CN101184016A - Universal resource management method under confusion type peer-to-peer network environment - Google Patents
Universal resource management method under confusion type peer-to-peer network environment Download PDFInfo
- Publication number
- CN101184016A CN101184016A CNA2007101687228A CN200710168722A CN101184016A CN 101184016 A CN101184016 A CN 101184016A CN A2007101687228 A CNA2007101687228 A CN A2007101687228A CN 200710168722 A CN200710168722 A CN 200710168722A CN 101184016 A CN101184016 A CN 101184016A
- Authority
- CN
- China
- Prior art keywords
- node
- resource
- strong
- peer
- over
- 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
Images
Landscapes
- Computer And Data Communications (AREA)
Abstract
The invention discloses a general resource management method in a hybrid peer-to-peer network environment, comprising a plurality of steps that: (1) all the resources of a node P are described uniformly by the node P, the internal resource of the higher-level service corresponds to the description of the general resource; (2) the node P is added into a hybrid topology, then the node is determined as a role of a strong node or a weak node; (3) resource releasing, resource searching and resource deleting are respectively carried out according to the different management operations. The invention has the advantages that: the resource management opens to a plurality of services based on peer-to-peer network technology due to the standard definition of the general description for the resource; in addition, the effective management of the uniform resource is realized in the dynamic network environment by the construction of a peer-to-peer network environment with the mutual mixture of distributed Hash table and delayed aggregation multi-analytical annular topology; a plurality of services are supported by using open resource management method, so the waste of network bandwidth and node calculation ability is reduced, moreover, the dynamic property of the system and the effectiveness of resource searching is guaranteed.
Description
Technical field
The invention belongs to computer application field, is the universal resource management method in a kind of confusion type peer to peer environment.
Background technology
The peer-to-peer network technology is that of rising in recent years utilizes terminal desktop PC computing capability and terminal network bandwidth to carry out the new technology of file-sharing between the user, it has plurality of advantages such as using flexibility, extensibility, robustness, load balancing, so universal a kind of good technical basis that provides of some network multimedia services such as streaming media service, voice service also is provided simultaneously for its.Can share existing data mutually and need not all focus on the single server end based on the node of peer-to-peer network and obtain data, thereby alleviate the burden of server and network greatly.
In the service based on the peer-to-peer network technology, data are no longer obtained from single server, and the user need adopt distributed method to safeguard own resource and search for and want the resource obtained in network.At present, carry out the conservation of resources and search mainly contains two kinds of methods by the peer-to-peer network technology: a kind of is to utilize distributed hashtable (Distributed Hash Table) to have to organize user node and resources shared structurally, carries out the issue and the search of resource then by hash algorithm; Another kind is the method that adopts inundation (Flooding) request, need be with certain structure organization node, and near the node the user utilizes in network is with forwards, and the diffusing resources search message replies to the requestor after the node that has resource is received message.All there is the shortcoming of oneself in these two kinds of methods: the expense of the conservation of resources is very big in the distributed hash table method, does not support the search of multikey value, and is subjected to the influence of network dynamic easily; The inundation request can produce bigger network traffics in the resource searching process, for the search poor effect of scarce resource.The peer-to-peer network of confusion type then is a kind of compromise method, node in the network is divided into weak node and strong two levels of node, some weak nodes of strong node administration adopt the distributed hash table method to safeguard and searching resource between strong node, then adopt the method for inundation request between weak node.
At present, adopt the method for confusion type based on the service of peer-to-peer network technology more, but these services itself all be sealing with independently, there are a large amount of resource and operating procedures that repeat each other, therefore, for the conservation of resources itself, caused the waste of the network bandwidth and terminal node computing capability.
Summary of the invention
The object of the present invention is to provide the universal resource management method under a kind of confusion type peer to peer environment, this method utilizes open method for managing resource to support multiple service, reduce the waste of the network bandwidth and node computing capability, and guaranteed the dynamic of system and the validity of resource searching.
Universal resource management method under the confusion type peer to peer environment provided by the invention, its step comprises:
(1) node P unifies to describe to all resources of self, the internal resource of top service is corresponded to generic resource describe;
(2) node P joins in the peer-to-peer network according to following flow process:
(2.1) initialization, acquiescence node P is weak node; Node P connects the startup server, and obtains the road sign node listing, and the road sign node is the particular server that is arranged in different regions in the network in advance, is to be used as the reference point of setting up the network virtual coordinate system;
(2.2) network speed between each road sign node in measured node P and the above-mentioned road sign node listing finds nearest road sign node LM[M], to road sign node LM[M] ask the tabulation less than the node composition of the threshold value of default from node P, as strong node listing; Postpone to be less than the strong node that postpones threshold value if exist in this tabulation, enter step (2.3), otherwise change step (2.7) over to from node P;
(2.3) number that postpones to be less than the strong node that postpones threshold value from node P in the strong node listing that obtains of node P equals or during more than three, enters step (2.4), otherwise enter step (2.6);
(2.4) node P is connected by force with setting up with its nearest strong node S, if the weak interstitial content of institute's carry is smaller or equal to the higher limit of system's appointment under the node S, then node P sets up TCP with node S and links by force, and choose two time near strong Node B 1, B2 is as the strong node of reserve of oneself, communication mode with UDP is kept in touch, and enters step (3), otherwise enters step (2.5);
(2.5) when interstitial content reaches above-mentioned higher limit a little less than the carry of node S, node S notifies node P or certain carry weak node Q under S to change strong node at random, enters step (2.7);
(2.6) bandwidth, the disposal ability of inspection node P or Q if satisfy the requirement that becomes strong node, enter step (2.7), otherwise change step (3) over to;
(2.7) promoting node P or Q is strong node, this node is set up concordance list, this concordance list comprises the doubly linked list of identifier Hash table and time-sequencing, a distributed hashtable is safeguarded in this node and other each strong node cooperation, adds many parsings ring structure topology that postpones to assemble of other strong node maintenances simultaneously;
(3) node P carries out resource management according to the bookkeeping difference according to following flow process:
(3.1) decision operation type is if issue/deletion resource enters step (3.2); If searching resource enters step (3.8);
(3.2) whether decision node P is strong node, if enter step (3.3); Otherwise enter step (3.5);
(3.3) node P sets up/deletes the identifier resource index entry in concordance list;
(3.4) node P issues/deletes resource in distributed hashtable, changes step (3.17) then over to;
(3.5) node P sends issue/deletion resource request to its strong node S of agency;
(3.6) after node S receives issue/removal request of node P, foundation/deletion identifier resource index entry in concordance list;
(3.7) node S issues/deletes resource in distributed hashtable, changes step (3.21) then over to;
(3.8) whether decision node P is strong node, if enter step (3.9), otherwise change step (3.13) over to;
(3.9) node P searches in concordance list, obtains Search Results;
(3.10) node P waits for the response message of other nodes that have this resource postponing to spread search on many parsings ring structure topology of assembling, and the stand-by period is set; Node P searches in distributed hashtable, has the response message of other nodes of this resource by the time;
(3.11) when node P receives response message, enter step (3.17), otherwise change step (3.12) over to;
(3.12) when the node P stand-by period uses up, enter step (3.17), otherwise change step (3.11) over to;
(3.13) node P sends the searching resource request to its agent node S; Node P enters the Search Results state of waiting for; Node S searches in concordance list, obtains Search Results, and the result is told to node P;
(3.14) node S waits for the response message of other nodes that have this resource postponing to spread search on many parsings ring structure topology of assembling; Node S searches in distributed hashtable, has the response message of other nodes of this resource by the time;
(3.15) when node S receives response message, tell node P, enter step (3.17), otherwise change step (3.16) over to;
(3.16) when the node S stand-by period uses up, tell node P, enter step (3.17), otherwise change step (3.15) over to;
(3.17) according to user's input, need to judge whether the ending resource management, then finish if desired, otherwise, return step (3.1) and carry out next step resource management action.
The present invention makes the management of resource open to multiple service based on the peer-to-peer network technology by resource is carried out the definition of universal description standard.In addition, the present invention is by making up distributed hashtable and the peer to peer environment that the many parsings ring topology that postpones gathering mixes mutually, is implemented under the dynamic network environment the effective management to unified resource.Particularly, the present invention has following characteristics:
(1) opening of resource management
No matter be to be applied to which kind of service, the resource of being managed among the present invention is abstract resource description, and based on these descriptions, comprise that result that bookkeepings such as issue, search, deletion produce is abstract to resource, service is inner to obtain serving inner resource itself to these abstract reverse resolutions that carry out again.Therefore, different services does not need to understand the inner resource of other services, safeguards but follow same unified the description, thereby has guaranteed the opening of resource management.
(2) dynamic
In the confusion type peer to peer environment, weak node can keep strong annexation with the strong node of management of oneself, can choose two strong nodes of reserve simultaneously and keep weak connection.If weak node leaves system, the strong strong node that connects can know that connection disconnects and the identity Dynamic Maintenance resource change situation to act on behalf of.Conversely, if strong node leaves, so weak node will be chosen a strong node of reserve as the manager after learning the connection disconnection, and sets up strong the connection.Strong node self also can be in postponing many parsings ring topology of assembling regularly and other the most contiguous strong node intercommunication node survival condition, if the situation that strong node leaves, other strong nodes can upgrade the ring topology of oneself, and no longer are sent to the node of having left away when the diffusing resources administrative messag.
(3) validity
The definition that generic resource is described is open, therefore can support the resource type of all services with expanding.In addition, by the formation of confusion type peer-to-peer network and the local resource tissue of strong node, three level resource administrations have been realized: the concordance list management, the management of many parsings ring topology and the distributed hashtable that postpone to assemble are managed, avoided search to invalid resource, avoided Hash search, avoided inundation search, and guaranteed the validity of resource management scarce resource to hot resource.
Description of drawings
Fig. 1 is the main flow chart of the inventive method
Fig. 2 is the unified flow process of describing of the generic resource of the inventive method;
Fig. 3 is the confusion type topology schematic diagram of the inventive method;
Fig. 4 is that the node of the inventive method adds flow process;
Fig. 5 is the strong node resource concordance list organization chart of the inventive method;
Fig. 6 is the resource management flow process of the inventive method;
Fig. 7 is the schematic diagram of the resource issue of the inventive method;
Fig. 8 is the schematic diagram of the resource deletion of the inventive method;
Fig. 9 is the schematic diagram of the strong node searching resource of the inventive method;
Figure 10 is the schematic diagram of the weak node searching resource of the inventive method.
Embodiment
The present invention is directed to the total resource management demand of multiple service in the peer-to-peer network, proposed the universal resource management method under the specific admixture network environment, the present invention is done detailed explanation below in conjunction with accompanying drawing and example.
As shown in Figure 1 be the main flow chart of any node P in the our bright method, node P must pass through following step if carry out universal resource management under the peer to peer environment of chamber mixing:
(1) P must earlier unify to describe (Fig. 2) to all resources of self, the internal resource of top service is corresponded to generic resource describe;
(2) add flow process (Fig. 4) by node, P need add the confusion type topology (Fig. 3) as the entire method lower support.After adding the confusion type topology, node P determines own role as the still weak node of strong node, enters step (3) afterwards again;
(3) this step then is resource management flow process (Fig. 6).The operation of universal resource management has resource issue, resource searching and resource to delete three classes respectively.
Fig. 2 is exactly that node P describes flow process to the unified resource that internal resource carries out.It is by extracting the various features of internal resource, the title, content, amount of capacity, the type of affiliated service, the identifier that comprise resource correspond to URL(uniform resource locator), resource name, resource type, resource content, content size, content Hash value and the resource tag several sections that generic resource is described by combinatorial operation again.
Concrete implementation step is as follows:
(1.1) node P selects a local resource, reads the title of this resource, the type of affiliated service and amount of capacity;
(1.2) if this resource capacity size greater than the threshold values (50 Mbytes) of default, enters step (1.3), otherwise changes step (1.4) over to; This threshold values obtains by continuous experiment, and one can make overhead and resource characteristic extract the best empirical value that accuracy reaches balance;
(1.3) node P reads size in the resource at random and is the partial content of default threshold values, enters step (1.5);
(1.4) node P reads the full content of resource;
(1.5) node P carries out Hash operation to the content of reading, and obtains cryptographic Hash;
(1.6) identifier that different services are added on resource in the node P read-out system;
(1.7) service and the cryptographic Hash that belong to according to resource of node P generates the unified finger URL (URL) of resource;
With a certain file-sharing service is example, and the unified process of describing is described.URL(uniform resource locator) is defined as " file-sharing (COS): // address: port/filename ", can only issue at node P under the prerequisite of different title resources, and finger URL can guarantee the uniqueness of global resource.Resource name is exactly a filename, the resource service type is exactly file-sharing, resource content is a file at the threshold values (50 Mbytes) of resource capacity with interior content, content size is a file size, content Hash value is that resource content is carried out result after the Hash operation, and resource identifier is the label of file.
As shown in Figure 3, the confusion type peer-to-peer network of forming based on distributed hashtable and the many parsings ring structure that postpones to assemble has constituted the supporting network of universal resource management.The terminal node of forming this network is divided into two kinds, and a kind of is strong node, as the node of numbering 1 to 6 among the figure; Another kind is weak node, as the node of numbering 7 to 9 among the figure.Weak node and strong node are that computing capability, the network bandwidth according to terminal comes and line duration is divided, and purpose is in order to allow more capable more stable node bear more responsibility, to increase stability of network.Therefore, the responsibility of strong node is the weak node of a management part, and the weak node of agency is exercised resource management action.In concrete enforcement, we are provided with computing capability and surpass 2GHZ, and bandwidth surpasses the node that 1Mbps and average online hours reach 2 hours and can be used as strong node.Among the figure, weak node 7,8,9 is chosen from them and is postponed the agency of minimum strong node 6 as resource management action, and has set up strong the connection with 6.In addition, weak node also can be chosen other and postpone two minimum strong nodes node in support, and keeps weak connection, has chosen strong node 2 and 5 as node 9 and has been reserve.
Be made into a big ring by the distributed hashtable algorithm groups between the strong node, node issued among index hash all nodes on encircling of resource, and can accurately be searched according to hash algorithm.Meanwhile, strong node has also been safeguarded many parsings ring structure topology that postpones to assemble separately, and other the strong nodes around being used for safeguarding also are divided in their differences according to network delay in the different endless belt.(can referring to patent application " a kind of game platform system ", application number 200710052729.3) based on reciprocity overlay network.As shown in Figure 3, node 5 is just in the endless belt of node 4.In concrete the enforcement, we are arranged to the delay of endless belt border representative respectively: 15ms, 50ms, 100ms and 300ms (the delay threshold value on these endless belt borders can collective fluctuate in an acceptable scope, but will guarantee that it increases successively).
And organize a confusion type peer to peer environment as shown in Figure 3, and need node P according to the whole topology of the first adding of flow process as shown in Figure 4, concrete implementation step is as follows:
(2.1) initialization, acquiescence P is weak node;
(2.2) node P connects the startup server, and obtains the road sign node listing; Road sign node (Landmark) is the particular server that is arranged in different regions in the network in advance, is to be used as the reference point of setting up the network virtual coordinate system; Any one node can by with they test the speed, obtaining one is the virtual coordinates of coordinate figure with the network delay, and measures the position of this node in network with this coordinate.Suppose to have three coordinate points A, B, C, if the network delay of P and they is respectively 10 milliseconds, 20 milliseconds, 5 milliseconds so, the network virtual coordinate of P is exactly (10,20,5) so.If the road sign interstitial content is many more, the accuracy of measurement of this virtual coordinates is also just high more, but considers the cost and the coordinate complexity of road sign node, and the scope of actual deployment should be between 3 to 7.
(2.3) network speed between each road sign node in measured node P and the above-mentioned road sign node listing, find nearest road sign node LM[M], to road sign node LM[M] ask from the strong node listing of node P less than the road sign threshold value (its span is generally 300～400 milliseconds) of default; Postpone to be less than the strong node that postpones threshold value (as being set at 50 milliseconds) if exist in this tabulation, enter step (2.4), otherwise change step (2.8) over to from node P;
(2.4) number that postpones to be less than the strong node that postpones threshold value from node P in the node P tabulation of obtaining equals or during more than three, enters step (2.5), otherwise enter step (2.7);
(2.5) node P sets up with nearest strong node S and is connected by force, if the weak interstitial content of institute's carry (is system's designated value smaller or equal to the upper limit under the node S, be at least 20), then P sets up TCP with S and is connected by force, choose near strong Node B 1 simultaneously two times, B2 keeps in touch with the communication mode of UDP as the strong node of reserve of oneself, enter step (3), otherwise enter step (2.6); Selecting two backup node is in order to prevent S to vamose, to cause the operation upheaval of node P.
(2.6) the weak interstitial content that is connected as node S has reached the upper limit, and the weak node Q that node P or certain are connected under the S can notifiedly at random change strong node into, enters step (2.7);
(2.7) check bandwidth, the disposal ability of node P or Q,, enter step (2.8), otherwise change step (3) over to if satisfy the requirement that becomes strong node (as bandwidth more than or equal to 1 mbit/, and disposal ability is more than or equal to 2 GHzs);
(2.8) promoting node P or Q is strong node, this node is set up concordance list, this concordance list comprises the doubly linked list of identifier Hash table and time-sequencing, a distributed hashtable is safeguarded in this node and other each strong node cooperation, adds many parsings ring structure topology that postpones to assemble of other strong node maintenances simultaneously;
Like this, just can set up based on the distributed hashtable and the confusion type peer-to-peer network of the many parsings ring structure topology that postpones to assemble.
Strong node is the succedaneum of resource management action, it is managing the resource information of the weak node of having jurisdiction over, and support to issue, the resource operation of deletion and multiple search, therefore how to carry out index and managerial demand and depend on as shown in Figure 5 data structure for the resource information of weak node.The concordance list of strong node depends on the identifier of unified resource description and sets up.At first the various identifiers of resource are put into a Hash table, the identifier 1 among the identifier Xiang Rutu of hash-collision, identifier 2 and identifier 3 usefulness slide fastener methods solve, and the resource index item that will have certain identifier is put into the vertical tabulation of this identifier.Resource index resource description pointed safeguards by the doubly linked list of a time correlation that in addition the resource that is updated can be put into begin chain, like this, the resource of last-of-chain be exactly the time at most, if overtime situation takes place, just can remove forward from afterbody.
The flow chart of resource management according to the bookkeeping difference, is divided into resource issue, deletion and three sub-processes of search as shown in Figure 6.And issue and delete two operations owing to be corresponding fully, so flow process is about the same.Strong node at first in the resource index table of this locality the identifier according to operated resource add resource index (issue resource) or remove resource index (deletion resource), and then in distributed hashtable, the resource index on the distant-end node is added or deletion action accordingly.Weak node then sends the method for request by the strong node of agency to oneself, finish the corresponding operating that the local resource concordance list is set up/removed and distributed hashtable adds/removes successively by the strong node of agency.It is as follows that node P carries out the formalized description of resource management:
(3.1) decision operation type is if issue/deletion resource enters step (3.2); If searching resource enters step (3.8);
(3.2) whether decision node P is strong node, if enter step (3.3); Otherwise enter step (3.5);
(3.3) node P sets up/deletes the identifier resource index entry in concordance list;
(3.4) node P issues/deletes resource in distributed hashtable, changes step (3.21) then over to;
(3.5) node P sends issue/deletion resource request to its strong node S of agency;
(3.6) after node S receives issue/removal request of node P, foundation/deletion identifier resource index entry in concordance list;
(3.7) node S issues/deletes resource in distributed hashtable, changes step (3.21) then over to;
(3.8) whether decision node P is strong node, if enter step (3.9), otherwise change step (3.13) over to;
(3.9) node P searches in concordance list, obtains Search Results;
(3.10) node P spreads search on many parsings ring structure topology of postpone assembling, and waits for the response message of other nodes that have this resource, be provided with the stand-by period (for system according to demand with the numerical value below 30 seconds of ability to bear decision); Node P searches in distributed hashtable, has the response message of other nodes of this resource by the time;
(3.11) when node P receives response message, enter step (3.17), otherwise change step (3.12) over to;
(3.12) when the node P stand-by period uses up, enter step (3.17), otherwise change step (3.11) over to;
(3.13) node P sends the searching resource request to its agent node S; Node P enters the Search Results state of waiting for; Node S searches in concordance list, obtains Search Results, and the result is told to node P;
(3.14) node S waits for the response message of other nodes that have this resource postponing to spread search on many parsings ring structure topology of assembling; Node S searches in distributed hashtable, has the response message of other nodes of this resource by the time;
(3.15) when node S receives response message, tell node P, enter step (3.17), otherwise change step (3.16) over to;
(3.16) when the node S stand-by period uses up, tell node P, enter step (3.17), otherwise change step (3.15) over to;
(3.17) according to user's input, need to judge whether the ending resource management, then finish if desired, otherwise, return step (3.1) and carry out next step resource management action.
Fig. 7 and Fig. 8 are respectively the schematic diagrames of resource issue and resource deletion.Node 1 to 5 is strong nodes among the figure, and all the other are weak node, and the thick line that among the figure node 1 to 5 is coupled together has been represented distributed hashtable, then is the weak node of its management with all nodes in the cloud that strong node is connected.The process of strong node 4 issue resources had only for two steps among Fig. 7: issue resource in the local resource concordance list, issue resource in distributed hashtable.The process of weak node 7 issue resources had three steps: act on behalf of strong node 3 request issue resources to it, node 3 is issued resource earlier in the local resource concordance list, issue resource again in distributed hashtable.Generic resource delete procedure shown in Figure 8 is almost consistent with issuing process.The process of strong node 4 deletion resources had two steps: delete resource in the local resource concordance list, delete resource in distributed hashtable.The process of weak node 7 issue resources has three cloth: to the strong node 3 request deletion resources of its connection, node 3 is deleted resource earlier in the local resource concordance list, delete resource again in distributed hashtable.
Shown in Figure 9 is the schematic diagram that strong node carries out distributed search, and wherein node 1 to 5 is strong nodes, and node 12,13,14,35 is mutual on the other side's ring structure topology, and all strong nodes are all in distributed hashtable.Node 1 is the promoter of searching resource, in search procedure its behavior as follows, this promptly is the detailed implementation of step 3.10 and 3.11:
(A1) node 12,3,4 three strong nodes transmission resource searchings request towards periphery, the life span of searching request bag is 5;
(A2) node 2 is received the resource searching request, and finding does not have other strong node neighbours except node 1, does not continue the searching request of big vast model node 1; The own local Hash table of node 2 search is found the resource that oneself do not have node 1 to search for, does not return any information;
(A3) node 3 is received the resource searching request, and the life span of searching request bag is subtracted 1, becomes 4, continues to send this searching request bag to node 5; The own local resource index table of node 3 search is found the resource that oneself do not have node 1 to search for, does not return any information;
(A4) node 4 dispositions are identical with (A2);
(A5) node 5 is received the resource searching request, and finding oneself does not have other strong node neighbours except node 3, does not continue the searching request of inundation node 1; The own local resource index table of node 5 search is found the resource that node 1 will be searched for, and returns Search Results to node 1.
(A6) in distributed hashtable, node 1 is initiated searching request by algorithm to node 3 by the Hash record, and the resource that node 3 no nodes 1 will be searched for continues to propagate searching request by algorithm to node 5 based on the Hash record;
(A7) node 5 is received searching request, directly returns the resource that will search for to node 1.
Shown in Figure 10 then is the distributed process that weak node is initiated searching request, and wherein node the 1,2,3,4, the 5th, and strong node, node 6 are weak nodes of node 1 agency, and node 12,13,14,35 is mutually on the other side's loop network.The process of weak node 6 searching resources is as follows, and this promptly is the detailed implementation of (3.17) and (3.18):
(B1) node 6 sends searching request to node 1, and the life span of searching request bag is 5;
(B2) node 1 subtracts 1 to the life span of searching request bag, becomes 4, sends the resource searching request to node 2,3,4; The own local resource index table of node 1 search is found the resource that oneself do not have node 6 to search for, does not return any information;
(B3) node 2 is received the resource searching request, and finding does not have other strong node neighbours except node 1, does not continue the searching request of big vast model node 6; The own local resource index table of node 2 search is found the resource that oneself do not have node 6 to search for, does not return any information;
(B4) node 3 is received the resource searching request, and the life span of searching request bag is subtracted 1, becomes 4, continues to send this searching request bag to node 5; The own local resource index table of node 3 search is found the resource that oneself do not have node 6 to search for, does not return any information;
(B5) node 4 dispositions are identical with (2);
(B6) node 5 is received the resource searching request, and finding oneself does not have other strong node neighbours except node 3, does not continue the searching request of big vast model node 6; The own local resource index table of node 5 search is found the resource that node 6 will be searched for, and returns Search Results to node 6.
(B7) in distributed hashtable, node 1 is initiated searching request to node 3, and the resource that node 3 no nodes 6 will be searched for continues to propagate searching request to node 5;
(B8) node 5 is received searching request, directly returns the resource that will search for to node 6.
Example:
Utilize the universal resource management method under the confusion type peer to peer environment set forth in the present invention, the laboratory provides 1 to start server, three road sign servers, and the logical PC of 10 Daeporis, every PC organizes according to the rule adding topology of Poisson distribution, moves two kinds of services based on the peer-to-peer network technology of instant messaging and file-sharing then on every PC simultaneously.The hardware configuration of PC is as follows:
| Machine name | CPU | Internal memory | Hard disk | The network bandwidth |
| Start server | PIV 2.0G | 2G | 300G | 100M |
| |
PIV 2.0G | 2G | 40G | 100M |
| |
PIV 2.0G | 1G | 200G | 100M |
| |
PIV 2.0G | 1G | 200G | 100M |
| PC 1-10 | PIIII 1.7G | 256M | 40G | 10M |
PC 1-10 is called node 1, node 2 ... node 10.
Each selected in this example constant threshold is as follows: the resource capacity threshold value is 50 Mbytes; The delays of resolving the endless belt border representative of ring structure topology are arranged to 15 milliseconds respectively more, and 50 milliseconds, 100 milliseconds and 300 milliseconds; It is 300 milliseconds that road sign postpones threshold value; Postponing threshold value is 50 milliseconds; Strong node can be limited to 20 on the weak number of nodes of carry; The requirement that becomes strong node be bandwidth more than or equal to 1 mbit/, and disposal ability is more than or equal to 2 GHzs; Stand-by period is 30 seconds.
The foundation of confusion type peer to peer environment the and wherein example operation of universal resource management method is as follows:
(1) sets up the startup server
Start server and carry out the TCP monitoring, and keep the UDP standby condition at 9000 ports at 8000 ports.When starting, road sign server and terminal node all to land the startup server, the started server maintenance table of a road sign server, the ID and the IP address of having write down the road sign server.When client node connected the startup server, starting server can return to client to the road sign nodal information in the table.
(2) set up the road sign server
The road sign server keeps the UDP standby condition at 9100 ports, and connects after startup and start server, the ID and the IP address of road sign server 1,2,3 that started server record.When terminal node starts and by after starting server and obtaining the road sign server address, can test the speed to all road sign servers.The road sign server can be oneself to be the center, safeguard many parsings ring structure topology that postpones to assemble, with the strong node that tested the speed according to delayed maintenance in postponing endless belt, after new terminal tests the speed, the strong node that the road sign node can be sought in endless belt and new node closes on, and reply to newly added node.
(3) adding of terminal node
Certain according to the rule of Poisson distribution, starts terminal node 1 to node 10 constantly, adds the confusion type topology.Node 1 is as the node of first startup, it can obtain road sign server 1,2,3 by starting server earlier, after testing the speed, discovery is nearest from road sign server 2, and the strong node that does not have answer, so promoting automatically as first node in the topology is strong node, and registration in road sign server 2.
The node of back has formed confusion type peer to peer environment as shown in Figure 2 after adding, and node 10 is managed by strong node 5 as weak node in addition.
(4) universal resource management operation
All moved instant messaging and two kinds of services of file-sharing on all nodes simultaneously, the resource of instant messaging service is the subscriber data of service, and the resource of file-sharing then is a file.Definition by the unified resource descriptor is extracted, and the subscriber data that makes communication is according to { user ID, resource name }, { instant messaging, COS }, { pet name, resource identifier }, { instant messaging: // customer location/user ID, URL(uniform resource locator) } mode one to one sets up resource description; The file of file-sharing is then according to { file name, resource name }, { file-sharing, COS }, { file content, resource content }, { file-sharing: // document location/file name, URL(uniform resource locator) }, the mode of { label of file, resource identifier } is set up resource description.
Through repeatedly test, adopt the universal resource management method under the confusion type equity environment that the present invention discussed, can support multiple service based on the peer-to-peer network technology, and can effectively realize the issue, deletion of resource and accurately, search for generally, the dynamic for network has excellent adaptability simultaneously.
Claims (2)
1. the universal resource management method under the confusion type peer to peer environment, its step comprises:
(1) node P unifies to describe to all resources of self, the internal resource of top service is corresponded to generic resource describe;
(2) node P joins in the peer-to-peer network according to following flow process:
(2.1) initialization, acquiescence node P is weak node; Node P connects the startup server, and obtains the road sign node listing, and the road sign node is the particular server that is arranged in different regions in the network in advance, is to be used as the reference point of setting up the network virtual coordinate system;
(2.2) network speed between each road sign node in measured node P and the above-mentioned road sign node listing finds nearest road sign node LM[M], to road sign node LM[M] ask the tabulation less than the node composition of the threshold value of default from node P, as strong node listing; Postpone to be less than the strong node that postpones threshold value if exist in this tabulation, enter step (2.3), otherwise change step (2.7) over to from node P;
(2.3) number that postpones to be less than the strong node that postpones threshold value from node P in the strong node listing that obtains of node P equals or during more than three, enters step (2.4), otherwise enter step (2.6);
(2.4) node P is connected by force with setting up with its nearest strong node S, if the weak interstitial content of institute's carry is smaller or equal to the higher limit of system's appointment under the node S, then node P sets up TCP with node S and links by force, and choose two time near strong Node B 1, B2 is as the strong node of reserve of oneself, communication mode with UDP is kept in touch, and enters step (3), otherwise enters step (2.5);
(2.5) when interstitial content reaches above-mentioned higher limit a little less than the carry of node S, node S notifies node P or certain carry weak node Q under S to change strong node at random, enters step (2.7);
(2.6) bandwidth, the disposal ability of inspection node P or Q if satisfy the requirement that becomes strong node, enter step (2.7), otherwise change step (3) over to;
(2.7) promoting node P or Q is strong node, this node is set up concordance list, this concordance list comprises the doubly linked list of identifier Hash table and time-sequencing, a distributed hashtable is safeguarded in this node and other each strong node cooperation, adds many parsings ring structure topology that postpones to assemble of other strong node maintenances simultaneously;
(3) node P carries out resource management according to the bookkeeping difference according to following flow process:
(3.1) decision operation type is if issue/deletion resource enters step (3.2); If searching resource enters step (3.8);
(3.2) whether decision node P is strong node, if enter step (3.3); Otherwise enter step (3.5);
(3.3) node P sets up/deletes the identifier resource index entry in concordance list;
(3.4) node P issues/deletes resource in distributed hashtable, changes step (3.17) then over to;
(3.5) node P sends issue/deletion resource request to its strong node S of agency;
(3.6) after node S receives issue/removal request of node P, foundation/deletion identifier resource index entry in concordance list;
(3.7) node S issues/deletes resource in distributed hashtable, changes step (3.21) then over to;
(3.8) whether decision node P is strong node, if enter step (3.9), otherwise change step (3.13) over to;
(3.9) node P searches in concordance list, obtains Search Results;
(3.10) node P waits for the response message of other nodes that have this resource postponing to spread search on many parsings ring structure topology of assembling, and the stand-by period is set; Node P searches in distributed hashtable, has the response message of other nodes of this resource by the time;
(3.11) when node P receives response message, enter step (3.17), otherwise change step (3.12) over to;
(3.12) when the node P stand-by period uses up, enter step (3.17), otherwise change step (3.11) over to;
(3.13) node P sends the searching resource request to its agent node S; Node P enters the Search Results state of waiting for; Node S searches in concordance list, obtains Search Results, and the result is told to node P;
(3.14) node S waits for the response message of other nodes that have this resource postponing to spread search on many parsings ring structure topology of assembling; Node S searches in distributed hashtable, has the response message of other nodes of this resource by the time;
(3.15) when node S receives response message, tell node P, enter step (3.17), otherwise change step (3.16) over to;
(3.16) when the node S stand-by period uses up, tell node P, enter step (3.17), otherwise change step (3.15) over to;
(3.17) according to user's input, need to judge whether the ending resource management, then finish if desired, otherwise, return step (3.1) and carry out next step resource management action.
2. universal resource management method according to claim 1 is characterized in that: step (1) is described all resources according to following step:
(1.1) node P selects a local resource, reads the title of this resource, the type of affiliated service and amount of capacity;
(1.2), otherwise change step (1.4) over to if this resource capacity size greater than the threshold values of default, enters step (1.3);
(1.3) node P reads size in the resource at random and is the partial content of default threshold values, enters step (1.5);
(1.4) node P reads the full content of resource;
(1.5) node P carries out Hash operation to the content of reading, and obtains cryptographic Hash;
(1.6) identifier that different services are added on resource in the node P read-out system;
(1.7) service and the cryptographic Hash that belong to according to resource of node P generates the unified finger URL of resource.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007101687228A CN100518129C (en) | 2007-12-10 | 2007-12-10 | Universal resource management method under confusion type peer-to-peer network environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007101687228A CN100518129C (en) | 2007-12-10 | 2007-12-10 | Universal resource management method under confusion type peer-to-peer network environment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101184016A true CN101184016A (en) | 2008-05-21 |
| CN100518129C CN100518129C (en) | 2009-07-22 |
Family
ID=39449108
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2007101687228A Expired - Fee Related CN100518129C (en) | 2007-12-10 | 2007-12-10 | Universal resource management method under confusion type peer-to-peer network environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100518129C (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101742702A (en) * | 2009-12-31 | 2010-06-16 | 中兴通讯股份有限公司 | Mobile internet service sharing system and method |
| WO2010069245A1 (en) * | 2008-12-16 | 2010-06-24 | 华为终端有限公司 | Method, apparatus and processing system for obtaining and reporting information of instance identifier |
| CN101924793A (en) * | 2010-08-17 | 2010-12-22 | 北京易视腾科技有限公司 | P2P streaming media-based secondary coding play method and system |
| WO2011000209A1 (en) * | 2009-06-29 | 2011-01-06 | 日电（中国）有限公司 | Server, method and system for providing node information for p2p network |
| CN102164186A (en) * | 2011-05-31 | 2011-08-24 | 互动在线（北京）科技有限公司 | Method and system for realizing cloud search service |
| CN101399778B (en) * | 2008-10-30 | 2011-08-31 | 北京邮电大学 | Node organizing method and resource discovery method in wireless P2P network |
| WO2012025052A1 (en) * | 2010-08-27 | 2012-03-01 | Dong Fang | System and method for internet search based on point-to-point network |
| CN102668505B (en) * | 2009-10-21 | 2015-02-18 | 瑞典爱立信有限公司 | Method and arrangement for locating services in a peer-to-peer network |
| CN104937902A (en) * | 2012-12-14 | 2015-09-23 | 微软技术许可有限责任公司 | Content source selection in a p2p network |
| CN106375130A (en) * | 2016-09-28 | 2017-02-01 | 苏州迈科网络安全技术股份有限公司 | Smart bandwidth management method and system based on Linux |
| CN108449374A (en) * | 2017-10-16 | 2018-08-24 | 同济大学 | Web3D virtual scenes equity transmission data dispatching method and platform |
| US10129334B2 (en) | 2012-12-14 | 2018-11-13 | Microsoft Technology Licensing, Llc | Centralized management of a P2P network |
| US10284641B2 (en) | 2012-12-14 | 2019-05-07 | Microsoft Technology Licensing, Llc | Content distribution storage management |
| US10391387B2 (en) | 2012-12-14 | 2019-08-27 | Microsoft Technology Licensing, Llc | Presenting digital content item with tiered functionality |
| CN111369777A (en) * | 2020-04-21 | 2020-07-03 | 深圳芯珑电子技术有限公司 | Maintenance method and maintenance system for low-voltage user electricity consumption information acquisition system |
| CN111629051A (en) * | 2020-05-26 | 2020-09-04 | 北京邮电大学 | Performance optimization method and device for industrial internet identification analysis system |
| CN114285602A (en) * | 2021-11-26 | 2022-04-05 | 成都安恒信息技术有限公司 | Distributed service safety detection method |
-
2007
- 2007-12-10 CN CNB2007101687228A patent/CN100518129C/en not_active Expired - Fee Related
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101399778B (en) * | 2008-10-30 | 2011-08-31 | 北京邮电大学 | Node organizing method and resource discovery method in wireless P2P network |
| WO2010069245A1 (en) * | 2008-12-16 | 2010-06-24 | 华为终端有限公司 | Method, apparatus and processing system for obtaining and reporting information of instance identifier |
| US8661134B2 (en) | 2008-12-16 | 2014-02-25 | Huawei Device Co., Ltd. | Method and apparatus for obtaining and reporting instance label information and system for processing instance label information |
| WO2011000209A1 (en) * | 2009-06-29 | 2011-01-06 | 日电（中国）有限公司 | Server, method and system for providing node information for p2p network |
| CN102668505B (en) * | 2009-10-21 | 2015-02-18 | 瑞典爱立信有限公司 | Method and arrangement for locating services in a peer-to-peer network |
| WO2010148843A1 (en) * | 2009-12-31 | 2010-12-29 | 中兴通讯股份有限公司 | System and method for mobile internet service sharing |
| CN101742702B (en) * | 2009-12-31 | 2014-03-19 | 中兴通讯股份有限公司 | Mobile internet service sharing system and method |
| US8819137B2 (en) | 2009-12-31 | 2014-08-26 | Zte Corporation | System and method for sharing mobile internet service |
| CN101742702A (en) * | 2009-12-31 | 2010-06-16 | 中兴通讯股份有限公司 | Mobile internet service sharing system and method |
| CN101924793B (en) * | 2010-08-17 | 2013-03-13 | 北京易视腾科技有限公司 | P2P streaming media-based secondary coding play method and system |
| CN101924793A (en) * | 2010-08-17 | 2010-12-22 | 北京易视腾科技有限公司 | P2P streaming media-based secondary coding play method and system |
| WO2012025052A1 (en) * | 2010-08-27 | 2012-03-01 | Dong Fang | System and method for internet search based on point-to-point network |
| US9330177B2 (en) | 2010-08-27 | 2016-05-03 | Dong Software Studio Beijing | System, method and device for internet search based on peer-to-peer network |
| CN102164186A (en) * | 2011-05-31 | 2011-08-24 | 互动在线（北京）科技有限公司 | Method and system for realizing cloud search service |
| CN102164186B (en) * | 2011-05-31 | 2013-10-30 | 互动在线（北京）科技有限公司 | Method and system for realizing cloud search service |
| CN104937902A (en) * | 2012-12-14 | 2015-09-23 | 微软技术许可有限责任公司 | Content source selection in a p2p network |
| US9781056B2 (en) | 2012-12-14 | 2017-10-03 | Microsoft Technology Licensing, Llc | Content source selection in a P2P network |
| US10129334B2 (en) | 2012-12-14 | 2018-11-13 | Microsoft Technology Licensing, Llc | Centralized management of a P2P network |
| US10284641B2 (en) | 2012-12-14 | 2019-05-07 | Microsoft Technology Licensing, Llc | Content distribution storage management |
| US10391387B2 (en) | 2012-12-14 | 2019-08-27 | Microsoft Technology Licensing, Llc | Presenting digital content item with tiered functionality |
| CN106375130A (en) * | 2016-09-28 | 2017-02-01 | 苏州迈科网络安全技术股份有限公司 | Smart bandwidth management method and system based on Linux |
| CN108449374A (en) * | 2017-10-16 | 2018-08-24 | 同济大学 | Web3D virtual scenes equity transmission data dispatching method and platform |
| CN108449374B (en) * | 2017-10-16 | 2020-08-14 | 吉林吉动盘古网络科技股份有限公司 | Web3D virtual scene peer-to-peer transmission data scheduling method and platform |
| CN111369777A (en) * | 2020-04-21 | 2020-07-03 | 深圳芯珑电子技术有限公司 | Maintenance method and maintenance system for low-voltage user electricity consumption information acquisition system |
| CN111629051A (en) * | 2020-05-26 | 2020-09-04 | 北京邮电大学 | Performance optimization method and device for industrial internet identification analysis system |
| CN111629051B (en) * | 2020-05-26 | 2021-12-03 | 北京邮电大学 | Performance optimization method and device for industrial internet identification analysis system |
| CN114285602A (en) * | 2021-11-26 | 2022-04-05 | 成都安恒信息技术有限公司 | Distributed service safety detection method |
| CN114285602B (en) * | 2021-11-26 | 2024-02-02 | 成都安恒信息技术有限公司 | Distributed service security detection method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN100518129C (en) | 2009-07-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100518129C (en) | Universal resource management method under confusion type peer-to-peer network environment | |
| EP2058980B1 (en) | A method, system and device for establishing a peer to peer connection in a p2p network | |
| Lua et al. | A survey and comparison of peer-to-peer overlay network schemes | |
| Ramanathan et al. | Finding good peers in peer-to-peer networks | |
| CN101378409B (en) | Optimal operation of hierarchical peer-to-peer networks | |
| Joseph | An extendible open source P2P simulator | |
| Guerraoui et al. | Gosskip, an efficient, fault-tolerant and self organizing overlay using gossip-based construction and skip-lists principles | |
| Scheideler et al. | A distributed and oblivious heap | |
| Sacha et al. | Using aggregation for adaptive super-peer discovery on the gradient topology | |
| CN111800516B (en) | Internet of things equipment management method and device based on P2P | |
| Medrano-Chávez et al. | A performance comparison of Chord and Kademlia DHTs in high churn scenarios | |
| CN101626336A (en) | Method for improving P2P overlay network interaction performance by using trust mechanism | |
| Li et al. | A chord-based novel mobile peer-to-peer file sharing protocol | |
| KR20110070447A (en) | Method of distributed discovery of communication participant in data distribution service | |
| Mei et al. | PP-COSE: a P2P community search scheme | |
| Pietiläinen et al. | Experiments in mobile social networking | |
| Sotiriadis et al. | Using self-led critical friend topology based on P2P chord algorithm for node localization within cloud communities | |
| Sacha et al. | A gradient topology for master-slave replication in peer-to-peer environments | |
| Zeng et al. | A novel lookup and routing protocol based on CAN for structured P2P network | |
| Sethom et al. | Palma: A P2P based Architecture for Location Management | |
| Adamsky et al. | Locust: Highly Concurrent DHT Experimentation Framework for Security Evaluations | |
| Ktari et al. | A peer-to-peer social network overlay for efficient information retrieval and diffusion | |
| Gatani et al. | Notice of Violation of IEEE Publication Principles: Reinforcement learning for P2P searching | |
| Gatani et al. | An adaptive routing mechanism for efficient resource discovery in unstructured p2p networks | |
| CN114338676A (en) | Block synchronization method and device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090722 Termination date: 20111210 |