CN107819628B - A kind of certificate server escape system and method based on the backup of reciprocal loop - Google Patents

Abstract

The present invention relates to an authentication server circular escape system, comprising a server group S={S ₁ , S ₂ ... S _N }, wherein the server group S includes N interconnected authentication servers S ₁ , S ₂ ... S _N ; each authentication server S _i in the server group S has a first backup server S _j and a second backup server _{S k} in the server group S; The first backup server, and the second backup server of a certain authentication server S _q ; the first backup relationship of all authentication servers in the server group S forms a ring, and the second backup relationship forms an inverse ring of the first backup relationship. Through the authentication server escape system and method, it is possible to effectively and quickly restore business applications and improve the ability to resist disaster strikes, while reducing the topology complexity of the escape system.

Description

An authentication server escape system and method based on reciprocal loop backup

technical field

The invention relates to the technical field of authentication servers, in particular to an authentication server escape method and system.

Background technique

With the development of society and the advancement of technology, computer networks play an increasingly important role in daily life. The authentication server is an important node in the network system, and the security of the authentication server network system is increasingly prominent. For example, in the authentication network of the power network, 802.1x authentication is adopted. This authentication method generally adopts dual-machine hot backup. If one authentication server fails, the remaining authentication server can take over from the failed device to provide corresponding services. If the authentication servers are all paralyzed, then start the escape server.

However, in the current power communication network system, the two authentication servers and the escape server generally have the same physical location in the computer room (for example, in the computer room of the same county or city), so when the computer room environment is not ideal, the authentication server and The escape server will lose all or part of its ability to work, which will have a major impact on the continuity of network communication. Moreover, in order to ensure the smooth flow of network communication, at least three servers are required to implement authentication and escape functions in the power network system, namely the authentication server, backup authentication server and escape server, resulting in high authentication costs.

The application number submitted by the State Grid System is 201710388516.1, and the technical solution introduced in the patent application titled "An Authentication Server Ring Escape System and Method" can effectively improve the anti-risk ability of the authentication server system. When the authentication server is hit, Business usage can be restored quickly. However, the topology structure of the server group S and the setting of the escape strategy in this technical solution are relatively complicated and difficult to implement.

In this application, the entire content of the patent application with application number 201710388516.1 is incorporated.

Contents of the invention

In order to solve the above problems, the present invention relates to a circular escape system for authentication servers, which includes a server group S={S ₁ , S ₂ ... S _N }, the server group S includes N authentication servers S ₁ connected to each other, S ₂ ... S _N ; each authentication server S _i in the server group S has a first backup server S _j and a second backup server S _k in the server group S; The first backup server of the authentication server S _p , and the second backup server of a certain authentication server S _q ; the first backup relationship of all authentication servers in the server group S forms a ring, and the second backup relationship constitutes the inverse of the first backup relationship. ring. Through the authentication server escape system and method, it is possible to effectively and quickly restore business applications and improve the ability to resist disaster strikes, while reducing the topology complexity of the escape system.

The invention also provides a circular escape method for the authentication server.

Description of drawings

Fig. 1 is a schematic structural diagram of an authentication server circular escape system;

Fig. 2 is a flowchart of an authentication method for authenticating a device to be authenticated accessing a network by using an authentication server.

Detailed ways

In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail with reference to the accompanying drawings. This description presents, by way of illustration and not limitation, specific embodiments consistent with the principles of the invention in sufficient detail to enable those skilled in the art to practice the invention without departing from this disclosure. Other embodiments may be utilized and the structure of elements may be changed and/or substituted within the scope and spirit of the invention. Accordingly, the following detailed description should not be read in a limiting sense.

As shown in Fig. 1, according to one aspect of the present invention, a circular escape system for authentication servers is provided, including a server group S={S ₁ , S ₂ ... S _N }, and the server group S includes N ( For example N>=3, preferably N>=6) authentication servers S ₁ , S ₂ . . . _SN connected to each other.

In the schematic structural diagram shown in Figure 1, only the situation with 6 authentication servers (N=6) is shown, and in the specific implementation, the specific technical solution of the present invention is mainly introduced by taking 6 authentication servers as an example , but those skilled in the art know that the escape system that can also be realized when N takes other values will also fall within the protection scope of the present invention.

Each authentication _server S _i in the server group S has a first backup server S _j and a second backup server S _k in the server group S; backup server, and a second backup server of a certain authentication server S _q . The authentication server S2 shown in Figure 1 is an example, it has the first backup server S3 (pointed by the solid line arrow), also has the second backup server S1 (pointed by the dotted line arrow), and S2 is also the first backup server of S1 at the same time. The second backup server of S3.

The first backup relationship of all authentication servers in the server group S forms a ring, and the second backup relationships form an inverse ring of the first backup relationship. As shown in Figure 1, the first backup relationship is a ring {S1, S2, S3, S4, S5, S6, S1} formed by solid lines, and the second backup relationship is a ring formed by dotted lines {S1, S6, S5, S4, S3, S2, S1}. Obviously, the second backup relationship is an inverse ring of the first backup relationship.

It can be seen that as a general requirement, S _j and S _k are different authentication servers; S _j and S _q are the same authentication server; S _k and S _p are the same authentication server. Still taking the authentication server S2 shown in FIG. 1 as an example, the four related servers S3, S1, S1, and S3 satisfy the above relationship.

According to the present invention, the N authentication servers are physically separated by a certain distance. As shown in Figure 1 as an example, in the electric power communication network system, six authentication servers can be respectively placed in computer rooms in six different regions. When an authentication server (such as S2) fails to work, the terminal corresponding to S2 can be migrated to its first backup server S3 or second backup server S1 for authentication, since S2, S3 and S1 physically belong to three Different regions, so the probability of being unable to work at the same time (such as power failure or server software and hardware failure) is very low. In this way, the continuity of terminal network communication is guaranteed. In addition, since any authentication server (such as S2) has two backup servers, that is, the terminal that should be authenticated by S2 can be authenticated by three authentication servers, S2, S3 and S1, so in the circular escape system, the loss Any two authentication servers will not affect the authentication of the system.

In one implementation of the present invention, the authentication server (S2) includes two network cards, wherein the first network card is used for communication when authenticating the corresponding terminal, and the second network card is virtualized as two network cards. The virtual network card is used for communication when S2 is used as the terminal corresponding to the first backup server (S1) for authentication, and the second virtual network card is used for communication when S2 is used as the terminal corresponding to the second backup server (S3) for authentication. This design allows only one server with two network cards to be reserved in each physical computer room. Generally, each authentication server is responsible for the load of terminals in the corresponding area, and the backup server is used less frequently, so there is no need to set up backup separately Servers to save costs without compromising performance.

The circular escape system of the present invention also includes an escape server. When a large area of systemic authentication server fails to work in the server group S, the escape server starts to prompt the unauthenticated terminal to escape (that is, modify the authority of the network node device, allowing the terminal to directly log in to the local area network or the Internet without authentication). Obviously, for the N authentication servers in the server group S, only one escape server is needed.

According to the present invention, the communication connection paths among the authentication servers in the server group S are P={p ₁ , p _{2 .} . . p _N }, and each communication connection path p _i is used to connect two authentication servers. For example, in Fig. 1, p ₁ may be the communication connection path between S1 and S2, p ₂ may be the communication connection path between S1 and S2, ..., and so on.

The communication speed of each communication connection path in P={p ₁ ,p ₂ ...p _N } is B={b ₁ ,b ₂ ...b _N }, and those skilled in the art know that the communication speed can be adopted by each authentication server It can be characterized by any one or weighted combination of network loans, network hops, average network transmission speed in a specific period of history.

The physical distance between the authentication servers is D={d ₁ , d _{2 .} . . d _N }, preferably, the physical distance is a spatial linear physical distance.

According to the present invention, when setting the backup relationship of the server group S, according to the communication speed B and the physical distance D, when constructing the first and second backup relationships, the weighted weight W of the ring and reverse ring communication paths formed by it should be The value of is the highest, so as to ensure to the greatest extent that in the actual application process, the number of authentication servers that go down is the least. Obviously, due to the reversibility of the communication speed B and the physical distance D, the weighted weight of the circular communication path is consistent with the weighted weight of the inverse circular communication path.

According to the present invention, the weighted weight

Among them, min(D) and max(D) are the minimum and maximum physical distances between any two servers in the server group S, and min(B) is the minimum communication speed between any two servers in the server group S value.

According to the present invention, since the communication speed B and the physical distance D have different physical units, they are first normalized by the functions f1 and f2 when used, and the value ranges of the communication speed B and the physical distance D are limited to [ 1,2] to facilitate subsequent operations.

According to the present invention, the first type of authentication server is an authentication server that is easily disrupted by natural conditions, such as but not limited to an earthquake fault zone, a flood discharge area, or an area prone to landslides; the second type of authentication server is not easily An authentication server that is down due to interference from natural conditions.

The first and second types of authentication servers are designated by users, for example, by experts; or they are automatically obtained by the circular escape system, for example, according to the statistical data of historical earthquakes, floods and other natural disasters, or according to the geographical location and the earthquake fault zone and flood discharge area. Geographical comparisons are obtained.

According to the present invention, it is hoped that the first-type authentication servers do not form a backup relationship as much as possible, so if a certain communication path p _i is connected to two first-type authentication servers, the weight of the communication path will be set lower, further Yes, in the present invention, the weight of the communication path is only related to the physical distance of the communication path, so as to avoid simultaneous downtime of the two authentication servers connected by the communication path as much as possible. Correspondingly, according to the present invention, it is hoped that the backup relationship between the first type of authentication server and the second type of authentication server should be formed as much as possible, so for the communication path connecting the first type and the second type of authentication server, a higher weight is set; and the weight is also There is a positive correlation with the communication speed, so that when the authentication server goes down, it can be migrated to the backup server as soon as possible. Finally, according to the present invention, for the two second-type authentication servers, it is mainly considered that they can migrate as soon as possible when they are down, so the communication speed is used as the main basis for weight setting, and fine-tuning is performed according to the distance between the two second-type authentication servers. It is hoped that the distance between the two will be avoided to a certain extent, and at the same time, it will be disturbed by natural conditions.

Compared with the background patent application with application number 201710388516.1, in the ring escape system of the present invention, on the one hand, the ring and the reverse ring are in one-to-one correspondence, which simplifies the topology; on the other hand, when determining the ring backup relationship, through Weighted weights to emphasize avoiding multiple downtime authentication servers at the same time, instead of only emphasizing the migration speed from downtime servers to backup servers in the background patent application.

As shown in Figure 2, the present invention also provides a method for circular escape using the above-mentioned circular escape system, comprising the following steps:

Step S100, judging how many authentication servers in the server group are down at the same time, if it is judged that there is no downtime, go to step S200; if it is judged that one is down, go to step S300; If it is judged that there are three downtimes, execute step S500; if it is judged that there are four or more downtimes, then execute step S600.

Step S200, each authentication server authenticates the corresponding terminal.

Step S300, migrating the terminal of the downtime authentication server _Se to the backup server with a smaller load among the first and second backup servers of _Se for authentication. When an authentication server goes down, for example, server S2 goes down, the terminal corresponding to S2 is migrated to the terminal with a smaller load in the first backup server S3 or the second backup server S1, and after S2 resumes work, the The corresponding terminal migrates back to S2. The load is the number of terminals authenticated by the authentication server.

Those skilled in the art know that various existing technical solutions in the prior art can be used for the migration of the terminal, so how to do the migration will not be introduced in the present invention. It is also clear to those skilled in the art that the basis for judging the load can be calculated based on the rated load, or the actual load, or the difference between the rated load and the actual load, but is preferably calculated based on the actual load , that is, select the one with the smaller actual load in S3 and S1 to complete the migration of S2.

Step S400, judging whether the downtime authentication servers S _e and S _f have a backup relationship, that is, Se is the first or the second backup server is S _f ; if there is a backup relationship, then migrate the terminals of Se _and S _f to Authentication is performed in a backup server different from S _e and S _f ; if there is no backup relationship, execute step S450. Still take Figure 1 as an example. For example, when S1 and S2 are down, simply, the terminal corresponding to S1 cannot be migrated to S2 (because S2 is also down), and the terminal corresponding to S2 cannot be migrated to S1 (because S1 is also down. machine), therefore, only the terminal corresponding to S1 can be migrated to S6, and the terminal corresponding to S2 can be migrated to S3.

Step S450, migrate the terminal of S _e to the backup server with a smaller load among the first and second backup servers of S _e for authentication, and migrate the terminal of S _f to the first and second backup servers of S _f Authentication is performed on a smaller backup server. Still taking Figure 1 as an example, for example, when S1 and S4 are down, S1 can choose S2 and S6 for migration, and S4 can choose S3 and S5 for migration. Similar to the method in step S300, the one with a smaller load is selected as the migration object of S1 in S2 and S6, and the one with a smaller load is selected as the migration object of S4 in S3 and S5, so as to further achieve load balance.

According to one aspect of the present invention, preferably, the following steps are also performed in step S450:

Judging whether the authentication servers S _e and S _f have the same backup server, if not (for example, S1 and S4 are down), then proceed according to the above processing method; if there are the same backup servers, for example, when S1 and S3 are down, S1 The first backup server of S3 and the second part server of S3 are both S2. If the load of S2 is smaller than the second backup server S6 of S1 and smaller than the first backup server S4 of S3, it is easy to migrate the terminals of S1 and S3 to S2, thereby making the load of S2 too large.

Therefore, in this case, the present invention preferably requires that the backup servers for migration of S _e and S _f are not the same backup server. An optional way is that, when the load of S2 is less than the second backup server S6 of S1, and smaller than the first backup server S4 of S3, the load of S _e and S _f (that is, S1 and S3 in this example) is randomly selected. Select one terminal (such as S1) to migrate to S2, and another terminal (such as S3) to migrate to S4. But preferably, obtain the historical loads of S _e and S _f (ie S1 and S3 in this example), and migrate the terminal of an authentication server with a larger historical load to S2, so that the overall load of the escape system is more balanced .

According to one aspect of the present invention, the processing process of simultaneous downtime of three authentication servers is as follows:

Step S500, obtain the downtime authentication server set {S _u , S _v , S _w }, if the first and second backup servers S _v , _{S w} of a certain downtime authentication server Su are down (such as S1 , S2, S3 are both down, and at this time the first and second backup servers S1 and S3 of S2 are both down), then start the escape server, perform an escape operation on the terminal corresponding to _Su , and transfer the terminal corresponding to S _v , S _w The terminals are respectively migrated to corresponding authentication servers; otherwise, step S520 is performed.

Step S520, if there is an authentication _server {S _u , S _v } with a backup relationship _in { _{S u ,} S _v , S w }, then migrate the terminals of _Su and S _v to different The authentication is performed in the backup server; the terminal of S _w is migrated to the backup server with a smaller load among the first and second backup servers of S _w for authentication; otherwise, step S540 is performed. For example, S1, S2, and S5 are down. Since there is a backup relationship between S1 and S2, the terminal of S1 can only be migrated to S6, the terminal of S2 can be migrated to S3, and S5 will choose the one with the lighter load among S4 and S6 for migration. .

Step S540, the terminals of {S _u , S _v , S _w } are respectively migrated to the backup servers with a smaller load among the first and second backup servers of {S _u , S _v , S _w } for authentication.

According to another preferred aspect of the present invention, since the probability of simultaneous failure of three authentication servers in the circular escape system is extremely small, the escape server is started in step S500, and the escape operation is performed on the terminal corresponding to the downtime authentication server to All terminals are allowed to access the network.

Finally, if it is judged that more than four (including four) are down, execute step S600, start the escape server, and perform an escape operation on the terminal corresponding to the downtime authentication server, so as to allow all terminals to access the network.

The method disclosed in the present invention includes one or more steps for achieving the purpose of the present invention, and the method steps can be interchanged with each other without departing from the scope of the present invention. In other words, unless a specific order of steps is required for proper operation of the embodiment, the order of specific steps may be modified without departing from the scope of the spirit of the invention. While the present invention has primarily been described with respect to specific embodiments and applications, those skilled in the art will understand that the invention is not limited thereto. According to the methods and systems disclosed in the present invention, various modifications, changes and changes will be apparent to those skilled in the art without departing from the spirit and scope of the present invention.

Claims (4)

1. An authentication server circular escape system, comprising a server group S={S ₁ , S ₂ ... S _N }, said server group S including N interconnected authentication servers S ₁ , S ₂ ... S _N , where N>=6; each authentication server S _i in the server group S has a first backup server S _j and a second backup server S _k in the server group S; A first backup server of an authentication server S _p , and a second backup server of a certain authentication server S _q ; it is characterized in that: The first backup relationship of all authentication servers in the server group S forms a ring, and the second backup relationship forms an inverse ring of the first backup relationship; the ring escape system also includes an escape server for terminals that cannot be authenticated by the server group S escape; Among them, the communication connection path between the authentication servers in the server group S is P={p ₁ , p ₂ ...p _N }, and the communication speed of each communication connection path is B={b ₁ , b ₂ ...b _N }, the physical distance between each authentication server is D={d ₁ , d ₂ ... d _N }; the ring and inverse ring formed by the first and second backup relationships make the value of the weighted weight W the highest; Among them, min(D) and max(D) are the minimum and maximum physical distances between any two servers in the server group S, and min(B) is the minimum communication speed between any two servers in the server group S value; the first type of authentication server is an authentication server that is susceptible to downtime due to interference from natural conditions; the second type of authentication server is an authentication server that is not prone to downtime due to interference from natural conditions. specified, or obtained automatically by the circular escape system. 2. A method for circular escape using the circular escape system according to claim 1, characterized in that it comprises the following steps: Step S100, judging how many authentication servers in the server group are down at the same time, if it is judged that there is no downtime, go to step S200; if it is judged that one is down, go to step S300; if it is judged that there are two downtimes, go to step S400; Step S200, each authentication server authenticates the corresponding terminal; Step S300, migrating the terminal of the downtime authentication server S _e to the backup server with a smaller load among the first and second backup servers of S _e for authentication; Step S400, judging whether the downtime authentication servers S _e and S _f have a backup relationship; if there is a backup relationship, then migrate the terminals of S _e and S _f to backup servers different from S _e and S _f for authentication; If there is no backup relationship, execute step S450; Step S450, migrate the terminal of S _e to the backup server with a smaller load among the first and second backup servers of S _e for authentication, and migrate the terminal of S _f to the first and second backup servers of S _f Authentication is performed on a smaller backup server. 3. The circular escape method according to claim 2, characterized in that said step S100 also includes, if it is judged that there are three downtimes, execute step S500; Step S500, obtain the downtime authentication server set {S _u , S _v , S _w }, if the first and second backup servers S _v , _{S w} of a downtime authentication server Su are down, start The escape server performs an escape operation on the terminal corresponding to _Su , and migrates the terminals corresponding to Sv _{and Sw} to a backup server different from _Su for authentication; otherwise, execute step S520; Step S520, if there is an authentication _server {S _u , S _v } with a backup relationship _in { _{S u ,} S _v , S w }, then migrate the terminals of _Su and S _v to different Perform authentication in the backup server; migrate the terminal of S _w to the backup server with a smaller load among the first and second backup servers of S _w for authentication; otherwise, perform step S540; Step S540, the terminals of {S _u , S _v , S _w } are respectively migrated to the backup servers with a smaller load among the first and second backup servers of {S _u , S _v , S _w } for authentication. 4. The circular escape method according to claim 3, characterized in that said step S100 also includes, if it is judged that more than four machines are down, then execute step S600; Step S600, start the escape server, and perform an escape operation on the terminal corresponding to the downtime authentication server, so as to allow the terminal to access the network.