CN1326362C - Network partitioning method for large-scale complicated network - Google Patents
Network partitioning method for large-scale complicated network Download PDFInfo
- Publication number
- CN1326362C CN1326362C CNB2004100182896A CN200410018289A CN1326362C CN 1326362 C CN1326362 C CN 1326362C CN B2004100182896 A CNB2004100182896 A CN B2004100182896A CN 200410018289 A CN200410018289 A CN 200410018289A CN 1326362 C CN1326362 C CN 1326362C
- Authority
- CN
- China
- Prior art keywords
- node
- network
- subnet
- shortest path
- centroid
- 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.)
- Expired - Fee Related
Links
Images
Abstract
The present invention relates to a network partitioning method for large-scale complicated networks, which is used for the technical field of networks. The present invention comprises a network preprocessing method, a subnetwork judging method and a network partitioning method. Firstly, in the network preprocessing method, a topological diagram of a large-scale network; is improved into a diagram whose diagram is at least 2; then, the large-scale network is further partitioned into a plurality of annular subnetworks and fan-like subnetworks by the combination of the network judging method and the network partitioning method; the whole network is partitioned into subnetworks by the network partitioning method; the subnetworks are classified into the annular subnetworks with closed external links and the fan-like subnetworks with non-closed external links according to the subnetwork judging method. The large-scale network of the present invention adopts a distributed intelligent recovery strategy by using each subnetwork as a unit, so the survivability of large-scale complicated networks can be flexibly and conveniently achieved; in addition, multiple failures can be simultaneously processed by the subnetworks, which saves time for failure elimination.
Description
Technical field
The present invention relates to a kind of network architecture dividing method of communication network, especially relate to a kind of large complicated network of network dividing method.Be used for networking technology area.
Background technology
The quick growth of Network impels network technology to obtain significant progress, and simultaneously along with the various new switching technologies and the introducing of transmission technology, the scale of network and complexity are also increasing day by day.This has higher requirement also for corresponding network survivability technology.The survivability of network is the focus of network technology research always, the survivability technology of network and the topological property of network self are closely-related, because network size was less in the past, relevant network survivability technology is carried out unified management by Master Control Center to whole network mostly based on centralized control strategy.Network makes that to the inexorable trend of more extensive development centralized survivability technology is unable to do what one wishes a bit.The distributed fault recovery technology is implemented the optimization allotment of Internet resources at the locality of fault in subrange, can realize fast fault diagnosis and recovery, thereby become the mainstream technology of large complicated network survivability.
Find one piece of article that L.Guangzhi etc. deliver " the distribution type restoration strategy in a kind of MESH of being effectively applied to network recovery " (L.Guangzhi by literature search in international network meeting in 2003; W.Dongmei, C.Kalmanek, and R.Doverspike.Efficient distributedrestoration path selection for shared mesh restoration[C] .IEEE Trans.Networking, 11 (5), Oct.2003,761-771.) article, just this problem has been done comparatively comprehensively summary.The enforcement distribution type restoration mainly contains the work aspect two in large complicated network: the one, and utilize the topological relation of network catenet to be partitioned into the subnet of being convenient to distribution type restoration.The 2nd, implement rational failure diagnosis and recovery algorithms in subnet or between subnet.Obviously, a kind of method of network division fast and effectively is the prerequisite and the basis of realizing the distribution type restoration strategy.At present, mostly adopt the mode of information broadcast to carry out distributed control, its distinct disadvantage is that resource occupying is big, and information interaction mechanism is poor, also do not have a kind of suitable effective network division method at present both at home and abroad, this is restricting the development of distributed management and control always.
Summary of the invention
The objective of the invention is to deficiency, a kind of suitable effective large complicated network of network dividing method is provided, make it to be applicable to the distributed management and the control of large complicated network at the previous designs scheme.
The present invention is achieved by the following technical solutions, the inventive method comprises the network preprocess method, three parts of the decision method of subnet and sub-network dividing method, at first adopt the network preprocess method that the topological diagram of catenet is improved as an edge connectivity and be no less than 2 figure, decision method and sub-network dividing method further combined with subnet is partitioned into several annular subnets and fan-shaped subnet with described catenet then, wherein utilize the sub-network dividing method that whole network is divided into subnet one by one, according to the decision method of subnet, subnet is divided into the annular subnet with closed outer shroud link and has two kinds of the fan-shaped subnets of the outer shroud link of non-closure.
Through described network preliminary treatment, it is 3 figure that the topological diagram of catenet becomes an edge connectivity, concrete, be 1 node for degree, i.e. landing top, directly from topology, remove, for degree is 2 node, in topology, two adjacent nodes of this category node are carried out with the limit direct-connected, for monolateral cut set, the cut set that promptly has only a limit is merged into a new node with described monolateral and two adjacent node.
Below the present invention is further illustrated:
The decision method of described subnet may further comprise the steps:
1) from network, finds out all adjacent nodes of node n, and classify all adjacent nodes of node n as a S set (n) in order by the order of connection;
2) node among the pair set S (n) is pressed from first to last in order, the node that partners in twos is right, as by sequence number before and after it by classifying (1 as, 2), (2,3) ..., in original catenet, adopt single source shortest path algorithm to find the solution shortest path to each to node, can find out between every pair of node that shortest path is end to end just can to constitute a track;
Remove all links in the described track, utilize single source shortest path algorithm to find the solution first node among the S (n) and the shortest path between the tail node again, if do not have to separate and to judge that then with described node n be Centroid, and just can constitute a fan-shaped subnet as the outer shroud link with described track, if have shortest path between first node among the S (n) and the tail node, but among the S (n) any a pair of node between shortest path greater than half of first node among described track and the S (n) and the shortest path sum between the tail node, can judge also that then with described node n be Centroid, and can constitute a fan-shaped subnet as the outer shroud link with described track, if exist between first node among the described S (n) and the tail node any a pair of node among shortest path and the S (n) between shortest path all be not more than first node among described track and the S (n) and half of the shortest path sum between the tail node, can judge that then with described node n be Centroid, and can constitute an annular subnet as the outer shroud link with first node among described track and the S (n) and the shortest path between the tail node;
Described sub-network dividing method may further comprise the steps:
A. be that the breadth First traversal is carried out to whole network in the summit with node in the network, thereby the range optimization traversal that can obtain whole network generate tree;
B. search for the described range optimization traversal that obtains in the previous step and generate tree, utilize the decision method of subnet, find out the Centroid of an annular subnet, and from described range optimization traversal generation tree, remove this annular subnet Centroid, and all nodes on the outer shroud link of the Centroid of this annular subnet;
C. the method that continues B set by step generates tree to described range optimization traversal and searches for, its search order is to child node from father node, from left sibling to right node, generate tree until complete described range optimization traversal of search, to find out the Centroid of all annular subnets;
D. removing all and use node fully in network, use node to be meant that each bar adjacent link of this node used by an annular subnet at least fully, is exactly to obtain a rest network like this;
E. in rest network, select the Centroid of a maximum node of connected node number, and in rest network, remove the Centroid of this fan-shaped subnet as a fan-shaped node, and all nodes on the outer shroud link of the Centroid of this fan-shaped subnet;
F. continue set by step the method for E described rest network is searched for, until complete described rest network of search, to find out the Centroid of all fan-shaped subnets.
So far, the Centroid of all annular subnets and the Centroid of all fan-shaped subnet have been found, center and corresponding outer shroud link thereof with described annular subnet can constitute an annular subnet, center and corresponding outer shroud link thereof with described fan-shaped subnet can constitute a fan-shaped subnet, and whole network division finishes.
The present invention compares effect with foregoing existing program be positive and tangible, compare following characteristics: the subnet segmentation strategy that is proposed can be cut apart catenet effectively, and then employing distributed intelligence recovery policy, can solve the survivability problem that is faced of large complicated network flexibly and easily, and on this partitioning algorithm basis, each subnet parallel processing simultaneously many places fault is for the time has been won in fault recovery.
Description of drawings
Fig. 1 is the example of a large complicated network
Fig. 2 is the pretreated network topology structure of large complicated network
Fig. 3 preferably travels through the generation tree for the range of pretreated network
Subnet tabulation after the large complicated for this reason network utilisation the inventive method of Fig. 4 is cut apart
Embodiment:
With Fig. 1, Fig. 2, Fig. 3 and Fig. 4 is that example describes implementation procedure of the present invention in detail, and its concrete implementation process is as described below:
The first step is the network preliminary treatment, described network preprocess method is the topological diagram with catenet, concrete, it is 3 figure that the topological diagram of described catenet is become an edge connectivity, it for degree 1 node, it is landing top, directly from topology, remove, it for degree 2 node, in topology, two adjacent nodes of this category node are carried out direct-connectedly,, promptly have only the cut set on a limit for monolateral cut set with the limit, described monolateral and two adjacent node is merged into a new node, and network shown in Figure 1 through pretreated network topology as shown in Figure 2.
Network shown in Figure 2 further adopts the cutting procedure of segmentation strategy of invention as described below: with node in the network 1 is that the breadth First traversal is carried out to whole network in the summit, thereby the range optimization traversal that can obtain whole network generates tree; Corresponding range optimization traversal generates tree as shown in Figure 3 among the embodiment; The described range optimization traversal that obtains in the search previous step generates tree, utilize the decision method of subnet of the present invention, find out the Centroid of first annular subnet, node 5, and from described range optimization traversal generation tree, remove node 5, and all nodes on the outer shroud link of node 5; The method that continues B set by step generates tree to described range optimization traversal and searches for, its search order is to child node from father node, from left sibling to right node, generate tree until complete described range optimization traversal of search, to find out the Centroid of all annular subnets, its set is for { 5,7,18,21}; Removing all and use node fully in network, use node to be meant that each bar adjacent link of this node used by an annular subnet at least fully, is exactly to obtain a rest network like this; Select a maximum node of connected node number in rest network, node 15 is as the Centroid of fan-shaped node, and node 15 and its outer shroud link have just constituted a fan-shaped subnet, and so far, whole network division finishes, and concrete subnet situation as shown in Figure 4.
Utilize the present invention easily large complicated communication network to be divided into the subnet that has independent recovery capability one by one, thereby can utilize the distributed management mechanism of subnet can realize the distributed intelligence recovery policy, can solve the survivability problem that is faced of large complicated network flexibly and easily, and utilize the present invention, each subnet parallel processing simultaneously many places fault is for the time has been won in fault recovery.
Claims (6)
1, a kind of large complicated network of network dividing method, it is characterized in that, comprise the network preprocess method, the decision method of subnet and sub-network dividing method, at first adopt the network preprocess method that the topological diagram of catenet is improved as an edge connectivity and be equal to or greater than 2 figure, decision method and sub-network dividing method further combined with subnet is partitioned into several annular subnets and fan-shaped subnet with catenet then, wherein utilize the sub-network dividing method that whole network is divided into subnet one by one, according to the decision method of subnet, subnet is divided into the annular subnet with closed outer shroud link and has two kinds of the fan-shaped subnets of the outer shroud link of non-closure.
2, large complicated network of network dividing method as claimed in claim 1 is characterized in that the decision method of described subnet may further comprise the steps:
1) from network, finds out all adjacent nodes of node n, and classify all adjacent nodes of node n as a S set (n) in order by the order of connection;
2) node among the pair set S (n) is pressed from first to last in order, the node that partners in twos is right, by sequence number before and after it by classifying (1 as, 2), (2,3) ..., in original catenet, adopt shortest path algorithm to find the solution shortest path to each to node, find out track of an end to end formation of shortest path between every pair of node;
3) remove all links in the described track, utilize shortest path algorithm to find the solution first node among the S (n) and the shortest path between the tail node again, if do not have to separate and judge that then with described node n be Centroid, and just can constitute a fan-shaped subnet as the outer shroud link with described track, if have shortest path between first node among the S (n) and the tail node, but among the S (n) any a pair of node between shortest path greater than half of first node among described track and the S (n) and the shortest path sum between the tail node, can judge also that then with described node n be Centroid, and can constitute a fan-shaped subnet as the outer shroud link with described track, if exist between first node among the described S (n) and the tail node any a pair of node among shortest path and the S (n) between shortest path all be less than or equal to first node among described track and the S (n) and half of the shortest path sum between the tail node, judge that then with described node n be Centroid, and can constitute an annular subnet as the outer shroud link with first node among described track and the S (n) and the shortest path between the tail node.
3, large complicated network of network dividing method as claimed in claim 2, its feature are being that what described shortest path algorithm adopted is single source shortest path algorithm.
4, large complicated network of network dividing method as claimed in claim 2 is characterized in that, described sub-network dividing method may further comprise the steps:
A. be that the breadth First traversal is carried out to whole network in the summit with node in the network, thereby the range optimization traversal that obtains whole network generate tree;
B. search for the described range optimization traversal that obtains in the previous step and generate tree, utilize the decision method of subnet, find out the Centroid of an annular subnet, and from described range optimization traversal generation tree, remove this annular subnet Centroid, and all nodes on the outer shroud link of the Centroid of this annular subnet;
C. the method that continues B set by step generates tree to described range optimization traversal and searches for, and generates tree until complete described range optimization traversal of search, to find out the Centroid of all annular subnets;
D. in network, remove all and use node fully, use node to be meant that each bar adjacent link of this node used by an annular subnet at least fully, so just obtain a rest network;
E. in rest network, select the Centroid of a maximum node of connected node number, and in rest network, remove the Centroid of this fan-shaped subnet as a fan-shaped subnet, and all nodes on the outer shroud link of the Centroid of this fan-shaped subnet;
F. continue set by step the method for E described rest network is searched for, until complete described rest network of search, to find out the Centroid of all fan-shaped subnets.
5, large complicated network of network dividing method as claimed in claim 4, its feature are being that generating the search order of seeking annular subnet Centroid in the tree in described range optimization traversal is to child node, from left sibling to right node from father node.
6, large complicated network of network dividing method as claimed in claim 1, it is characterized in that, through described network preliminary treatment, it is 3 figure that the topological diagram of described catenet becomes an edge connectivity, concrete, it for degree 1 node, it is landing top, directly from topology, remove, for degree is 2 node, in topology, two adjacent nodes of this category node are carried out with the limit direct-connected, for monolateral cut set, the cut set that promptly has only a limit is merged into a new node with described monolateral and two adjacent node.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB2004100182896A CN1326362C (en) | 2004-05-13 | 2004-05-13 | Network partitioning method for large-scale complicated network |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CNB2004100182896A CN1326362C (en) | 2004-05-13 | 2004-05-13 | Network partitioning method for large-scale complicated network |
Publications (2)
Publication Number | Publication Date |
---|---|
CN1571363A CN1571363A (en) | 2005-01-26 |
CN1326362C true CN1326362C (en) | 2007-07-11 |
Family
ID=34479438
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CNB2004100182896A Expired - Fee Related CN1326362C (en) | 2004-05-13 | 2004-05-13 | Network partitioning method for large-scale complicated network |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN1326362C (en) |
Families Citing this family (11)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101466101B (en) * | 2007-12-20 | 2010-08-18 | 北京沃泰丰通信技术有限公司 | Networking method of mobile network and device thereof |
CN102137405B (en) * | 2010-06-01 | 2014-04-02 | 华为技术有限公司 | Method and device for constructing network topology |
CN103336810B (en) * | 2013-06-26 | 2016-08-31 | 国家电网公司 | A kind of Topology Analysis of Power Distribution Network method based on multi-core computer |
CN104168132B (en) * | 2014-06-18 | 2017-12-12 | 华为技术有限公司 | Method for diagnosing faults, equipment and system |
CN104079433B (en) * | 2014-06-30 | 2017-05-31 | 中国矿业大学 | A kind of method for improving complex network structures robustness |
US9313663B2 (en) * | 2014-08-21 | 2016-04-12 | Futurewei Technologies, Inc. | Automatic division of internet protocol radio access networks to interior gateway protocol areas |
US20170180208A1 (en) * | 2015-12-22 | 2017-06-22 | Intel Corporation | Organically composable iot networks |
CN106603369B (en) * | 2016-12-20 | 2019-06-11 | 浪潮天元通信信息系统有限公司 | A method of each network element institute is calculated automatically into shortest path loop chain |
CN106815662A (en) * | 2017-03-28 | 2017-06-09 | 鲁可儿 | A kind of express mail service collecting and distributing centre and method for solving |
WO2020113437A1 (en) * | 2018-12-04 | 2020-06-11 | 区链通网络有限公司 | Graph structure processing method and system, network device and storage medium |
CN111490897B (en) * | 2020-02-27 | 2021-04-20 | 华中科技大学 | Network fault analysis method and system for complex network |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0951196A2 (en) * | 1998-02-24 | 1999-10-20 | AT&T Corp. | Optical layer quasi-centralized restoration |
-
2004
- 2004-05-13 CN CNB2004100182896A patent/CN1326362C/en not_active Expired - Fee Related
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0951196A2 (en) * | 1998-02-24 | 1999-10-20 | AT&T Corp. | Optical layer quasi-centralized restoration |
Non-Patent Citations (4)
Title |
---|
Efficient distributed restoration path selection for shared meshrestoration Guangzhi Li、Dongmei Wang、Kalmanek,C.Doverspike,R,IEEE/ACM Transactions on Networking,Vol.11 No.5 2003 * |
Efficient distributed restoration path selection for shared meshrestoration Guangzhi Li、Dongmei Wang、Kalmanek,C.Doverspike,R,IEEE/ACM Transactions on Networking,Vol.11 No.5 2003;光虚拟专用网技术的研究 谢新梅,宋荣方,光通信技术,第11卷 2003;虚拟专用网的生存性方案 贺辉、范戈,光通信技术,第9卷 2003 * |
光虚拟专用网技术的研究 谢新梅,宋荣方,光通信技术,第11卷 2003 * |
虚拟专用网的生存性方案 贺辉、范戈,光通信技术,第9卷 2003 * |
Also Published As
Publication number | Publication date |
---|---|
CN1571363A (en) | 2005-01-26 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
ALRikabi et al. | Secure Chaos of 5G Wireless Communication System Based on IOT Applications. | |
CN1326362C (en) | Network partitioning method for large-scale complicated network | |
US9515845B2 (en) | Utility communication method and system | |
US11271774B2 (en) | Redundant multicast tree in a fat tree network topology with anisotropic routing | |
CN103493441A (en) | Generating a loop-free routing topology using routing arcs | |
CN103346979B (en) | Flow distribution method and equipment in SPBM network | |
DE102010056369A1 (en) | Routed Split Multilink Trunking for IPv6 | |
CN102420797B (en) | Topology mapping method and system | |
CN102065006B (en) | Method for recovering inter-domain failure of cross-domain label switched path | |
CN1323522C (en) | Method for determining relation between routers at fringe of client site and virtual private network | |
CN107690168A (en) | A kind of expansible networking method of wireless sensor network | |
CN113810225B (en) | In-band network telemetry detection path planning method and system of SDN (software defined network) | |
CN107135160B (en) | Spanning tree fault-tolerant method based on network failure node | |
CN1302637C (en) | A multiple spanning tree protocol bridge priority coordination method | |
CN105515984B (en) | A kind of more means of communication route planning methods of multipath | |
Sarma et al. | Near-optimal random walk sampling in distributed networks | |
CN102045228A (en) | Discovery method and system of loop network protection information | |
CN109474908A (en) | A kind of aeronautical Ad hoc networks method of task based access control driving | |
CN107689916A (en) | A kind of method and system for seeking complete risk sharing link group disjoint paths pair | |
Indukuri et al. | Dominating sets and spanning tree based clustering algorithms for mobile ad hoc networks | |
Nair et al. | Bandwidth sensitive fast failure recovery scheme for Metro Ethernet | |
CN112910689A (en) | Clock network topology construction method and system | |
CN108111418B (en) | Communication system, data transmission method and node | |
CN108989201A (en) | A kind of end-to-end path configuration method for SNCP tangent rings business | |
Chaturvedi et al. | Improved algorithms for construction of connected dominating set in MANETs |
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: 20070711 Termination date: 20100513 |