KR100304367B1 - Authentication method for network system - Google Patents

Abstract

The present invention relates to an authentication method of a network system, and in particular, an object of the present invention is to eliminate an authentication failure due to a time deviation by additionally transmitting a separate bit for indicating a period during which the authentication due to the time deviation may fail. . The present invention for this purpose comprises the steps of setting the bit flag when the difference between the client's time and the representative value is less than the maximum time deviation or the difference between the validity period and the representative value and the client time is equal to or less than the maximum time deviation; Setting the value obtained by subtracting the maximum time deviation from the client time to the client time when the bit flag is set, and encrypting the client recognition name, the password, and the representative value calculated from the reset client time using the hash function. And transmitting the cipher value together with the recognition name and the bit flag on the client side, and subtracting twice the maximum time deviation from the server time if the bit flag included in the transmission data from the client is set. Calculating the representative value and calculating the representative value; Reading a password included in the encryption key and encrypting the recognition name, the password, and the representative value calculated using the hash function; determining whether the received encryption value matches the encryption value calculated above; If the result of comparing the encryption value in the configuration is configured to perform the step that the server determines that the authentication success.

Description

Authentication method for network systems {AUTHENTICATION METHOD FOR NETWORK SYSTEM}

TECHNICAL FIELD The present invention relates to authentication techniques, and more particularly to an authentication method in a network system.

In general, the user ID and password-based user authentication mechanism used in the network system uses a fixed password, so it is easy to leak passwords, and there is no way to protect against hackers intercepting passwords on the network.

In fact, most of the hacking incidents occurred at home and abroad through IP address and password theft, and the user's password can be easily found by illegally tapping through the sniffer program.

Therefore, a password leakage prevention technique must be secured in the user authentication technique, and typical user authentication techniques can be classified into a password use method, an encryption method, and a non-encryption method.

1) User authentication using passwords is the mechanism adopted by most real authentication systems, but is particularly vulnerable to external exposure, speculation, eavesdropping and replay.

2) Cryptographic methods include common key methods such as Kerberos and public key methods such as X.509, but in actual system implementation, they are not used because they require a lot of effort and cost in key management.

3) Non-encryption methods include one-time passwords, challenge-response, personal characteristics, and address-based authentication.

In conclusion, most user authentication systems that are actually implemented use one-time password authentication or challenge-response authentication.

These methods have been researched at home and abroad, and dozens of certified products have been produced and applied to various applications.

On the other hand, although challenge-response authentication can be included in the category of one-time password authentication in a broad sense, challenge-response authentication is a two-pass authentication protocol that is different from the one-pass authentication protocol of one-time password authentication and implemented in a real system. Considerations vary greatly in time, so the two methods are generally classified differently.

The operation process of the challenge-response authentication scheme is the same as the signal flow diagram of FIG. 1, which will be described below.

To use a network system, a client sends its password to a server, which applies the password to a specific function, calculates the "challenge", and sends the "challenge" to the client. 'And the password above are calculated together to calculate the expected response value R.

At this time, the client calculates a response by calculating a password and 'challenge' from the server and transmits the response to the server.

Accordingly, the server compares the response value (Response) in the client with the response value (R) calculated by itself, and if the comparison result is determined to be a legitimate user to allow the client to use the network system.

In other words, the challenge-response authentication method requires a function to receive the challenge and calculate the response on the client side, and to generate and transmit the challenge and calculate the response on the server side. It is not easy to apply to native number and password based authentication protocol.

However, in the non-encryption method, the one-pass password authentication method, which is a one-pass authentication protocol, performs authentication by directly transmitting the authentication value generated by the client to the server, and thus it can be applied without any change to the existing authentication protocol based on the user ID and password. Do.

That is, the one-time password authentication method can be easily implemented by adding only the authentication value generation function to the client program and the verification value calculation function to the server program without changing the existing authentication protocol.

Therefore, when the non-cryptographic methods are applied to the existing user ID and password based authentication system, the one time password authentication method can be implemented much faster and easier.

One-time password authentication involves 1) a random password list (requires location synchronization within the password list), 2) a pseudo-random sequence number (requires synchronization of the state of the sequence generator), and 3) time (of time). Synchronization required).

However, in the case of a large number of client / server environments, a method using a random password list and a sequence number causes many problems in maintenance, but a method using time can be used without any maintenance problem.

Therefore, the operation of the one-time password authentication method will be described taking the case of using time as an example.

The authentication method using time is performed in the same process as the signal flow of FIG. 2, and the authentication process is separately described to the client and server sides as follows.

(Client)

1. Client's time ( Read).

2. The client's time ( ) Is converted into seconds and the corresponding representative value ( Calculate

3. Client Identification Number ( ), password( ) And the representative value calculated above ( ) To the hash function ( ) Using the encryption value ( Calculate

4. The encryption value calculated above ) In the Client Identification Number ( Add) to the server.

(server)

1. Client Identification Number ( ) Is an encrypted value ( ), The server's time ( Read).

2. The server's time ( ) Is converted into seconds and the corresponding representative value ( Calculate

3. The client's unique number ( Password of the client corresponding to Read the representative value above With a hash function ( ) To calculate the encrypted value ( Calculate

4. Two encryption values ( ), ( ).

Accordingly, when the algorithm and the same operation are executed in the client and the server, if the comparison result is matched, the server informs the client that the authentication is successful and allows the use of the network system.

In general, in the case of the authentication method using time, all the time is converted into seconds.

Here, the client and server side takes a representative value within the unit period of time and uses it as an input value during authentication. In general, the representative value within the unit period selects the start point of the interval.

For example, when the time is divided by a predetermined unit as shown in FIG. 3, if the time at which the authentication process starts is 'aa: bb: cc', 'aa: bb: 00' is taken as the representative value.

In particular, when all unit periods are constant, this unit period is called an effective period and the same representative value is maintained during this valid period.

If the valid period is set to '1' minutes when the new representative value starts as a change point, the change point is a time point that becomes '0' seconds.

However, the authentication method using time should consider time synchronization, password reuse within valid time, and time deviation.

In other words, time synchronization must be guaranteed to match the authentication data.

Time synchronization can be out of sync if the client and server are geographically remote, but this can be solved by installing the Greenwich Mean Time program on both the client and server.

Then, appropriate measures for reuse of the same password are required within the validity period, that is, the period in which the current one-time password is continuously used.

This is because the one-time password generated by the client within the validity period is always the same, so if a hacker taps a password on the network and attempts to reconnect within the validity period, it can cause problems in the authentication system if the hacker succeeds in authenticating the user.

However, this problem can be solved by maintaining the most recent authentication success time for each user.

In other words, if the connection is attempted again within the valid period after authentication, the server may search for the most recent authentication success time and reject the authentication when the valid period has not expired.

In addition, there is always a problem of more serious time deviation in the time-based authentication scheme.

That is, the time deviation may occur between the client and the server due to the time deviation due to the calibration period, the authentication data calculation time of the client, and the transmission time.

Therefore, when a time deviation occurs, a time period during which authentication is not performed properly may occur because time synchronization is not guaranteed.

At this time, the probability of authentication failure may be calculated by the following calculation (1).

---- Formula (1)

Therefore, there is a problem in that the probability of failure increases more and more as the time deviation increases with time.

Hash function used by common algorithms ( ) Uses a fast computational algorithm, so if the transmission time is less than the time deviation, the problem is caused by using a time-dependent clock and extending the validity period to reduce the probability of authentication failure in the authentication algorithm. Can be solved.

However, the conventional technique is expensive to use a clock having a small time deviation, and further, extending the validity period may threaten safety due to repeated use of the representative value within the validity period.

Accordingly, the following three methods have been proposed as solutions for the authentication failure due to time deviation, but each has problems.

The first method is to re-authenticate when authentication fails due to time deviation without taking any further action. This method has a problem of causing performance degradation of the authentication system.

Secondly, when authentication fails, the server performs authentication by comparing several values before and after the current authentication value. In this method, the user's password is regarded as multiple passwords in the server and used in comparison. have.

Third, when the authentication fails, the server re-transmits the time value to the client and performs authentication again. This method has three-pass authentication, and there is still a problem in that the probability of authentication failure due to time deviation still exists.

Accordingly, the present invention eliminates authentication failure due to time deviation by additionally transmitting a separate bit for indicating a period in which authentication due to time deviation is likely to fail to solve the problem of time deviation in the prior art. It is a first object to provide an authentication method of a network system which is invented to be.

In addition, the present invention has a second object to correct the time deviation in consideration of the transmission time.

1 is a signal flow diagram illustrating a challenge-response authentication process.

2 is a signal flow diagram illustrating a password authentication process using a general time.

3 is an exemplary view illustrating a representative value setting range in FIG. 2.

4 is an exemplary view showing a time deviation in an embodiment of the present invention.

5 is a table showing a comparison between the present invention and the prior art.

In order to achieve the above object, the present invention reads the time of the client, converts it into seconds, calculates a representative value, clears a bit flag indicating a period in which authentication failure may occur due to time deviation, and Setting a bit flag for indicating a period of time in which there is a possibility of authentication failure if the difference between the time and the representative value is less than the maximum time deviation or the sum of the validity period and the representative value and the client time difference is less than or equal to the maximum time deviation; Setting the value obtained by subtracting the maximum time deviation from the client time to the client time when the bit flag is set, and encrypting the client recognition name, the password, and the representative value calculated from the reset client time using the hash function. Transmitting the cipher value together with the distinguished name and the bit flag. Performed on the client side; Receiving cipher data transmitted from the client and subtracting twice the maximum time deviation from the server time to set the server time when the bit flag is set, calculating the representative value from the server time set above; Reading the password of the recognized recognition name and encrypting the recognition name, the password and the representative value calculated using the hash function, determining whether the encryption value received and the encryption value calculated above match; If the comparison result in the encryption value is matched, the server judges the successful authentication by performing the step of correcting the time deviation, characterized in that it is determined to be a legitimate user.

The process of performing the authentication algorithm of the client may include setting the current client time by adding the minimum delay time after reading the client time.

In the above, the effective time is set so as to be equal to or greater than four times the sum of the variation of the time deviation and the delay time.

Hereinafter, the present invention will be described in detail with reference to the drawings.

The algorithm of the present invention also performs the same process as the signal flow diagram of FIG.

In general, in the time-based authentication algorithm, the clock embedded in each of the client and server has an error of a crystal oscillator, and the error value may be changed by a clock calibration period.

That is, each client and server clock has errors due to the clock's crystal oscillator and calibration cycle.

Due to this error, when performing authentication between any client / server using the one-time password authentication system using the internal clock, another authentication value is generated, which is one factor of authentication failure.

The client uses a token, commonly called a one time password generator.

if, ' Is the maximum error of the clocks that are embedded in the client and server, respectively, 'Is determined by the maximum error of the crystal oscillator of the watch and the maximum value of the computer's clock calibration period.

As shown in Figure 4, the absolute time ( For the client and server clocks, The maximum time deviation (T) between client and server is given by the following equation, and the absolute time between client and server is the same.

For example, the frequency stability of the crystal oscillator built into the watch is 'e' ppm ( ) And the maximum error (when the watch's maximum calibration cycle is one week) ) And the maximum time deviation T are obtained as follows.

That is, if the stability of the crystal oscillator is within 10 ppm, the maximum time deviation T is 13 seconds or less.

Therefore, the following theoretical background is introduced in the conventional authentication system using time to implement a complete authentication system using time.

Assembly Z, for any natural number V, T that is Characteristic function , Step function Define as

here, Is the largest integer not greater than y.

That is, the present invention is an algorithm for eliminating the probability of authentication failure due to time deviation occurring in the existing authentication system using time by additionally transmitting a separate 1-bit indicating the period in which authentication by time deviation is likely to fail. In this case, the method is classified into a method of correcting only the time deviation and a method of correcting the time deviation in consideration of the transmission time.

First, the time deviation correction algorithm in the present invention will be described.

T: Maximum time deviation between client and server clock

V: Validity Period

: 1-bit flag indicating the period during which time-based authentication may fail

(Client)

1. Client's time ( Read).

2. Client's time ( ) Is converted into seconds and the corresponding representative value ( Calculate

3. Bit flag ( ) Is cleared to '0'. )

4. Client time ( ) And client time ( Representative value calculated by ) 'S car Is less than the maximum time deviation (T) between the client and server clocks, or the effective time (V) and the difference value ( If the difference is less than or equal to the maximum time deviation T, then the bit flag ( ) Is set to '1'.

5. Bit flag ( ) Is set to '1', the client time ( ) Subtracts the maximum time deviation (T) from the Reset to).

{ }

6. Client Identification Number ( ), password( ) And the representative value calculated above ( ) To the hash function ( ) Using the encryption value ( Calculate

7. The encryption value calculated above ) In the Client Identification Number ( ) And bit flags ( ) Together and send it to the server.

(server)

1. On the server, the client's unique number ( ) And bit flags ( ) Is an encrypted value ( ), The server's time ( ).

2. Bit flag ( ) Is set to '1', the server time ( Subtract twice the maximum time deviation (T) from the new server time ( Set to).

( )

3. The newly set server time above ) Is converted into seconds and then the representative value ( Calculate

4. Unique number ( Password of the client corresponding to ) Above the representative value ( ) And a hash function ( ) To calculate the encrypted value ( Calculate

5. Two encryption values ( ), ( ).

Accordingly, when the algorithm and the same operation are executed in the client and the server, if the comparison result is matched, the server informs the client that the authentication is successful and allows the use of the network system.

However, since the time deviation correction algorithm does not consider the transmission time, authentication does not always succeed, and the probability of authentication success is affected by the validity period and the time deviation of the clock.

That is, the above time deviation correction algorithm is based on the theorem set below. , You can always authenticate with a normal user.

First, the natural numbers x, y, V, and T , And If you satisfy , Which proves as follows.

, Because of There is a natural number n that satisfies.

At this time, By the definition of equation (5) to be.

therefore, Because of Being established This holds true.

In other words, Because of It can be seen that this holds true.

The natural numbers x, y, V, and T , And If you satisfy Is established and proved as follows.

Because of Suitable to meet and Is present.

At this time, ego Because of , , Are established respectively.

therefore, , It can be seen that is established.

However, if the client calculation, transmission, delay, and token input time are simply referred to as delay time, the delay time greatly affects the stability of the algorithm.

At this time, if the delay time is constant, an effective authentication algorithm can be derived, but if the delay time is not constant, there is a problem in the design of the authentication algorithm.

Therefore, in the present invention, the variable (for the delay time) , ), We design the authentication algorithm.

: Minimum Client Calculation, Transmission, and Delay Token Entry

: Client Calculation, Transfer, and Delay Token Entry Maximum Times

That is, the time deviation correction algorithm considering the transmission time in the present invention is as follows.

(Client)

1. Client's time ( Read the minimum delay time ( ) Add the new client time ( ).

2. The client time set above ) Is converted into seconds and the corresponding representative value ( Calculate

3. Bit flag ( ) Is cleared to '0'. )

4. Client time ( ) And client time ( Representative value calculated by Car Is less than the maximum time difference (T) of the client and server clocks, or the effective time (V) and the difference value ( If the difference is less than or equal to the maximum time deviation T, then the bit flag ( ) Is set to '1'.

5. Bit flag ( ) Is set to '1', the client time ( ) Subtracts the maximum time deviation (T) from the Reset to).

( )

6. Client Identification Number ( ), password( ) And the representative value calculated above ( ) To the hash function ( ) Using the encryption value ( Calculate

7. The encryption value calculated above ), The client's unique number ( ) And bit flags ( ) Together and send it to the server.

(server)

1. On the server, the client's unique number ( ) And bit flags ( ) Is an encrypted value ( ), The server's time ( ).

2. Bit flag ( ) Is set to '1', the server time ( Subtract twice the maximum time deviation (T) from the new server time ( Set to).

( )

3. The newly set server time above ) Is converted into seconds and then the representative value ( Calculate

4. Unique number ( Password of the client corresponding to ) Above the representative value ( ) And a hash function ( ) To calculate the encrypted value ( Calculate

5. Two encryption values ( ), ( ).

Therefore, when the same operation as the above algorithm is executed in the client and the server, when the comparison result is matched, the server informs the corresponding client that the authentication is successful and permits the use of the network system.

That is, the algorithm can reduce the difference between the client time input value and the server authentication time by adding the minimum delay time to the client time in advance.

However, even in the above algorithm, the variation of delay time ( ) Cannot be corrected, but assuming that the sum of the time deviation and the delay variation of the clocks in the client and server is 'T'. In the above algorithm, the one-time authentication algorithm using time is very effective.

On the other hand, comparing the performance of the algorithm of the present invention and the conventional algorithm is shown in the table of FIG.

That is, as shown in the table of FIG. 5, the existing authentication algorithms are very unstable when there is a difference in token calculation time and transmission time, that is, delay time, but the algorithm of the present invention has a high probability of authentication failure due to time deviation of a clock. In addition, the failure probability according to the delay time can be seen that there is no authentication failure within the expected maximum delay time.

As described in detail above, the present invention adds only 1-bit to user authentication information without changing the protocol of the conventional one-time password authentication system so that authentication failure due to time deviation does not occur, thereby improving reliability and processing time. There is an effect that can achieve a high speed.

Claims (3)

A first step of reading the time of the client to calculate the representative value, a second step of clearing a bit flag indicating the period of time that the authentication may be failed due to time deviation, and a difference between the client time and the representative value is less than the maximum time deviation. Or setting a bit flag for indicating a period of time in which authentication failure may occur when the sum of the validity period and the representative value and the difference between the client time is less than or equal to the maximum time deviation, and the client if the bit flag is set above. Resetting the representative value by resetting the value obtained by subtracting the maximum time deviation from the time to the client time; and encrypting the client recognition name, the password, and the representative value calculated in the fourth step by using a hash function. The client side performs the fifth step of transmitting the cipher value together with the recognition name and the bit flag. .; A sixth step of receiving cipher data transmitted from the client and subtracting twice the maximum time deviation from the server time to reset the server time if the bit flag is set; and calculating the representative value from the server time reset as described above. A seventh step of reading a password of the received recognition name and encrypting the recognition name, the password, and the representative value calculated in the seventh step using a hash function; and the received encryption value and the encryption calculated above. The server may perform the ninth step of determining whether the values match and the tenth step of determining that the authentication is successful if the encryption value comparison result is matched to correct the time deviation to determine whether the user is a legitimate user. The authentication method of a network system. The method of claim 1, wherein the first step includes reading the client time and adding the minimum delay time to the current client time. 2. The authentication method of a network system according to claim 1, wherein the validity period of the third step is set to be equal to or greater than four times the sum of the variation of the time deviation and the delay time.