WO2021248740A1 - Mimic router execution entity scheduling method, and mimic router - Google Patents

Abstract

Disclosed is a mimic router execution entity scheduling method and a mimic router, pertaining to the technical field of network communications. The invention aims to solve the problem of the prior art in which scheduling policies that are based on the reliability of execution entities only consider the reliability of each single execution entity and do not take into account the reliability of the entire system or correlations between execution entities, and accordingly adds two decision-making factors to existing scheduling policies, that is, the reliability of an execution entity group and correlations between execution entities. The method comprises: acquiring states and data information of execution entities; calculating respective reliability levels of the execution entities; calculating, according to the reliability levels of the execution entities, reliability levels of execution entity groups obtained by randomly combining execution entities having a reliability level greater than a threshold, and calculating correlations between the execution entities; generating a scheduling policy by integrating the above calculations; and lastly determining online or offline conditions of the execution entities. The invention improves the accuracy of detecting abnormal information in a system, thereby improving system security.

Description

A mimic router executive body scheduling method and mimic router Technical field

The invention relates to the technical field of network communication, and more specifically, to a mimic router executive body scheduling method and a mimic router.

Background technique

The cyberspace domain is the same as nature, including known unknown risks and unknown unknown threats. In the face of these cyber threats, vulnerabilities and backdoors, the commonly used cyber security defense technologies include authentication technology, digital encryption technology, firewall technology, and intrusion. Detection system, virtual private network technology, hacker deception technology, honeypot, etc. However, with the advent of the "Internet of Everything", cloud computing, and big data era, the continuous advancement of microelectronics and virtualization technologies, and the continuous improvement of cyberspace security defense performance, mimic defense theory has emerged. The mimic defense theory is a highly reliable, highly available, and highly credible information system. At present, the mimic defense theory is widely used in network core equipment such as storage, DNS servers, switches, routers, and WEB servers. The three important characteristics of mimic defense structure are dynamics, diversity, and redundancy, that is, to construct multiple functionally equivalent and structured multi-executive environments, and make full use of dynamics, Diversity and randomness hide various "dark functions" in the executive body, increase the difficulty for attackers to scan system vulnerabilities and backdoors, and make it difficult for attackers to establish a continuous and reliable attack chain.

The router is an important device connected to the network and plays the role of network data packet forwarding. Once the router is attacked, the network data is very likely to be tampered with or lost, which poses a great threat to the security of the network. In the face of security problems such as routers, such as deep hidden backdoors and many available loopholes, a mimic structure is added on the basis of traditional routers, that is, mimic routers. The mimic router adopts a dynamic heterogeneous redundancy mechanism, inheriting the three characteristics of mimic dynamics, heterogeneity, and redundancy. The key point of the three characteristics is how to achieve dynamics in the face of attacks, that is, to change the executive body presented to the outside from time to time. .

In the mimic router, the realization of the dynamics and diversity of the executive body is in the scheduling unit. The main function of the dynamic scheduling unit for the network scheduling unit is to manage the operation of the heterogeneous execution body pool and its function subpools. According to the scheduling strategy specified by the decision-making unit, it can schedule multiple heterogeneous function execution bodies to realize the dynamics of the function execution body. Sex and diversity increase the difficulty of scanning and discovery by attackers, and hide the visibility of unknown vulnerabilities and backdoors.

The key link in the design of dynamic scheduling unit is the scheduling strategy of the executive body. Current scheduling strategies, such as patent applications CN201611010128.1 or CN201910148526.7, are based on two random scheduling strategies based on executive credibility and executive weight. The scheduling strategy based on executive credibility currently only considers the credibility of a single executive It does not consider the credibility of the entire system and the correlation between the executive bodies, and the overall stability and security are not enough.

Summary of the invention

1. Technical problems to be solved

In view of the existing scheduling strategy based on the credibility of the executive body in the prior art, only the credibility of a single executive body is considered, and the credibility of the entire system and the correlation between the executive bodies are not considered. The present invention provides a The mimic router executive body scheduling method and the mimic router increase the credibility of the executive body group and the correlation between the executive bodies, improve the accuracy of the system to monitor abnormal information, and improve the security of the system.

2. Technical solution

The purpose of the present invention is achieved through the following technical solutions.

A mimic router executive body scheduling method includes the following steps:

Step 1: After the mimetic router is started, the executive body goes online, and the executive body data information is obtained. The executive body is divided into ready state, working state and abnormal state; according to the execution degree attacked or working situation, the execution body will be in three types after going online. Switch between states;

Step 2: Calculate the credibility of the executive body based on the number of times the executive body has been attacked and the online duration of the executive body;

Step 3: Calculate the correlation between the executive bodies according to the proportion of the executive bodies being attacked at the same time;

Step 4: Multiple executives with credibility and relevance that meet the conditions form an executive combination, calculate the average credibility of the executive combination, and calculate the average credibility of the executive combination to meet the condition. Combination, calculate the correlation coefficient of the executive body combination;

Step 5: According to the calculation results in step 2 to step 4, the decision execution body goes online or offline.

Furthermore, in step 5, the scheduling strategy of the execution entity going online or offline is: selecting the execution entity and/or the combination of the execution entity with a high credibility value and a small correlation coefficient as the work execution entity. The executive body of the working state is selected through the credibility calculation, and the reliability of the entire scheduling system is considered. When scheduling, the executive body with similar structure, strong correlation and low average credibility should not be used as the working executive body at the same time. That is to say, when selecting the executive body as the working state executive body, select the executive body with high credibility according to the calculation result. The high credibility of the executive body indicates that the executive body is less attacked and the state is stable and more reliable; correlation coefficient The small description shows that the structural similarity between the executive bodies is low, and they can deal with many different types of attacks at the same time.

Furthermore, when calculating the relevance of the executive body combination in step 4, it is necessary to first calculate the credibility average M of the executive body combination.

Where 1≤k≤j, j is the number of the specified number of work executives, and R _i is the credibility value of a single executive; if the credibility average of the executive combination is lower than the critical credibility average, reselect Execution body combination; calculate the correlation of all execution body combinations that meet the credibility condition, that is, for all execution body combinations that meet the credibility condition, calculate each execution body in each execution body combination and other executions For the correlation between entities, select the maximum correlation coefficient from it to form a combination of the maximum correlation coefficient, and select the executive combination with the smallest correlation coefficient from the combination of the maximum correlation coefficient as the work executive combination.

When the average credibility of the executive body combination is lower than the critical credibility average, the current scheduling system is considered to be unreliable, and the work executive is re-selected. Otherwise, the current scheduling system is considered reliable. When multiple executive body combinations meet the credibility conditions and judged to be a reliable executive body combination, the correlation of each executive body combination is calculated, that is, the relationship between each executive body in each executive body combination and the other executive bodies is calculated separately For the correlation, select the maximum value of the correlation coefficient among all the execution bodies in the combination of execution bodies, and the maximum value of the correlation coefficient of each execution body combination forms a combination of the maximum correlation coefficient, and select from the combination of the maximum correlation coefficient The executive body combination with the smallest correlation coefficient is regarded as the work executive body combination, and the executive body combination is judged to be the optimal executive body combination.

Still further, the step of calculating the reliability of the formula R _i is 2

Wherein ε is the infinitely small value, t _i being the number of attacks on the line for the period, T _i is the total length on the line, i is the number of executable id. The greater the value of R _i , the greater the credibility of the executive body, on the contrary _{, the smaller the value of R i} , the smaller the credibility of the executive body.

Furthermore, the calculation formula of _{correlation q AB in step 3 is}

Among them, T _AB is the time when the executive body A and the executive body B are in the working state at the same time, and t _AB is the number of times that the time difference between the executive body A and the executive body B is less than the time critical difference Δt. When calculating Δt as small as possible. If T _AB =0 or t _AB =0, it is considered that executive body A and executive body B are not related, and q _AB =0. Set a correlation coefficient critical value q, if the correlation coefficient of q _AB ≥ q executive body combination is large, then in the scheduling process, A and B will not be used as work executive bodies at the same time, otherwise, A and B can be used as work executive bodies at the same time body.

Furthermore, an executive whose credibility is lower than the critical credibility value requires offline cleaning and data rollback. At this time, the executive is placed in an abnormal state. The executive body in the abnormal state no longer participates in the work. In order to improve the safety of the system, the execution body in the abnormal state is implemented with a cleaning strategy. Setting a threshold confidence executable offline cleaning is R, R _i when the executable reliability when R is equal to less than 10, this body is set to perform an abnormal state waiting for offline cleaning, data rewind operation; when When R _{i is} greater than R or after cleaning, the executive body is placed in a ready state.

Furthermore, when selecting the execution body as the work execution body, the execution body in the ready state is preferentially selected. If there is no executive body in the ready state in the state system, select among the executive bodies in the working state. The executive body in the ready state is generally just launched or cleaned, and has high reliability. For example, there is no executive body in the ready state in the system during scheduling, or the executive body in the ready state does not meet the conditions of credibility or relevance, then it is in the working state To choose from the executive body, the credibility and relevance of the executive body need to be calculated at this time.

Furthermore, after the mimic router is started, it selects the executive body as the working state according to the online sequence of the executive body, and the remaining executive bodies are placed in the ready state. After the execution body is online, according to the work requirements, all the execution bodies may not be placed in the working state. At this time, the execution bodies that have been successfully online and have not been placed in the working state are placed in the ready state.

Furthermore, in the scheduling cycle: if the work executive is attacked, the work executive is selected by calculating the credibility and correlation; if the work executive is not attacked, the credibility and correlation are calculated after the scheduling period ends. Select the executive body as the work executive body. In a scheduling cycle, if the executive body is not attacked, or the credibility and relevance of the executive body still meet the conditions after the attack, the scheduling will be performed at the end of the current scheduling cycle; if the executive body is attacked, pass the credibility and The calculation of the correlation is scheduled, and the work execution body is reselected.

In the scheduling method of the present invention, when the credibility value of the executive body reaches the credibility critical value, the status of the executive body is modified to an abnormal state, and then the abnormal state executive body is offline cleaned and data rolled back, and the cleaned executive body is set to Ready state, when the execution body of the working state is attacked or the cycle time arrives, the state of the execution body is adjusted according to the scheduling strategy and the credibility of the execution body is calculated to increase the credibility of the execution body group and the correlation between the execution bodies. The reliability of the executive body group guarantees the reliability of the executive body set in a working state during a scheduling period, and the correlation between the executive bodies prevents the executive bodies with strong scheduling correlation from working at the same time, so as to make full use of the heterogeneous characteristics of the executive bodies to ensure System security, improve the accuracy of system monitoring abnormal information.

A mimic router includes a dynamic scheduling subsystem, a route arbitration subsystem, a protocol proxy subsystem, and a configuration management protocol proxy subsystem. The dynamic scheduling subsystem uses the mimic router executive scheduling method in combination. The scheduling subsystem dynamically adjusts the status of the executive body and sends the status of the executive body to each subsystem. Each subsystem determines whether the executive body participates in the business ruling based on the status information of the executive body. After the ruling, the executive body that produces inconsistent results will be Relevant information is fed back to the scheduling subsystem for the scheduling subsystem to determine the credibility of the executive body.

In a mimic router, an executive body is highly likely to be compromised by an attacker, but multiple executive bodies with the same function and structure but different structures are less likely to be compromised by the same attack method at the same time. Therefore, in the face of an attacker’s attack, increase the system The performance of active defense adopts the organizational form of heterogeneous groups. Heterogeneous groups can be composed of multiple virtual machines, multi-vendor routing simulation software, or a combination of virtual and real devices. The scheduling method of the present invention performs scheduling of work executives by calculating the credibility of the executive body, the credibility of the heterogeneous groups and the correlation between the executive bodies, which better reflects the dynamics of the heterogeneous groups and improves the system performance safety.

3. Beneficial effects

Compared with the prior art, the advantages of the present invention are:

On the basis of the existing scheduling strategy to calculate the credibility of a single executive body, the present invention increases the calculation and comprehensive consideration of the two determinants of the correlation between the executive bodies and the credibility of the executive body group. The risk of the attack, calculate the credibility value of the heterogeneous executor, the credibility value of the heterogeneous group, and the correlation between the heterogeneous executors. The system finally decides whether the executor is subject to an abnormal attack or cycle based on the above calculation data. The scheduling strategy in the case of flexible scheduling.

On the premise of calculating the credibility of a single executive body, calculate the relevance of the executive body to avoid the probability of putting together highly correlated executive bodies; the credibility calculation of the executive body group ensures that it is in a working state during a scheduling period The average value of the credibility of the executive body set not only ensures the safety of a single executive body, but also ensures the credibility and relevance of the combined executive body, that is, the executive body group, from the overall consideration. Through the scheduling method of the present invention, the scheduling strategy is determined and calculated from the perspective of the correlation between the executive bodies and the executive body group, which improves the accuracy of the mimic router to monitor abnormal information and also improves the safety of the system.

Description of the drawings

Figure 1 is a block diagram of the scheduling strategy of the present invention;

Figure 2 is a state switching diagram of the executive body of the present invention;

Figure 3 is a flow chart of the scheduling of the present invention.

detailed description

The present invention will be described in detail below in conjunction with the drawings and specific embodiments of the specification.

Example 1

This embodiment first introduces a mimic router, including dynamic scheduling subsystems, routing arbitration subsystems (ipv4, ipv6), protocol proxy subsystems (bgp, ospf, and other protocols), and configuration management protocol proxy subsystems (netconf, telnet, ssh) , Snmp and other protocols), the dynamic scheduling subsystem dynamically adjusts the status of the executive body, and sends the status of the executive body to the routing arbitration subsystem, the protocol proxy subsystem, and the configuration management subsystem. Each subsystem is based on the status information of the executive body. Decide whether the executive body participates in the business adjudication. After the adjudication, for the executive body that produces inconsistent results, the relevant information is fed back to the dynamic scheduling subsystem for the dynamic scheduling subsystem to determine the feasibility of the executive body.

The following describes in detail the scheduling method used by the scheduling subsystem in the mimic router of this embodiment.

After the executive body goes online, there are three states: ready state, working state and abnormal state.

Ready state: After the mimic defense system is started, the executive body that has been successfully started and has not participated in the scheduling of the various modules of the system, the executive body after offline cleaning from the abnormal state, and the executive body whose credibility value is greater than the credibility threshold are all set to Ready state.

Working status: The executive bodies participating in the arbitration and scheduling of the subsystems of the system are all set to working status.

Abnormal state: Execution entities whose credibility value is less than or equal to the credibility critical value, and those that are abnormal due to external factors such as software and hardware are all set to an abnormal state.

After the mimic defense system is activated, the executive body also starts to go online, and the online executive body realizes the switching of three states under corresponding circumstances. As shown in Figure 2, the executive body is in the working state during normal work. When the reliability of the executive body is higher than the critical value after a scheduling period, it switches to the ready state and waits for the next cycle of the scheduling strategy to continue working; when the executive body is available The executive body switches to the abnormal state after the reliability reaches the critical value or external factors interfere with the abnormal state. The executive body in the abnormal state will undergo offline cleaning and data rollback. The cleaned executive body will be placed in the ready state, and the executive body in the ready state will be set to the ready state. After the calculation is in accordance with the scheduling strategy, it is converted to a working state.

The dynamic scheduling strategy is shown in Figure 1, and is specifically divided into the following steps:

Step 1: Obtain the executive body data information sent on each subsystem;

After the initial startup of the system, the executive bodies go online one by one, and the scheduling system chooses to put the executive bodies in the working state or the ready state according to the task requirements. At this time, if the executive body is attacked and an abnormality occurs, the data and information related to the abnormal state are collected for later calculation of the credibility and correlation of the executive body.

Step 2: Calculate the credibility of the abnormal executive body;

According to the number of times the work executive has been attacked and the length of time it has been online, calculate the credibility of the executive after being attacked, and set a credibility threshold for the off-line cleaning of the executive. When the credibility is less than or equal to the critical value, the executive Set to an abnormal state for offline cleaning and data rollback; when the credibility is greater than the critical value or after cleaning, the executive is placed in the ready state. When the execution body in the working state is abnormal, the execution body with high reliability is selected as the working execution body in the working state or the ready state.

Step 3: Calculate the correlation between the executive bodies;

Each executive body is independent and heterogeneous. There is no correlation between the executive bodies, and the difficulty of being attacked is also different. Therefore, considering the reliability of the entire system, try not to make the structure similar when scheduling, and the correlation is strong. The executive body at the same time as the work executive body. In order to find the possible similarities between executive bodies, you can conduct multiple scheduling experiments and summarize the regular characteristics of each executive body or between the executive bodies being attacked.

In multiple scheduling experiments, record the executor id that is working at the same time during each experiment and the executor id that is attacked. According to the recorded multiple sets of data, calculate the execution body combination that is attacked at the same time in multiple scheduling experiments. Proportion, record the online duration of the executive IDs that are working at the same time and the number of online executions of the executive IDs that are simultaneously attacked by the same type, so as to find out the possible similarities between the executives. If the proportion of multiple executives being attacked at the same time is large, the correlation coefficient of the combination of executives is relatively large, and it can be used as a work executive while avoiding correlation during scheduling.

Step 4: Calculate the credibility of heterogeneous groups;

The average credibility of all work executives also affects the reliability of the scheduling system. Through step 3, the correlation of executives is calculated, and which executives are not suitable to be work executives at the same time. Under this premise, the calculation of the work executive group Reliability and correlation coefficient.

Calculate the credibility of the work executor group according to the average credibility of each executor in the work executor group. When the calculated credibility of the work executor group is greater than or equal to the credibility threshold, determine the work executor The group is reliable, otherwise the judgment is unreliable. When the group of work executives is judged to be unreliable, re-select the executives in the working state and the ready state and calculate the credibility of the group, and find out the group of executives that meet the credibility of the group to be greater than or equal to the critical value of the credibility.

For the executive body combination that meets the group credibility of the work executive body, calculate the maximum value of the correlation coefficient of the combination, and find the executive body combination that meets the critical value requirements of the group credibility and has the least correlation, that is, the best executive body combination as the work executive body The executive body combination.

Step 5: Integrate the scheduling strategy from step 2 to step 4, and schedule the online or offline of the decision-making executive body. Combining the executive body credibility calculated in step 2, the executive body correlation coefficient calculated in step 3, and the work executive group credibility and correlation coefficient calculated in step 4, select the best executive body combination work.

The following specifically introduces the scheduling method described in this embodiment:

The scheduling system obtains the abnormal data information sent by other subsystems on the mimic router, and calculates the credibility of the executive body, the credibility of the heterogeneous group and the correlation between the executive bodies according to the abnormal data information of the executive body, and the correlation between the executive bodies is calculated according to the calculated The value is generated by the scheduling system to generate the corresponding scheduling strategy.

The executive body is divided into three data sets according to different states, among which:

c ₁ : Execution body data set in the ready state c ₁ ={m ₁₁ ,m ₁₂ ,m ₁₃ ,...,m _1n }, including n execution bodies in the ready state;

c ₂ : Execution body data set in working state c ₂ ={m ₂₁ ,m ₂₂ ,m ₂₃ ,...,m _2j }, including j execution bodies in working state;

c ₃ : Execution body data set in abnormal state c ₃ ={m ₃₁ ,m ₃₂ ,m ₃₃ ,...,m _3k }, including k execution bodies in abnormal state;

N: is all the online and offline executives in the executive pool, n+j+k≤N.

The scheduling method shown in Figure 3, after the mimic router is initially started, a certain number of executives are taken from the executive pool and placed in the working state set c ₂ according to the following rules, and the states of all the executives in _{c 2} All set to working state. The activation and ready status of the work state executive body are set by the following rules:

(i) Within the specified time, if all the execution bodies in the execution body pool have been successfully launched, according to the time sequence of the execution bodies going online in turn, the specified number of execution bodies will be selected as the execution bodies in the working state, and the remaining execution bodies Set the state to the ready state and put it into c ₁ ;

(ii) Within the specified time, if the number of successfully online executions does not reach the stipulation, the executions that are actually successfully online in the execution pool will be used as the executions of the working state. If the remaining executions are successfully online, they will be placed in the ready state. c ₁ .

After the system is running, periodic scheduling or scheduling under abnormal conditions is carried out according to the attack situation of the executive body during the working period. The periodic scheduling is to set a scheduling period T. During the operation of the system, if all work execution entities are not attacked, after the scheduling time T has elapsed, the work execution entities are reselected. The scheduling in the abnormal situation means that if the executive body in the current working state is attacked, the working executive body must be reselected. Regardless of whether it is periodic scheduling or scheduling under abnormal conditions, the following scheduling strategy must be followed for scheduling.

The specific scheduling policy rules are as follows: if _{there is an exception in the executive body in c 2} or periodic scheduling is required, the executive body that has _{not worked in c 1} is preferentially selected as the working executive body. For the executive body that has participated in the scheduling, it is determined whether to continue as a work executive body by calculating the credibility of the executive body and the correlation between the executive bodies. The calculation rules for the credibility of the executive body, the correlation between the executive bodies, and the system credibility are as follows:

(1) Calculate the credibility of the executive body

First, record during the operation of the system, according to the number of times the work executive has been attacked and the online duration, calculate the reliability R _i of the executive after being attacked:

Wherein ε is the infinitely small value, t _i being the number of attacks on the line for the period, T _i is the total length on the line, i is the number of executable id.

The greater the value of R _i , the greater the credibility of the executive body, on the contrary _{, the smaller the value of R i} , the smaller the credibility of the executive body. Set the credibility critical value of an executor offline cleaning to R, when R _{i is} less than or equal to R, put the executor into an abnormal state and put it in c ₃ for offline cleaning and data rollback operations; When R _{i is} greater than R or after cleaning, the executive is placed in the ready state and placed in c ₁ ; when the _{work executive in c 2} is abnormal, the executive in _{c 1} and c ₂ is selected as the work according to the strategy Executive body.

(2) Calculate the correlation coefficient between executive bodies.

While the executive body is working, the types of attacks that may be received are different. Therefore, in order to ensure that the attack on the executive body is probably the same type of attack, the time difference of the execution body is limited, that is, the executive body _A is attacked at time t A, and B The executive body _{is attacked at t B.} In order to ensure that A and B are attacked by the same kind of attack, add a time critical difference Δt (Δt is as small as possible), that is, if |t _A -t _B |≤Δt, then Think that A and B suffered the same kind of attack. In the next many experiments, count _{the time that A and B are in the c 2} set at the same time (assumed as T _AB ), and the number of times A and B are attacked and satisfy the condition |t _A -t _B |≤Δt (assumed as t _{AB ), the correlation coefficient q AB} after being attacked when A and B work at the same time, the _{calculation formula of q AB} is as follows:

Set a correlation coefficient critical value q, if the correlation coefficient of q _AB ≥ q executive body combination is large, then in the scheduling process, A and B will not be used as work executive bodies at the same time, otherwise, A and B can be used as work executive bodies at the same time body.

(3) Calculate the credibility of the group

The average credibility of all work executives also affects the reliability of the scheduling system. First, exclude the executives in the abnormal state set c ₃ that are to be cleaned offline, and calculate the correlation between the executives through step (2), and find which executives are not suitable to be work executives at the same time. Under this premise, the work executives are calculated. The credibility of the group and the correlation coefficient.

The specific algorithm is as follows: According to the credibility value of each executive body, calculate the average credibility M,

Among them, 1≤k≤j, j is the number of the specified number of work executives. Set W as the credibility critical value of the overall scheduling system. When M≥W, the scheduling system is considered to be reliable, otherwise it is considered unreliable, then _{reselect the executive body from c 1} and c ₂ and calculate the average credibility Degree, find all executive body combinations that meet the condition of M≥W.

Respectively calculate the correlation of _{the executive body combinations M 1} , M ₂ , M ₃ ,... under the above-mentioned credibility conditions. Get the maximum correlation coefficient of each executive body combination

i, j are the executor id (i≠j) in the combination, n is the number of executors in the combination, the correlation calculation of the executor combination is to calculate each executor in each executor combination and the other executors separately correlation between, that is to say when the constant i, is sequentially changed j, Q _i is calculated, and then change i, Q _i value is calculated sequentially. The maximum correlation coefficient of each executive body combination forms the maximum correlation coefficient combination. Select the smallest corresponding executive body combination in the maximum correlation coefficient combination as the work executive body combination, that is, calculate min(max(Q _i ) ₁ ,max(Q _i ) ₂ ,max(Q _i ) ₃ ,……), This combination is the best executive body combination with the highest reliability and stability.

Based on the existing scheduling strategy, this embodiment comprehensively calculates the credibility of the executive body, the correlation coefficient between the executive bodies and the average credibility between the executive body combinations, and the correlation coefficient between the executive bodies. The calculation avoids putting highly correlated executive bodies to work together, and the average credibility of the executive body combinations guarantees the average credibility of the executive body set in a working state during a scheduling period, and improves the accuracy of monitoring abnormal information Degree and the security of the system.

The above is a schematic description of the invention and its implementation. The description is not restrictive. The invention can be implemented in other specific forms without departing from the spirit or basic characteristics of the invention. What is shown in the drawings is only one of the embodiments created by the present invention, and the actual structure is not limited to this, and any reference signs in the claims should not limit the related claims. Therefore, if a person of ordinary skill in the art receives its enlightenment, and does not deviate from the purpose of this creation, without creative design, structural methods and embodiments similar to the technical solution should fall within the scope of protection of this patent. In addition, the word "comprising" does not exclude other elements or steps, and the word "a" before an element does not exclude the inclusion of "plurality" of the element. Multiple elements stated in the product claims can also be implemented by one element through software or hardware. Words such as first and second are used to denote names, but do not denote any specific order.

Claims (10)

1. A mimic router executive body scheduling method is characterized in that it comprises the following steps:

   Step 1: After the mimic router is started, the executive body goes online to obtain the executive body data information, and the executive body is divided into a ready state, a working state, and an abnormal state;

   Step 2: Calculate the credibility of the executive body based on the number of times the executive body has been attacked and the online duration of the executive body;

   Step 3: Calculate the correlation between the executive bodies according to the proportion of the executive bodies being attacked at the same time;

   Step 4: Multiple executives with credibility and relevance that meet the conditions form an executive combination, calculate the average credibility of the executive combination, and calculate the average credibility of the executive combination to meet the condition. Combination, calculate the correlation coefficient of the executive body combination;

   Step 5: According to the calculation results in step 2 to step 4, the decision execution body goes online or offline.
2. A mimic router executive scheduling method according to claim 1, characterized in that the scheduling strategy of the executive going online or offline in step 5 is: selecting executives with high credibility value and low correlation coefficient and/ Or a combination of executive bodies as work executive bodies.
3. A mimic router executive scheduling method according to claim 2, wherein when calculating the relevance of the executive combination in step 4, it is necessary to first calculate the credibility average M of the executive combination,

   

   Where 1≤k≤j, j is the number of the specified number of work executives, and R _i is the credibility value of a single executive; if the credibility average of the executive combination is lower than the critical credibility average, reselect Execution body combination; calculate the correlation of all execution body combinations that meet the credibility condition, that is, for all execution body combinations that meet the credibility condition, calculate each execution body in each execution body combination and other executions For the correlation between entities, select the maximum correlation coefficient from it to form a combination of the maximum correlation coefficient, and select the executive combination with the smallest correlation coefficient from the combination of the maximum correlation coefficient as the work executive combination.
4. A mimetic router executive scheduling method according to claim 1, wherein the formula for calculating the _{credibility R i in step 2 is}

   

   Wherein ε is the infinitely small value, t _i being the number of attacks on the line for the period, T _i is the total length on the line, i is the number of executable id.
5. A mimic router executive scheduling method according to claim 1, wherein the calculation formula of the _{correlation q AB in step 3 is}

   

   Among them, T _AB is the time when the executive body A and the executive body B are in the working state at the same time, and t _AB is the number of times that the time difference between the executive body A and the executive body B is less than the time critical difference Δt.
6. A mimic router executive scheduling method according to claim 1, wherein the executive whose credibility is lower than the critical credibility value requires offline cleaning and data rollback. At this time, the executive is set to Abnormal state.
7. A mimic router executive scheduling method according to claim 2, characterized in that when the executive is selected as the work executive, the executive in the ready state is preferentially selected.
8. A mimic router executive scheduling method according to claim 7, characterized in that after the mimic router is started, the executives are selected as the working state according to the online order of the executives, and the remaining executives are placed in the ready state.
9. A mimic router executive scheduling method according to claim 2, characterized in that, in the scheduling period: if the work executive is attacked, the work executive is selected by calculating the credibility and correlation; if the work executive is not Under attack, after the scheduling period ends, the executive body is selected as the work executive body by calculating the credibility and correlation.
10. A mimic router, comprising a dynamic scheduling subsystem, a routing arbitration subsystem, a protocol proxy subsystem, and a configuration management protocol proxy subsystem, wherein the dynamic scheduling subsystem uses any combination of any one as claimed in claims 1-9 A mimic router executive scheduling method described.

Images