CN103595734B - Based on the online social network fast repairing method that user-association structure divides - Google Patents

Abstract

The present invention discloses a quick repair method for an online social network based on user association structure division, 1) constructing a user association graph G according to user relationship data in an online social network; 2) calculating the message dissemination capabilities of nodes and node groups in the graph G; 3 ) Divide the graph G into the first-level user association structure group according to the message dissemination ability, and divide the second-level user association structure group on the basis of the first-level user association structure group; 4) In the user association structure group, the weight The largest node is marked as a key node, and the adjacent nodes between groups are marked as boundary nodes; 5) Follow the quick repair method of sending patches to key nodes and/or boundary nodes first, and then gradually patch other nodes, Complete the repair. This method is aimed at the online social network used on the intelligent terminal, and has the characteristics of quickly and effectively repairing the online social network vulnerability under the condition of bandwidth limitation, and preventing the large-scale spread and harm of the attack.

Description

A fast repair method for online social network based on user association structure division

technical field

The invention belongs to the technical field of network security, and in particular relates to an online social network fast repair method based on user association structure division in an intelligent terminal.

Background technique

With the continuous development of online social networks and smart terminals, more and more users begin to frequently use online social networks on smart terminals, and the impact of subsequent security issues is becoming more and more severe. Among them, attacks and privacy theft through security loopholes in online social networks are one of the most common threats, and this threat continues to intensify with the increasing dependence and frequency of use on smart terminals and social networks. The security protection of smart terminals and social networks and the protection of user privacy have brought severe challenges.

Online social networks, due to various loopholes that are unavoidable in their own implementation, complex network topology features, social engineering features, and open APIs for third-party implementation and development, etc., provide a good source for various vulnerability attacks and their rapid and widespread spread. With favorable conditions, examples of attacks on online social networks have emerged in an endless stream in recent years. With the popularization and large-scale use of smart terminals, users begin to use online social networks on smart terminals. When a vulnerability is attacked and an update patch needs to be released, the network bandwidth on the smart terminal cannot be compared with that of the PC terminal. , and cannot guarantee which network the user is using, and cannot guarantee the bandwidth. Due to the limitations of the network bandwidth of smart terminals and other issues, there is a problem that it is impossible to uniformly distribute repair patches in the shortest time, which provides favorable conditions for vulnerability attacks. How to arrange vulnerability repair and patch distribution, so that the vulnerabilities of online social networks on smart terminals can be repaired as soon as possible, and the damage of attacks can be minimized is one of the key issues that need to be studied and solved urgently.

At present, some working researchers have studied the suppression and repair methods of malicious codes in smart terminals and online social networks. Relatively representative related methods mainly include: In terms of the worm suppression and repair method for intelligent terminals, Professor Zhu Sencun of Pennsylvania State University and others proposed a social relationship-based mobile phone worm suppression method for worms that spread through mobile phones. The repair scheme, the main idea is to extract the social relationship graph in the mobile network and divide it into different groups, and control the worm in a specific area as much as possible by patching the members in the group, thereby suppressing the spread of the worm. Aiming at the method of suppressing the spread of worms in social networks, Nguyen et al. from the University of Florida proposed a method for suppressing the spread of worms in social networks, which takes into account the dynamic changes in the topology of social networks and utilizes the group structure of social networks Find out the core weak nodes in the network, and suppress the spread of worms by repairing these nodes.

The above method provides a good reference for the vulnerability repair and patch distribution of the smart terminal social network, but because the user relationship of the online social network is more complicated than the social relationship in the mobile phone network, if the current social relationship for the mobile phone is directly applied A large number of users will gather in the same user group, which cannot achieve a good partition suppression effect; and due to the limitation of the bandwidth of the mobile network of smart terminals, the repair and patch distribution methods of online social networks in the traditional Internet are limited. As a result, it is impossible to quickly and effectively implement patch distribution and vulnerability repair, and its time lag will bring more users to be attacked. Therefore, how to realize the rapid and effective method of issuing patches and repairing vulnerabilities in the online social network of smart terminals and reducing the damage of attacks is an important requirement for maintaining the security of online social networks of smart terminals at present.

Contents of the invention

Aiming at the problem of how to implement quick repairs to reduce the damage of attacks when there are loopholes and attacks in the online social network used on smart terminals, the purpose of the present invention is to propose a quick repair method for online social networks based on user association structure division in smart terminals , by analyzing the information dissemination of online social networks, users are divided into association structures according to the analysis results, and key nodes and boundary nodes in the association structure are identified and marked. The way of nodes and border nodes can effectively limit the spread of vulnerability attacks and achieve the purpose of rapid repair under the condition of limited bandwidth of smart terminals.

The technical scheme of the present invention is as follows: a method for quickly repairing an online social network based on user association structure division, the steps of which include:

1) Construct a user association graph G on the mobile smart terminal according to the user relationship data in the online social network;

2) calculating the message dissemination capability of nodes and node groups in the graph G;

3) dividing the graph G into first-level user association structure groups according to the message dissemination capability, and performing second-level user association structure group division on the basis of the first-level user association structure groups;

4) marking the node with the largest weight as a key node in the first-level user association structure group and/or the second-level user association structure group, and marking the adjacent nodes between the groups as boundary nodes;

5) Follow the quick repair method of sending patches to key nodes and/or border nodes first, and then gradually repair other nodes to complete the repair.

Furthermore, the message dissemination capability of the group U where n is the number of nodes in U, m is the number of connecting edges, and p _ij is the message propagation capability of nodes.

Furthermore, the p _ij is represented by obtaining the message propagation probability between nodes in the social network per unit time.

Furthermore, according to the message transmission capability, the user nodes with frequent message transmission are classified into the same structural group, so that the message transmission capability within the group is strong, and the message transmission capability between groups is low.

Furthermore, the intelligent mobile terminal is a smart phone, Pad, PDA.

Furthermore, the user relationship data includes the user, the friend relationship between users, the number of friends of the user, and the connection between users.

Further, the construction of the user association graph G={V, E, W _V }, wherein the node V represents the user, the edge E represents the relationship between users, and W _V is the node weight representing the number of associated friends of the user as the node weight .

Furthermore, the division method of the first-level user association structure group is:

1) Find the node v _i with the largest weight in the social network user association graph G, add v _i and its adjacent nodes that have not joined any group to the group U, and calculate the message delivery probability S(U) in the group U , and build a boundary set B(v) for all nodes adjacent to the group U;

2) For the nodes in B(v) adjacent to the group U, select the node v _j with the largest weight among them, and calculate the message delivery probability S(U+j) of the node j to join the group U after joining the group U, if S(U+j)>S(U), then add node j to group U, and expand B(v) until no new adjacent nodes can join group U;

3) Traverse the steps 1)-2) and divide all the nodes in the graph G into groups.

Furthermore, the specific method of dividing the second-level user association structure group is:

1) In the first-level user association structure group, find the node v _i with the largest weight, and then set the ratio value r according to the needs, find all the nodes in the group with weight w _vj >r*w _vi , for each The nodes satisfying this condition build a new node group; wherein, w _vi is the weight of the node with the largest weight in the first-level user association structure group, and w _vj is the weight of any node v _j in the first-level user association structure group;

2) According to the new node group obtained above, according to the division method of the first-level user association structure group, calculate the communication capabilities of their adjacent nodes joining the group and not joining the group, and complete the second in the first-level user association structure group. The division of user-level user association structure groups.

Furthermore, when a node is connected to two groups at the same time and meets the conditions for joining a group, it will be added to the group whose calculated message delivery probability S is larger.

Beneficial effect:

The present invention provides an online social network fast repair method based on user association structure division in an intelligent terminal, and its characteristics are as follows:

The present invention provides a layered and gradual repairing method to ensure fast repairing as much as possible, according to the problem that the slow repairing attack is easy to be implemented in a large area due to the limitation of the distribution of vulnerability repairs brought about by the network bandwidth of the intelligent terminal. Completion and limitations of the implementation of the attack. Based on the probability characteristics of information transmission among online social network users, the present invention divides users into groups with associated structures, so that the probability of message transmission between groups is lower than that within a group, and key nodes and boundary nodes are identified from groups. On the one hand, it is guaranteed to repair the user nodes with strong transmission ability and regional spread between groups, so that the attack cannot spread on a large scale; Vulnerabilities and attacks are fixed.

Description of drawings

Fig. 1 is a flow chart of the method of the present invention;

detailed description

The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. It should be understood that the described embodiments are only some of the embodiments of the present invention, not all of them. example. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts belong to the protection scope of the present invention.

A specific embodiment of the present invention is as follows, an online social network fast repair method based on user association structure division in a smart terminal, the steps of which are as follows:

The method flow process of the present invention is as shown in Figure 1, and the steps of the present invention are as follows:

1) Obtain the user relationship in the online social network that needs to be maintained, and construct a social network user association graph;

2) In the social network user association graph, by analyzing the connection edges of different user nodes, calculate the message dissemination capabilities of nodes and node groups;

3) Divide user association structure groups into two levels. According to the information dissemination ability, the first-level user association structure group, and then on the basis of the first-level division, find the associated user nodes whose information dissemination ability is greater than the threshold, and divide the second-level user association structure group;

4) In the divided user association structure, identify and mark the key nodes and boundary nodes in each group structure;

5) When repair is needed, the key nodes and border nodes are repaired first, and then the key nodes and border nodes gradually send repair information to their adjacent nodes, and the repair is completed by using the social network user relationship.

In step 1), the specific method of constructing the social network user association graph is: obtain the user and the user relationship of the online social network to be repaired, the user friend relationship is obtained by the social network service provider, and the server also Generally documented. With users as nodes, user relationships as edges, and the number of user friends as node weight parameters, a social network user association graph is constructed.

In step 2), the specific method for calculating the message dissemination capability of the node and the node group is: obtain the information dissemination quantity of the user node in the latest period, and calculate the information transmission probability per unit time;

The method to calculate the message passing ability of a node group composed of a node and some adjacent nodes is: through the message passing probability of the node connected to this node in the node group and the node, calculate the unit time between this node and these connected nodes The probability of node message transmission is used as a representation of the message transmission ability of this node and the node group; then calculate the transmission ability of each node in a node group and the subgroup composed of the remaining nodes of the group, and calculate the average value as the node The ability of the group to disseminate information.

In step 3), the specific method of dividing the first-level user association structure is: at first in the social network user association graph constructed in step 1), find the node with the largest weight, if there is more than one node, then Randomly select one of them as the initial node group, and add the adjacent nodes of the initial node to the group; then calculate the propagation ability of the new group after all the adjacent nodes of this group join the group, when the propagation of the new group When the capacity is greater than the original group, join the corresponding node, otherwise do not join; after completing the creation of a group, select the node with the most relationship from the remaining nodes to form another node group; repeat the steps to complete the first-level user association Structural division.

In step 3), the specific method of dividing the second-level user association structure is: in the first-level user group, find the node with the largest weight, and then set a proportional value r according to actual needs, in this group The nodes whose ownership weight is greater than the maximum node weight*r are selected. In this group, the nodes are divided into groups from the beginning of each group for the second-level division. For the nodes that may appear in the division process and are connected to multiple second-level group nodes , attribute it to the group with the strongest group propagation ability after joining the node, thus completing the second-level user association structure division.

In step 4), the key nodes and boundary nodes specifically refer to: key nodes refer to the initial nodes of all divided groups in step 3); boundary nodes refer to nodes that are connected to other external groups.

In an embodiment of the present invention, the social network user association graph refers to a graph structure in which users are nodes, friendships between users are edges, and the number of users' friends is the weight of nodes. In a social network, if there is a friend relationship between users, they may send messages to each other, thus indicating the possible delivery path of the message. The number of friends of a user and the number of messages passed indicate a user's ability to spread messages to the outside world.

In an embodiment of the present invention, according to the connection edges of different nodes and their weights, the message transmission capabilities between nodes, between a node and a group composed of multiple nodes, and between node groups can be calculated. We obtain the latest message propagation data for a period of time, and express the message delivery capability by calculating the message delivery probability per unit time. This data is updated once at a specific time as needed to keep the method valid.

In an embodiment of the present invention, we divide user association structure groups according to the message dissemination ability between nodes. When using online social networks on smart terminals, when repairs are required, the main problem is bandwidth limitations and delays in repairs that may lead to large-scale spread of attacks. Therefore, our idea is to find key nodes and repair them first, and then gradually repair them To all nodes of the entire social network. To achieve this goal, our goal of group division is to divide all user nodes into multiple groups, so that the nodes with strong message transmission ability among each other are in one group, and the message transmission ability between groups is as weak as possible. To achieve this goal, we implement user association structure group division in two levels.

In an embodiment of the present invention, the key nodes and border nodes we mark are important nodes that may cause a larger spread of attacks after the method considers that the group is divided. Since the characteristics of the institutional groups divided by this method are that the internal information transmission ability of the group is strong, and the transmission ability between the groups is weak, so the boundary node is considered to be one of the important nodes that may cause a larger amount of transmission because it may cause transmission between groups. In addition, in order to control the spread in the group as soon as possible and repair it faster, according to our division scheme, the initial node in the group is the node with the largest message delivery capability of the group, so it is called a key node and marked as an important node.

In an embodiment of the present invention, the repair sequence of the method is: firstly send repairs to key nodes and border node terminals, and then use the node relationships in the social network association graph to complete gradual repairs to other nodes in turn.

As shown in Figure 1, a schematic flow diagram of an online social network fast repair method based on user association structure division in an intelligent terminal, including steps:

1. Construct social network user association graph

Obtain the user relationship data in the online social network, including the user, the friend relationship between users, the number of user friends, and the connections between users, and construct a social network user relationship graph G={V, E, W _V }, where node V represents User, edge E represents the relationship between users, W _V is the node weight, representing the number of associated friends of the user is the node weight. The social network user association map can be regularly updated according to actual accuracy needs, that is, the latest user situation and relationship data are obtained from the social network server or service provider.

2. Calculate the message dissemination capability of nodes and node groups

After the social network user association graph is constructed, the idea of the present invention is to divide the user association graph into different user groups according to the information transfer capability between nodes. To achieve this goal, it is first necessary to define and calculate the message dissemination capabilities between nodes and node groups. In the present invention, we obtain the message delivery probability p _ij between two nodes in the corresponding online social network per unit time. Then, for a node group composed of multiple nodes, the calculation method of the propagation ability between a node i and the node group Uv is: assuming that there are n nodes in this group that have connection edges with i, then the distance between node i and the group Uv Propagation probability is: SU(i,U _V )=1-(1-pi _j ) ⁿ , where SU refers to the message delivery capability of a group, U _V refers to a divided group,

Furthermore, the message transmission capability within a node group is defined as the average value of the transmission capability of each node to the group composed of other nodes in the group, which is:

${\mathrm{<span\; class="notranslate">\; S</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}<span\; class="notranslate">\; =</span>\frac{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; \&Element;</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}\mathrm{<span\; class="notranslate">\; S</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; ,</span>{\mathrm{<span\; class="notranslate">\; u</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; =</span>\frac{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; \&Element;</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>{<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; j</span>}}<span\; class="notranslate">\; )</span>}^{\mathrm{<span\; class="notranslate">\; q</span>}}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; \&ap;</span>\frac{{<span\; class="notranslate">\; \&Sigma;</span>}_{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; \&Element;</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}}{\mathrm{<span\; class="notranslate">\; qp</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; j</span>}}}{\mathrm{<span\; class="notranslate">\; no</span>}}\frac{\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; p</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; j</span>}}}{<span\; class="notranslate">\; |</span>{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; u</span>}}<span\; class="notranslate">\; |</span>}<span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 2</span>}{\mathrm{<span\; class="notranslate">\; np</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}\mathrm{<span\; class="notranslate">\; j</span>}}}{\mathrm{<span\; class="notranslate">\; no</span>}}<span\; class="notranslate">\; ,</span>$

Among them, S _U represents the message delivery capability of the group, S(i, U _Ui ) refers to the message delivery capability of the group composed of node i and nodes in the group U other than i, n refers to the number of nodes in the group U, Vu is the set of all nodes in group U, Eu refers to the set of all edges in group U, |E _U | refers to the number of edges in U, |V _U | refers to the number of nodes in U, m refers to the edges in Eu quantity.

3. Divide to obtain the first-level user association structure group

In the present invention, we divide user association structure groups according to node propagation capabilities. Our structure group division idea is to divide user nodes with frequent message transmission into the same structure group according to the user message transmission probability, that is, try to make the message transmission ability within the group strong and the message transmission ability between groups low. In this way, it can be ensured that when an attack occurs, it is relatively easy to limit the damage to a certain area. The division method of the present invention is implemented in two stages. The division method of the first-level user association structure group is:

1) Find the node v _i with the largest weight in the social network user association graph G, add v _i and its adjacent nodes that have not joined any group to the group U, and calculate the message delivery probability S(U) in the group U , and build a boundary set B(v) for all the nodes adjacent to the group U, the adjacent nodes are the two nodes directly connected by the connecting edge;

2) For the nodes in B(v) adjacent to the group U, select the node v _j with the largest weight among them, and calculate the message delivery probability S(U+j) of the node j to join the group U after joining the group U, if S(U+j)>S(U), then node j will be added to group U, and B(v) will be extended. This calculation is performed until no new adjacent nodes can join the group U.

3) Repeat steps 1) and 2) until all nodes in graph G are divided into groups.

4. Divide to obtain the second-level user association structure group

Since there are some popular user nodes in the social network (popular user nodes are associated users, that is, user nodes with a lot of friends), these users are often related to each other, and it is inevitable that some group nodes are too concentrated and large. , in this case, in order to further control the spread of threats, the present invention divides the second-level user association structure group, and the specific method is:

1) In the first-level user association structure group, find the node v _i with the largest weight, and then set the ratio value r according to the needs, find all nodes with weight w _vj >r*w _vi in the group, and for each node that satisfies the If the conditions are not satisfied, the new group will not be constructed; among them, w _vi is the weight of the node with the largest weight in the first-level user association structure group, and w _vj is any node in the first-level user association structure group. the weight of node v _j ;

The ratio value r is set according to the actual situation in actual operation. The larger r is, the less area will be divided in the end, that is, when the bandwidth and other conditions are sufficient, when it is possible to send as much as possible at one time during repair, it can divide into several structural groups, and then send more repairs at the same time, and repair If the various conditions do not permit, of course we will divide some areas;

2) According to the new group obtained above, according to the division method of the first-level user association structure group, calculate the propagation ability of their adjacent nodes joining the group and not joining the group, and complete the second level in the first-level user association structure group For the division of user association structure groups, it should be noted that when a node is connected to two groups at the same time and meets the conditions for joining a group, it will be added to the calculated group with a larger S.

5. Identify and mark critical and boundary nodes

After completing the second-level division, the node with the largest weight in each group is marked as a key node, and according to the boundary set B (each group is marked as a boundary node with the nodes adjacent to other groups.

6. When repairing is required, the key nodes and boundary nodes should be repaired first

When a vulnerability occurs or the supply needs to be repaired, the order of repair is first to send patches to key nodes and border nodes for repair. In this way, repairing the border nodes first can avoid the large-scale spread of attacks between groups, and repairing key nodes can In the group, try to effectively control the nodes with the greatest influence and the strongest communication ability.

7. Repair other nodes step by step

Then the key nodes and border nodes gradually send the repair information to their adjacent nodes, and use the social network user relationship to transfer the patch and so on to complete the repair.

Claims (9)

1. A kind of online social network rapid repair method based on user association structure division, its step comprises: 1) Construct a user association graph G on the mobile smart terminal according to the user relationship data in the online social network; 2) calculating the message dissemination capability of nodes and node groups in the graph G; 3) dividing the graph G into first-level user association structure groups according to the message dissemination capability, and performing second-level user association structure group division on the basis of the first-level user association structure groups; The method for dividing the first-level user association structure group is: a) Find the node v _i with the largest weight in the social network user association graph G, add v _i and its adjacent nodes that have not joined any group to the group U, and calculate the message delivery probability S(U) in the group U , and build a boundary set B(v) for all nodes adjacent to the group U; b) For the nodes in B(v) adjacent to the group U, select the node v _j with the largest weight among them, and calculate the message delivery probability S(U+j) of the node j to join the group U after joining the group U, if S(U+j)>S(U), then add node j to group U, and expand B(v) until no new adjacent nodes can join group U; c) traverse the steps a)-b) and divide all the nodes in the graph G into groups; 4) marking the node with the largest weight as a key node in the first-level user association structure group and/or the second-level user association structure group, and marking the adjacent nodes between the groups as boundary nodes; 5) Follow the quick repair method of sending patches to key nodes and/or border nodes first, and then gradually repair other nodes to complete the repair. 2. the online social network rapid repair method based on user association structure division as claimed in claim 1, is characterized in that, the message dissemination ability of described group U where n is the number of nodes in U, m is the number of connecting edges, and p _ij is the message propagation capability of nodes. 3. The online social network fast repair method based on user association structure division as claimed in claim 2, characterized in that, said p _ij is expressed by obtaining the message propagation probability between nodes in social network unit time. 4. the online social network rapid repair method based on user association structure division as claimed in claim 1, is characterized in that, according to described message dissemination ability, the user node with frequent message delivery is divided into the same structure group, so that the users in the group The messaging capability is strong, and the group-to-group messaging capability is low. 5. the online social network rapid repair method based on user association structure division as claimed in claim 1, is characterized in that, described mobile intelligent terminal is smart phone, Pad, PDA. 6. The online social network quick repair method based on user association structure division as claimed in claim 1, is characterized in that, described user relation data comprises user, the friend association relation between users, the number of user friends, the connection between users. 7. the online social network rapid repair method based on user association structure division as claimed in claim 1, is characterized in that, described construction user association graph G={V, E, W _V }, wherein node V represents user, edge E represents the relationship between users, W _V is the node weight, and the number of associated friends of the user is the node weight. 8. the online social network rapid repair method based on user association structure division as claimed in claim 1, is characterized in that, the concrete method of second level user association structure group division is: 1) In the first-level user association structure group, find the node v _i with the largest weight, and then set the ratio value r according to the needs, find all the nodes in the group with weight w _vj >r*w _vi , for each The nodes satisfying this condition build a new node group; wherein, w _vi is the weight of the node with the largest weight in the first-level user association structure group, and w _vj is the weight of any node v _j in the first-level user association structure group; 2) According to the new node group obtained above, according to the division method of the first-level user association structure group, calculate the communication capabilities of their adjacent nodes joining the group and not joining the group, and complete the second in the first-level user association structure group. The division of user-level user association structure groups. 9. The online social network fast repair method based on user association structure division as claimed in claim 8, is characterized in that, when a node is connected with two groups simultaneously and satisfies the condition of joining the group, it is added to the calculated message In the group with larger transmission probability S.