operation situation, the access of resources in network catalogue can not dynamic assignment.Along with the expansion of network size, the performance of P2P network can only guarantee that such assurance is limited by improving constantly the super node performance.
The present invention proposes distributed network agency (DNA, Distributed Network Agent) and replaces super node that P2P Internet resources issue retrieval service is provided.Its superiority is that it not only has the characteristic such as efficient, fault-tolerant of third generation P2P network, and can be by adjusting host function in the network, according to network operation situation, dynamically adjust network host node quantity at the middle and upper levels, and and second layer host node signal intelligence; And Resource TOC is in the upper layer node distributed treatment because on average the resource of each node is limited, thereby can be fine the performance handled of assurance upper layer network.
In addition, such P2P network model also is to manage, and is controllable.Can divide different territories according to the geographical position; For different territories, control this territory quantity of node at the middle and upper levels according to the quantity of node; Each upper layer node is according to its performance characteristics, and control allows the maximum number of second layer node connection etc.
2, DNA chooses strategy:
Each is added host node in network, can with its performance abstract be a comparable value of utility, big more its performance of explanation of value of utility is good more.Among the present invention, design a kind of more linear value of utility, be about to the performance parameter of multidimensional, be converted into the value of utility of one dimension by linear function.
DNA be by performance in the territory relatively preferably node serve as, carry out processing such as distributed search, storage at upper layer network; For lower floor's network provides the centralized network agent functionality.DNA may become ordinary node under the certain condition, and ordinary node also may be transformed to DNA, corresponding the adjusting along with the variation of network.
At first extract this machine performance parameter, calculate value of utility; Game server is according to DNA quantity, and whether the decision of the ratio of total number of nodes increases new upper layer node.Increase as need, then compare the value of utility of all second layer nodes in this territory, find to have peaked main frame, making it is new DNA, returns new upper layer node tabulation to newly added node.Then directly return the DNA tabulation as not increasing.
This machine is surveyed all upper layer node in this territory according to function identity and upper layer node tabulation that game server returns, and each upper layer node is returned the quantity of institute's connected node, i.e. the size of institute's scope.According to the value of utility of each upper layer node and the ratio of its scope size, select one of the ratio maximum to connect, add network.
If this machine is chosen as DNA, then should open the network function of upper layer node, carry out distributed treatment, improve the distributed network function, if institute's scope unsaturation then the connection request of recipient node simultaneously, the saturated refusal connection message that then returns.
Detailed process as shown in Figure 3.
3, along with network change, the meshed network function adjusts accordingly:
Network change mainly comprises two aspects: mainly being the adding of node on the one hand, mainly is leaving of node on the other hand.The adition process of node also is the process of building of self-organizing P2P network, comprises the adding of DNA---communicate by letter between the DNA, carry out distributed treatment; The adding of lower level node---survey, select the upper strata agent node to connect.The process of leaving of node mainly is the process of network operation, comprises normally withdrawing from of node; Node withdraw from two aspects unusually.
Node changes and mainly to refer to the variation of " node identity ": being promoted by lower level node is DNA, or DNA deteriorates to lower level node, as shown in Figure 5.
Lower level node promotes to DNA, opens the DNA network function accordingly, monitors the connection request of other node; Communicate by letter with other DNA simultaneously, carry out distributed treatment according to the document routing algorithm.The ordinary node lifting mainly contains following two kinds of situations for DNA; Along with the expansion of network size, needing increases DNA quantity, according to the DNA selection algorithm, is elected as DNA again with the node of ordinary node adding network in the past; In addition, when common node connected in the network DNA exits network, in the territory preferably common node be chosen as standby DNA, withdraw from the back at DNA and initiatively upgrade to DNA, the execution corresponding function, other node then reconnects standby DNA.
It is less that DNA deteriorates to the situation of lower level node, the dynamic change of node causes the network upheaval in the network in order to reduce, DNA deteriorates to ordinary node and is controlled by server, the connection rate that connects DNA when ordinary node in the network is less than 50% the time, server can recomputate the DNA quantity that needs in the network, and according to the height of the value of utility of DNA, the not high DNA of notice value of utility withdraws from the DHT network, deteriorate to ordinary node, again obtain the DNA tabulation, survey all DNA, select to connect the highest DNA of average utility value.
Self-organizing P2P is a dynamic network, requires network that better dynamic self-adapting is arranged, and standby DNA will reach such purpose just.If certain DNA exits network, node will connect a node in this territory in this moment this DNA institute scope, be called secondary node; Secondary node is opened the upper layer network function, carries out finishing distributed treatment alternately with DNA.Choosing and safeguarding of secondary node carried out when this territory network condition changes, and as shown in Figure 4, the performance that requires secondary node simultaneously is that performance is more excellent in this territory.
Key characteristic: DNA, distributivity, dynamic.
Carry out distributed treatment between the DNA node, the distributivity requirement well is provided; The choice of dynamical of DNA, the requirement that agreement has well satisfied dynamic such as add, leave.
Two-layer self-organizing peer-to-peer network agreement:
1, land: node adds network, at first sends REQUEST message to game server, and request adds network; OK message is returned in the game server request of receiving, accepts request; Node is received the message that server returns, and sends value of utility to game server, if no message is returned, direct " historical DNA " functional module that connects is carried out in then Connection Service device failure; Server is received value of utility, according to the effectiveness of node, carries out DNA and chooses functional module, and the return node identity message is upgraded the DNA tabulation simultaneously, tabulates to the node return node; Node is received identity message, carries out the corresponding function module; Reception DNA tabulates, and writes down to show in node; At last, send END message to game server; Server is received END message, returns END message, this time sign off.Implementation as shown in Figure 1.
2, withdraw from: node exits network, and at first sends OVER message to game server; Server is received OVER message, withdraws from module according to the node ID XM, returns OVER message to node; Node is received communicating by letter of the OVER end of message and server.Next if current identity is a lower level node, the DNA of Xiang Suolian sends QUIT message; DNA receives QUIT message, carries out corresponding conservation of resources module according to node ID, returns QUIT message; Node is received QUIT message, and end is communicated by letter with DNA's.If current identity is DNA, then at first oneself is hung up, and send DNAQUIT message to each lower level node that is connected, send the information of standby DNA simultaneously; Lower level node is received DNAQUIT message, according to the information of standby DNA, connects secondary node, and end is communicated by letter with DNA's; At last, DNA carries out conservation of resources module, exits network.Implementation such as sequential chart are shown in Figure 2.
3, poll: DNA regularly sends polling message ISALIVING to the lower level node that is connected, if network need be chosen secondary node, will send secondary node information simultaneously; Lower level node is received ISALIVING message, illustrates that DNA is online, returns ALIVE message, proves own onlinely, otherwise that explanation institute connects DNA is not online, and the unusual exit function module of execution DNA is shown in Fig. 6 (b); DNA receives the lower level node online message, illustrate to connect lower level node online, otherwise XM withdraws from module unusually, shown in Fig. 6 (a).
4, issue retrieval:, can directly on distributed hashtable DHT, carry out if present node is DNA; Otherwise need to send FILESHARE or FILESEARCH message, send the file name information of issue or retrieval simultaneously to company DNA; DNA receives FILESHARE or FILESEARCH message, carries out corresponding functional modules, and object information is sent to requesting node; Requesting node receives execution result, and is shown to the user.
5, the change of node identity: node is received APPLY message, if present node is DNA, then returns the size in this territory, i.e. the ordinary node quantity of DNA connection; If present node is an ordinary node, then send the APPLYDNA/ID message authentication and whether chosen the DNA node again to server, ID promptly sends this node ID message to server.Receive that the POSITIVE message of server represents that this node has changed to DNA at server end, send the quantity of the common node that has connected to application node; The NEGTIVE message of receiving server represents that this node identity do not have change, sends-1 to application node, represent that no identity changes, and refuses connection request.
When the size in DNA territory changes, the ordinary node that DNA chooses value of utility maximum in the territory again is standby DNA, and ordinary node sends SECONDDNA/ID message in the territory, node is received this message, it is ID a number of comprising loopback IP address that expression has standby DNA updating message, the ID of the standby DNA that gives tacit consent on the ordinary node.Check the ID of new standby DNA,, when DNA withdraws from or lost efficacy, open the DNA network function, and mainly upgrade to DNA if this node ID then changes to standby DNA; If then do not write down this ID number, when the DNA that is connected lost efficacy or exitted network, reconnect standby DNA for this node ID.
Owing to the continuous adding of node arranged and withdraw from, when being connected rate, crosses when low the too much DNA of DNA number in the network, and server recomputates required DNA quantity in the network, and the DNA low to value of utility sends FALLBACK message.DNA receives this message, reconnects server and obtains up-to-date DNA tabulation, surveys, connects DNA again with the ordinary node identity; Simultaneously up-to-date DNA tabulation is transmitted to ordinary node in the territory, new DNA is surveyed, connected to ordinary node again.
U value calculative strategy
The U value is used for weighing the Main Feature of a computer, and general U value is big more, illustrates that performance is good more.During the computing node value of utility, the computing power parameter of need considering mainly comprises: 1, CPU quantity, 2, CPU speed, 3, the actual memory size, 4, the free memory size, 5 battery statuss, 6 battery electric quantities.The main performance linear expression of computer is come out note with a performance function
X wherein
iBe each performance parameter value; a
iBe the performance impact factor, its size is set according to actual conditions.Consider the difference of actual conditions, such as the difference of notebook, desktop computer, the difference of monokaryon, multinuclear, the performance impact factor of selection should be different, and classification is discussed below:
1, be divided into according to CPU quantity: monokaryon, two kinds of situations of multinuclear:
Order
Significantly the performance of multinuclear node is better than monokaryon, satisfies actual situation.
In addition, when handling the factor of influence of CPU speed and internal memory, done a main balance: by experiment, can know that when cpu type was " AMD 2000+ ", CPU speed was fixed value: 1262; When cpu type was " Intel Pentium4 ", CPU speed perseverance was 1999.Can see, CPU speed is the appearance (do not get rid of the CPU speed that exceeds this scope certainly, this allows) between 1000 to 3000 substantially, and the calculator memory size, be generally speaking at present at 256M between the 2048M, both are not on the same order of magnitude.At first calculator memory is handled, made that the memory size parameter is unit with the Mbytes; Because in the ordinary course of things, CPU speed can directly reflect the performance characteristics of CPU, so at both proportion degree that influences to computing power, we make factor of influence be respectively 0.618 and 0.382 again.For the situation of monokaryon, additionally consider the situation of free memory again, be calculated value can reflect computer use situation, also effectively distinguish the otherness of monokaryon, multinuclear.Through experiment, the setting of above-mentioned factor of influence can be good at reflecting the different performance of different main frames, can be used as a feasibility criterion of weighing host performance.
2, according to the difference of computer kind, at first will divide into notebook, desktop computer and other handheld devices etc., on the one hand the notebook performance may be a little less than the desktop computer of same parameter, and because it is portable, the line duration instability; Notebook has battery as guarantee on the other hand, can effectively reduce withdrawing from unusually that some power down causes.Can be from battery status, battery electric quantity waits and judges node is notebook or desktop computer.
Table 1, battery status classification chart
| Power supply status | Battery status | Battery capacity | Battery remaining power | |
| Note- |
1 | 1 | 100 | -1 |
| Monitor- |
1 | 128 | 255 | -1 |
As shown in table 1, wherein notebook computer is the situation that is inserted with power supply.Can judge from battery status, battery capacity two aspects is desktop computer or other equipment.
Consider the equipment situation that some are unknown, adopt the correction of the U value that previous step is calculated here, even U=g (U).When judging node and be desktop computer, U is carried out certain weighting, as make U increase certain ratio, general U will increase; When judging node and be notebook or other handheld devices, also U is carried out certain weighting, but weighting at this moment need take into full account the situation of battery electric quantity, as the situation that makes U reduce by battery electric quantity, reduce certain ratio, when general as battery electric quantity was not enough, U can reduce, and the electric weight abundance is considered increase U when AC power is arranged; When can not judging node type, or during battery electric quantity, U is constant generally speaking.So far, can obtain the criterion U value that can reflect host performance and classification difference.
At last, because except that desktop computer, the battery status of most of nodes, when battery electric quantity can be surveyed, withdraw from order to reduce the application exception that causes because of running down of battery, we regularly detect battery status, electric weight, when battery status is " electric weight is low " when electric weight less than 30% (through experimental analysis, this moment), display alarm information; When battery status be " electric weight exhausts substantially " (through experimental analysis, this moment the electric weight instability; Make electric weight be lower than at 9% o'clock, carry out identical operations) time, need initiatively withdraw from application program, make network change smoother.
Claims (1)
1. the self-organizing method based on composite structured peer-to-peer network is characterized in that this method comprises the steps;
Step 1. distributed network management is acted on behalf of choosing of (DNA) and is connected
A node adds network, according to network operation situation, is defined as DNA or ordinary node, and node adds network and is divided into:
1) at first extracts this machine performance parameter, calculate this machine value of utility according to U value calculative strategy;
2) game server is according to DNA quantity, and whether the decision of the ratio of total number of nodes increases new upper layer node; If total node number is S, stipulate that a DNA allows to connect N ordinary node at most, then required DNA quantity is the supremum of S/N in the network; The record of online DNA quantity by checking server end as can be known in the network; Greater than online DNA quantity, then compare the value of utility of all second layer nodes in this territory as required DNA quantity, find to have peaked main frame, making it is new DNA quantity, upgrades the DNA tabulation;
3) choose the result according to DNA and determine the node identity, and return new DNA tabulation to newly added node;
4) be ordinary node as this machine, then, survey the size of all the DNA institute scopes in this territory, calculate the value of utility of each DNA and the ratio of its scope size then according to the DNA tabulation, select a DNA of ratio maximum to send connection request, disconnection is communicated by letter with other DNA's;
5), then, join the DHT network that other DNA node forms, and open upper strata distributed network storage and processing capacity, lower floor's centralized network agent functionality according to the DNA tabulation if this machine is DNA;
Step 2, network node are dynamically adjusted strategy
The P2P network is the network of a dynamic change, according to the situation of the network operation, nodal function in the network is dynamically adjusted, and realizes the self-organizing characteristic of network;
1) ordinary node is changed to DNA: there have new node to add in network to be fashionable, needs to increase the DNA node, carries out DNA and choose strategy, and the node that is selected will be transformed to new DNA node by ordinary node; Server joins the DNA that newly chooses in the DNA tabulation, is convenient to ordinary node and sends probe message.New DNA receives probe requests thereby, will confirm with server, confirms successfully will open the DNA functional module;
2) standby DNA upgrades to DNA: when a DNA territory interior nodes situation changes, DNA will choose standby DNA again; DNA makes that the ordinary node of online value of utility maximum in the territory is standby DNA, when having node to add in the territory or withdrawing from, the ordinary node of value of utility maximum in the territory will be searched again, making it is new standby DNA, when DNA exits network or takes place when unusual, standby DNA will be automatically upgraded to DNA, send the DNA upgrading message to server, open the DNA functional module; Other ordinary node will directly send connection request to standby DNA in the territory;
3) DNA deteriorates to ordinary node: when network size reduces, ordinary node connects the connection rate of DNA less than 50% in the network, be that ordinary node quantity/DNA quantity is during less than N/2, server can recomputate the DNA quantity that needs in the network, and according to the height of the value of utility of DNA, the not high DNA of notice value of utility withdraws from the DHT network, deteriorate to ordinary node, again obtain the DNA tabulation, survey all DNA, select to connect the highest DNA of average utility value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910026197A CN101521679A (en) | 2009-04-03 | 2009-04-03 | Self-organizing method based on composite structured peer-to-peer network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN200910026197A CN101521679A (en) | 2009-04-03 | 2009-04-03 | Self-organizing method based on composite structured peer-to-peer network |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101521679A true CN101521679A (en) | 2009-09-02 |
Family
ID=41082050
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200910026197A Pending CN101521679A (en) | 2009-04-03 | 2009-04-03 | Self-organizing method based on composite structured peer-to-peer network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101521679A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013060221A1 (en) * | 2011-10-28 | 2013-05-02 | 中国移动通信集团公司 | Node addition method for p2p system, corresponding p2p system, peer network node degradation method and system, and peer network node |
| CN103095744A (en) * | 2011-10-28 | 2013-05-08 | 中国移动通信集团公司 | Method and system of peer-to-peer network node degradation and peer-to-peer network node |
| CN103227817A (en) * | 2013-03-28 | 2013-07-31 | 广东省电信规划设计院有限公司 | Super node selecting method and system based on P2P (peer-to-peer) network |
| CN103327058A (en) * | 2012-03-23 | 2013-09-25 | 财团法人工业技术研究院 | Method and system for adjusting tree coverage network of peer-to-peer streaming network |
| CN104283948A (en) * | 2014-09-26 | 2015-01-14 | 东软集团股份有限公司 | Server cluster system and load balancing implementation method thereof |
| CN114205363A (en) * | 2021-12-01 | 2022-03-18 | 浪潮云信息技术股份公司 | Cluster management method of distributed database and cluster of distributed management system |
-
2009
- 2009-04-03 CN CN200910026197A patent/CN101521679A/en active Pending
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2013060221A1 (en) * | 2011-10-28 | 2013-05-02 | 中国移动通信集团公司 | Node addition method for p2p system, corresponding p2p system, peer network node degradation method and system, and peer network node |
| CN103095744A (en) * | 2011-10-28 | 2013-05-08 | 中国移动通信集团公司 | Method and system of peer-to-peer network node degradation and peer-to-peer network node |
| CN103095744B (en) * | 2011-10-28 | 2016-05-25 | 中国移动通信集团公司 | Peer network node down method and system, peer network node |
| CN103327058A (en) * | 2012-03-23 | 2013-09-25 | 财团法人工业技术研究院 | Method and system for adjusting tree coverage network of peer-to-peer streaming network |
| CN103327058B (en) * | 2012-03-23 | 2016-05-04 | 财团法人工业技术研究院 | Method and system for adjusting tree coverage network of peer-to-peer streaming network |
| CN103227817A (en) * | 2013-03-28 | 2013-07-31 | 广东省电信规划设计院有限公司 | Super node selecting method and system based on P2P (peer-to-peer) network |
| CN103227817B (en) * | 2013-03-28 | 2016-08-17 | 广东省电信规划设计院有限公司 | P2P super node of network system of selection and system |
| CN104283948A (en) * | 2014-09-26 | 2015-01-14 | 东软集团股份有限公司 | Server cluster system and load balancing implementation method thereof |
| CN104283948B (en) * | 2014-09-26 | 2018-12-07 | 东软集团股份有限公司 | Server cluster system and its implementation of load balancing |
| CN114205363A (en) * | 2021-12-01 | 2022-03-18 | 浪潮云信息技术股份公司 | Cluster management method of distributed database and cluster of distributed management system |
| CN114205363B (en) * | 2021-12-01 | 2024-01-26 | 上海沄熹科技有限公司 | Cluster management method of distributed database and distributed management system cluster |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Hao et al. | Deep reinforcement learning for edge service placement in softwarized industrial cyber-physical system | |
| US7543020B2 (en) | Distributed client services based on execution of service attributes and data attributes by multiple nodes in resource groups | |
| Xiao et al. | Dynamic layer management in superpeer architectures | |
| Zhao et al. | Dynatops: A dynamic topic-based publish/subscribe architecture | |
| CN101521679A (en) | Self-organizing method based on composite structured peer-to-peer network | |
| Li | Data transmission scheme considering node failure for blockchain | |
| CN110138732A (en) | Response method, device, equipment and the storage medium of access request | |
| Wang et al. | Hierarchical cooperative caching in mobile opportunistic social networks | |
| CN108600316B (en) | Data management method, system and equipment of cloud storage service | |
| Lin et al. | Research on resource self-organizing model for cloud computing | |
| CN102067563A (en) | Methods and apparatus for optimal participation of devices in a peer-to-peer overlay network | |
| Alsenani et al. | SaRa: A stochastic model to estimate reliability of edge resources in volunteer cloud | |
| Achim et al. | Reputation based selection for services in cloud environments | |
| Fan et al. | Game balanced multi-factor multicast routing in sensor grid networks | |
| Zhang et al. | Energy minimization task offloading mechanism with edge-cloud collaboration in IoT networks | |
| Cheng et al. | SEEDS: A solar-based energy-efficient distributed server farm | |
| Eymann et al. | Self-organizing resource allocation for autonomic network | |
| Jiang et al. | Economics of peer-to-peer mobile crowdsensing | |
| CN113965587A (en) | Data acquisition method, device, equipment and medium of artificial intelligence platform | |
| Huang et al. | Power-aware hierarchical scheduling with respect to resource intermittence in wireless grids | |
| Sharifi et al. | Combing smart grid with community clouds: Next generation integrated service platform | |
| CN107612980B (en) | Adjustable and reliable consistency maintenance method in structured P2P network | |
| TW202327380A (en) | Method and system for federated reinforcement learning based offloading optimization in edge computing | |
| Caporuscio et al. | Decentralized architecture for energy-aware service assembly | |
| Zhang et al. | Adaptive Digital Twin Server Deployment for Dynamic Edge Networks in IoT System |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Open date: 20090902 |