Patent Application
20090282252
A client is intended to be authenticated with a server without transmitting his password in plain text. Even a password which is always encrypted with the same key may be intercepted and used by unauthorized persons.
With the conventional method, a random signed challenge text (signed challenge) which is encrypted by the server is requested by the client and is sent. The client decrypts said text, forms a digest and signs and encrypts the latter again (signed digest). The server checks the digests for a match. This method is very complicated.
In a manner similar to a TAN list (One Time Pad), a (finite) list containing keys which should each be used only once could be interchanged in another manner (for example paper) before the first authentication. However, this list would need to be stored or the keys would have to be copied by the user (vulnerable and laborious and prone to error).
The invention is explained in more detail below with the aid of the drawing which comprises one FIGURE.
At the outset (as part of a seed), both sides (client and server) agree on the same starting value and the same algorithm for generating keys.
As a result, the client and server are able to generate the same sequence of keys independently of one another. A sequence which can be generated in this manner is also referred to as a PRBS (Pseudo Random Binary Sequence). A PRBS can be generated, for example, with the aid of an LFSR (Linear Feedback Shift Register).
However, PRBS are strictly deterministic. Therefore, the method can be determined after a small number of results (keys) and an attacker then knows all further keys.
In addition, the sequence of keys is repeated exactly after a finite number of keys (period of the LFSR).
However, if interference is inserted into the sequence at regular intervals (that is to say into the formation of the keys with the aid of the LFSR), the result is neither periodic nor deterministic.
The interference in the sequence is achieved as follows (also see the FIGURE in this respect).
First of all, the client and server form a first key from the starting value (starting key) with the aid of the LFSR, said first key being referred to as the epsilon key ε below for the sake of better differentiation.
The epsilon key is not used for authentication since it is vulnerable.
Next, only the client generates a second key and sends it to the server. Only this second key is used for authentication. In order to understand the formation of the second key, which differs from the formation with the aid of the LFSR, a small digression into metrics is made.
Metrics means the manner in which the distance (delta) between two points is defined. Applied to numbers which are stored by the computer in bits, this means the difference between two numbers in the simplest case. The difference between two binary numbers depends on the significance of the individual bits in the numbers. The significance of the bits is normally determined by their transmission order from 0 (20) to 31 (231) (old metrics):
According to the old metrics, the distance between the numbers 3040593782 and 2503657302 has the value 536936480. If new metrics are now defined by allocating or agreeing on new significances for different bits, said numbers are at a different distance. If, for example, bit 16 receives the new significance of bit 0 (20), bit 29 receives the new significance of bit 1 (21) and bit 5 receives the new significance of bit 2 (22), the two numbers shown above are only at a distance of 7. The order of the other bits is unimportant for this but not the state (0 or 1) of the bits.
If a delta of up to 7 is considered, according to the new metrics, the number 3040593782 is situated in the delta of 2503657302, but the number 2503657303 is further than 7 away from 2503657302 since one bit which has a higher significance than 2 differs.
In the case of the invention, the client and server agree on the new metrics (as part of the seed) by agreeing on those bits (so-called interference bits) which, unlike their transmission order, have a different significance.
Assuming that 3040593782 is the first key (epsilon key) generated by the client and server, a key whose distance (delta) is within a predefined value with respect to the epsilon key according to the new metrics is then formed by the client as the next key (second key). This key is referred to as the delta key δ below. If the client thus generates the number 478651654 as the next key, it is a delta key since this number is situated in the delta of the epsilon key 3040593782.
Since only the client and server know the new metrics which have been amended according to the position of the interference bits, the server can thus authenticate the client using the second key received. If the second key sent by the client is within the delta, the client is authenticated. This key is used as the new seed on both sides and a further shift in the LFSR becomes the new first key (new epsilon key).
Only 32 bits are shown in the example illustrated above. In reality, the numbers are considerably greater, that is to say 2048 or 4096 bits with 8 or 16 interference bits, for example. Other operations can also be used as the difference for determining the distance.
If n is the number of interference bits, all 2n−1 delta keys form the delta, that is to say the area surrounding the epsilon in the case of known metrics, which delta is again known to both sides by virtue of the position of the interference bits and cannot be determined by the attacker.
Advantages of the new solution: a starting key must be interchanged once at the outset as part of the seed with the conventional method. Only one key from the new delta then needs to be sent for each new log-on.
| Number | Date | Country | Kind |
|---|---|---|---|
| 10 2006 039 327.9 | Aug 2006 | DE | national |
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/EP07/58203 | 8/7/2007 | WO | 00 | 5/13/2009 |
