DE102015123003A1 - Access-secured storage of information on a server - Google Patents

Abstract

The invention relates to the access-secured storage of electronically encoded information on a server, on which an entity authorized to access for the purpose of storing a data encoding the information or the disposal of information held in a memory of the server information is made possible. According to the invention, an information-encoding amount of data is transferred via a front end by specifying its identity and use of at least one authentication factor as an authorized entity and stored in a memory of a re-encoding unit of the server. The transferred data volume is encrypted by means of a key generated by the frontend and transferred to the re-encoding unit. At the same time a timestamp is generated and stored in a memory of the server. The amount of data stored in the memory of the re-encoding unit is then at least recoded continuously by means of a deterministic algorithm processed by a processing unit of the re-encoding unit until the server has at least one authentication factor via the front-end for the purpose of providing the information again successfully presented.

Description

The invention relates to a method for secure storage of electronically encoded information on or in a server, to which access via a network, in particular the Internet is made possible. Such a server may be, for example, a server forming part of a cloud. The aim of the presented method is the reliable and secure storage of secrets, which are associated with the information to be securely stored in the server. The relevant information does not necessarily have to be human-readable information. Rather, it may be information of the most arbitrary kind, which are electronically encoded and digitized for filing and processing by electronic systems.

Of course, in this respect, the information may also be statements which can be directly grasped or interpreted by a human, such as a text or a construction drawing or another graphic representation. Under an access-secured storage of such information is a storage of the information understood in such a way that they are only readable by authorized persons or systems again from a storage of the server to keep them safe to read out or that later information on this information, such as Example, for the purpose of deleting unneeded information, can be accessed.

As already indicated above, an authorized access to a secretive information encoding, stored in the memory module amount of data, so read the relevant amount of data directly by a person, but also by a technical system.

Taking into account the above premises, the solution according to the invention and its subsequent presentation in the description and in the claims assume an understanding according to which a server is a central resource in a network. Different persons as users and / or different computer-based systems preferably have access to this central resource via the network. Within the scope of this, the invention relates to ensuring that the users and / or systems possessing such access can store data-coded information, which may be trustworthy or constitute a secret, using the server as a central resource independently of one another Information is not accessible to all persons and / or systems who have access to the server, but always only to the entity (s) authorized to use it. The term entity here and in connection with the following description of the invention and in the context of the claims is representative of a person or a system which allows at least one access to the server via a network into which the relevant server is integrated is.

Here, the presented method is based on the assumption that the server itself is operated according to "security best practices" principles and, for example, protected against brute-force attacks by the implementation of appropriate protection mechanisms. It is further assumed in this context that persons / users and / or computer-based systems communicate with the server via a secure channel, for example via a channel secured using the Transport Layer Security (TLS) protocol.

The object of the invention is to provide a solution which ensures reliable protection of information stored on a server as a central resource accessible via a network insofar as only authorized persons and / or systems have access to them. For this purpose, a corresponding procedure must be specified.

The object is achieved by a method having the features of patent claim 1. Advantageous embodiments or further developments of the invention are given by the subclaims.

According to the proposed method for solving the problem, a data volume encoding the information to be protected is held in a memory of a re-encoding unit of the server, the designation selected for the aforementioned unit being assigned to that unit in the course of the procedure is derived from the following representations immediately resulting function. An authorized entity is granted access to the re-encoding unit for the purpose of storing the amount of data that encodes the information or information, or using information held in the memory of the re-encoding unit as a data set using at least one authentication factor via a front-end of the server.

According to the understanding previously explained with respect to the concept of the entity, a legitimate entity is a person or a system which is used by the user already mentioned at least one authentication factor is entitled to store a (preferably secretly) information-encoding amount of data in the memory of the re-encoding unit and optionally later access again to these, in order to have them, namely to read, overwrite, delete or manipulate it in any other way.

A data set of information transmitted to the server for storage by such a legitimate entity via the frontend is encrypted in the frontend by means of a key generated by the frontend according to the proposed method. The encryption can take place by means of a symmetric encryption method, so that the generated key is a key serving both encryption and possible later decryption. But also the use of an asymmetric or public key encryption method is possible, wherein a key pair comprising the aforementioned key is generated.

The encrypted amount of data is transferred from the frontend to the re-encoding unit, at the same time a time stamp indicating the time of this transfer is generated and stored in a memory of the server. In the re-encoding unit, the amount of data stored in its memory is then repeated, that is, continuously, as long as recoded by means of a deterministic algorithm processed by a processing unit of the server until the server has the at least one authentication feature for the purpose of having access to the access-secured information is presented again.

After the presentation of the at least one correct authentication factor, the entity presenting this authentication factor can use the front-end to hold the information previously held in the memory of the re-encoding unit so that it can in particular read it out, overwrite it or delete it.

The read-out of the information is made possible by the fact that the re-encoding unit from the hitherto constantly transcoded data volume with the help of the stored in the transfer of the encrypted amount of data to the re-encoding unit in the server back the original to them by the Frontend transferred encrypted data volume. This original encrypted dataset is then passed to the frontend and decrypted by it with the key (key for symmetric encryption or keys of a key pair) generated by it in connection with the storage of the information in the re-encoding unit. Depending on the implementation of the method, provision may be made for the data quantity representing the information to be continuously recoded or deleted after it has been read from the memory of the re-encoding unit of the server.

The timestamp is important in restoring for restoring the amount of data originally passed to the re-encoding unit, because the algorithm used to continually transcode the dataset can easily undo the transcoding. Depending on the algorithm, it may be possible to calculate the original encrypted amount of data that encodes the information, and thus possibly restore the actual sensitive information, more quickly if the time stamp is known.

Two pictorial and mental examples should explain this. For example, if the sensitive information is coded by data representing a body, ie, a three-dimensional object, and the algorithm is to rotate the object in space, then "turning back" to the original position is particularly easy, if known is that the object has already been rotated 357 ° in one direction. Rather than rotating it 357 ° backwards, it can instead be rotated by another 3 ° forward and is then again in non-recoded form, namely in the form of an encrypted amount of data encoding the original information. However, it should be noted that the latter is only an illustrative example and would probably not generate an image in concrete implementations of the method. In that regard, in an implementation comparable to this example, a continuous transcoding of the data quantity encoding the information to be stored with secure access can also take place using multi-dimensional number spaces, but for which the same applies with regard to the meaning of the time stamp.

For example, in a multi-dimensional space, multiple rotations could be made (forward, to the side) and the step sizes chosen differently (for example, one turn 0.27 degrees forward and 0.31 degrees to the right per calculation step), so that the rotated object is not at the same time in the original state with respect to the two or more axes of rotation. The timestamp nevertheless makes it possible to decide which rotations in the various dimensions must still be made in order to bring the object back to its original position. In addition to the twists can also displacements of the object take place in different dimensions.

In a specific embodiment of the invention, for example, the bodies or objects (as representatives of the encrypted data set) can be represented by vectors in n-dimensional space, and the transcoding operations are performed by matrix multiplications with one or more matrices. Such matrix multiplications can be performed quickly in hardware designed especially for this task, for example by the Graphics Processing Unit (GPU) of corresponding graphics cards.

As regards the generation of the time stamp, but in particular its storage, different possibilities come into consideration. Preferably, the timestamp stored in the server when the data encoded encoding the information is transferred to the re-encoding unit is generated by the frontend.

The generated timestamp can be passed to the re-encoding unit according to a first possible method design by the front end together with the encrypted data carrying the information and stored in a memory of the re-encoding unit. If, at a later time, the authentication factor providing the encoded information is then re-presented to the server via the front end, the process of permanently transcoding the amount of data in the re-encoding unit is stopped, and the latter can be based on the memory stored in its memory Timestamp immediately restore the amount of data originally passed to them with the information to be kept secret. It should be noted, however, that the amount of data recovered by the re-encoding unit still contains the information in encrypted form, and the actual information for issue or other form of acknowledgment becomes available only after the frontend has decrypted the restored amount of data.

According to a contrast preferred embodiment of the method, the time stamp generated by the front end is not transferred together with the encrypted amount of data to the re-encoding unit, but stored only in a memory of the front end. In this case, the time stamp is only transferred by the frontend to the re-encoding unit at a re-presentation of the authentication factor authorizing this via the encoded information. The re-encoding unit then calculates the time difference simply because of the difference between the time stamp and the current time (for example, from a real-time clock of the re-encoding unit) and can thus quickly and efficiently restore the originally received encrypted data with the coded information ,

Another possibility is that the time stamp is generated simultaneously in the frontend and in the re-encoding unit and stored both in a memory of the frontend and in the re-encoding unit. However, the frontend and the re-encoding unit must have a timebase synchronized with each other. Upon accessing the information stored in the memory of the re-encoding unit, the front-end passes the time stamp previously stored in its memory to the re-encoding unit for comparison with the time stamp stored in the memory of the re-encoding unit. The re-encoding unit releases the information stored in its memory in connection with a re-presentation of the at least one authentication factor only in the case of a match of the time stamp available.

Optionally, the method may also be designed such that the frontend in connection with the transfer of the encrypted data containing the information to be kept secret, further data are passed, which by the re-encoding unit in the continuous transcoding of the amount of data in the used algorithm. This may, for example, be one or more random numbers. But even the timestamp itself can, if it is passed through the frontend to the re-encoding unit, be included in the algorithm for continuous transcoding. This additional data is stored in the frontend with assignment to the at least one, the subsequent access to the stored data enabling authentication factor. If appropriate, the at least one authentication factor is not directly stored in the frontend within the scope of such an assignment, but only a hash value formed for it (reference authentication factor) which is compared with a hash value (verification authentication factor) generated for the at least one authentication factor presented in each attempt ,

The key used by the frontend to initially encrypt the amount of data that encodes the information (which may also be part of a key pair) and / or the additional ones passed to the re-encoding unit for use in repeated transcoding parameter For example, in the front-end, they can be stored under assignment to an identity tag of an entity authorized for later use via the information.

Another possibility is that the key and / or the additional parameters passed to the re-encoding unit for use in the repeated transcoding are generated by deriving from the at least one authentication factor used in the storage of the information in the server become. In this case, the key and / or the additional parameters for the repeated or continued transcoding of the information recording and for the first time encrypted in the frontend amount of data need not be stored on the server.

As far as the at least mentioned one authentication factor is concerned, this can be, for example, a PIN or even a digital image of a biometric feature, such as a fingerprint or the iris of a user to be detected by means of a camera. Of course, in view of the potential use of biometric features or their digital images in the process, a computer-based terminal used by a particular user to access the front-end of the server must be provided with appropriate means for detecting and extracting such features. Since the method is fundamentally based on the use of at least one authentication factor, the invention can also be implemented in practice or the method implemented such that the server front end for storing information worthy of protection and for later re-accessing this information or on the data set a combination of several of the above or other possible authentication factors must be presented.

As far as the server is concerned, from a hardware-technical point of view this can be a computer-based device, but also a plurality of devices / units arranged distributed over a central access device in the form of the hardware- and software-based frontend.

The method according to the invention can be further developed there or the server for carrying out the method can be set up so that the data, that is to say information-coding data sets, of several users / systems are, so to speak, continuously recoded in the re-encoding unit. In this case, the process of permanent transcoding can be parallelized by alternately transcoding different fragments of data sets of the different users or entities. Different calculation units (for example graphics cards) can be used for different users or for different parts of the encrypted data volume. In a special embodiment of the invention, there is additionally a logic which distributes the arithmetic operations to several re-encoding units.

The algorithm used for the continuous transcoding is designed in such a way that no, at least no short cycles are run through in which an encrypted data set transferred by the front-end to the re-encoding unit returns to its original state, which it currently has had the handover to the re-encoding unit. Thus, a potential attacker can not detect repetitions in the representation of the encoded data set encoding the information. The data available to a potential attacker, who has gained access to the server, at various arbitrary times accessible data or accessible, each presented in a different form of data thus appears amorphous for this potential attacker.

As a result, it becomes virtually impossible for the potential attacker to carry out structured attacks against the data volume held in the memory of the re-encoding unit, that is, in particular to gain access to the information encoded by this data volume. To do this, the potential attacker would first have to undo the repeated transcoding that took place at the time of his access in order finally to gain knowledge of the original encrypted data set and to decrypt it using the key that had come into his possession. For this he lacks any knowledge, in particular, he does not have the timestamp and thus the knowledge at what time the process of permanent or repeated transcoding was started. The key alone, if stored in the frontend, is in any case worthless to a potential attacker, since the attacker alone can not gain access to the sensitive information by means of the key alone. In the case of having the key, it also requires the encrypted sensitive information, that is the amount of data containing the information in the original encrypted form, which has not yet been transcoded to the re-encoding unit.

The following is based on the 1 again a possible sequence of the method according to the invention will be explained. The 1 shows a corresponding flow chart. Accordingly, the sequence of the procedure for the case of secure storage of information and its subsequent read-out by an authorized entity - in the example by a user - designed as follows.

First, the user establishes a secure connection to the front end of the server ("establishment of a secure communication channel") by means of a suitable computer-based terminal. For example, the user contacts the server via a web interface provided by the frontend on the Internet on an https page. For its part, the frontend contacted by the user establishes a secure channel for the action to be performed on the re-encoding unit of the server. The user then sets up a write request to the frontend on the server for the purpose of the information he intends to keep secret on the server. In connection with this, the user identifies with a user identifier, for example by means of a user ID (user ID) used with respect to the server, and moreover authenticates himself as an authorized user by means of at least one authentication factor, such as a PIN or a password.

The frontend checks the data provided by the user for identification and authentication ("authenticate users"). If these are correct, the data transferred to the server by the user as part of the write request is encrypted by means of a key generated by the frontend ("generate key [or key pair], encrypted information"). The encrypted data is transferred together with the user ID to the re-encoding unit ("Transfer for transcoding"). At the same time, a time stamp is generated by the frontend and stored in the server, which corresponds to the time of transfer of the encrypted amount of data to the re-encoding unit. Otherwise, insofar as the user was unable to authenticate successfully, a corresponding message issued by the terminal is returned by the frontend to the user's terminal.

In the first described case of a successful authentication of the user and the transfer of the encrypted amount of data encoding the information to be kept secret to the re-encoding unit, in the re-encoding unit the deterministic algorithm for the continuous transcoding of the received data amount (" continuous transcoding "). The encrypted data set transferred by the frontend to the re-encoding unit is then recoded continuously by the re-encoding unit, for example, using a multi-dimensional number space formed by random numbers, that is, as it were, constantly "scrambled". This is done until an authorized user (entity) successfully presents at least one authentication factor that identifies it as legitimate to the server via the frontend, whereupon the frontend communicates with the re-encoding unit and the process of transcoding is stopped or restored the amount of data originally transferred to the re-encoding unit is interrupted at least for a short time (the "scrambling" may also be continued in the case of an authorized access).

For later re-presentation of the authentication factor then available via the information which is to be securely stored, the user directs (again using a previously established secure connection) a read request to the frontend by means of his computer-based terminal suitable for this purpose. To do this, he hands over his user ID and the at least one authentication factor to the frontend. The latter checks the data and, in the case of its correctness, grants the user access to the stored information or to the encrypted amount of data that encodes this information. For this purpose, the front end in turn transmits the user ID and the time stamp stored for this user ID in the frontend, which indicates the time of the original transfer of the encrypted data volume to the re-encoding unit, and optionally further data (data which for the first time already with the encrypted data volume passed the re-encoding unit for use in continuous transcoding) to the re-encoding unit ("recovery request"). The re-encoding unit then interrupts the ongoing under the corresponding user ID and the time stamp process of permanent transcoding. Based on the timestamp and optional additional data, the re-encoding unit recovers the encrypted amount of data originally transmitted from the frontend and retransmits it back to the frontend. Here, this amount of data is decrypted by means of the key stored under the user ID and transmitted to the output ("read response") to the user's terminal.

Claims (10)

A method for access-secured storage of electronically encoded information on a server, wherein a legitimate entity access to the server for the purpose of storing an information encoded amount of data or disposing of that in a memory of the server as Data held information is given stating their identity and use of at least one authentication factor via a front end of the server, characterized in that - an information-encoding amount of data is stored in a memory of a re-encoding unit of the server by the authorized entity by the Frontend for storage on the server transferred amount of data is encrypted by means of a generated by the frontend key and passed to the re-encoding unit, at the same time a timestamp is generated and stored in a memory of the server, and - that received from the front end, in the memory of the re-encoding unit stored amount of data is at least as long recoded by means of a processed by a processing unit of the re-encoding unit deterministic algorithm until the server re-presents the at least one authentication factor on the front end rt and thereby the access-secured information for an entity presenting the at least one authentication factor is made available using the front-end. A method according to claim 1, characterized in that in the transfer of the information encoding encrypted amount of data to the re-encoding unit stored in the server timestamp is generated by the front end. A method according to claim 2, characterized in that the time stamp is stored in a memory of the front-end and / or the re-encoding unit. A method according to claim 1, characterized in that generated in the transfer of the information encoding encrypted amount of data to the re-encoding unit timestamps generated simultaneously in the front end and in the re-encoding unit and both in a memory of the front end and is stored in a memory of the re-encoding unit, wherein the front-end and the re-encoding unit have a timebase synchronized with each other, and wherein the front-end accesses the information stored in the memory of the re-encoding unit in its memory stored time stamp for comparison with the stored in the memory re-encoding unit timestamp passes to the re-encoding unit, which releases the information stored in its memory only in the case of a match of the timestamp available. A method according to claim 1, characterized in that the server is the at least one authentication factor on the front end presenting entity reading the access secured as a data held in the memory of the re-encoding unit information is enabled by the re-encoding unit the originally encrypted amount of data received from the frontend using the timestamp stored in the server when transferring the encrypted amount of data to the re-encoding unit from the amount of data currently held in the memory of the re-encoding unit and restored to the frontend, where it is decrypted by means of a key to be used and the resulting amount of data, which originally encoded the information originally transmitted to the server, is transmitted to the entity that reads it out. The method of claim 1 or 5, characterized in that the access-secured as data in the memory of the re-encoding unit information held by the server the at least one authentication factor on the frontend presenting entity are released for deletion. Method according to one of claims 1 to 6, characterized in that the Fronend to the re-encoding unit additional parameters are passed, which included by the re-encoding unit in the algorithm for repeated transcoding of the encrypted data received via the front end become. Method according to one of claims 1 or 7, characterized in that the key generated by the frontend, used to encrypt the amount of data encoding the information and / or the additional parameters passed to the re-encoding unit for use in the repeated transcoding the frontend to be stored under assignment to an identity number of an entity authorized to later use the information. Method according to one of claims 1 or 7, characterized in that the key for encrypting the data encoding the information and / or the additional, for use in the repeated transcoding to the Re-Encoding unit passed parameters from the frontend by derivation from the at least one authentication factor used in storing the information in the server is generated. Method according to one of claims 1 to 9, characterized in that in the re-encoding unit simultaneously independent Information that encodes information of different entities is continuously recoded.

Images