WO2008061389A1

WO2008061389A1 - Document management device and method

Info

Publication number: WO2008061389A1
Application number: PCT/CH2007/000587
Authority: WO
Inventors: Patrick Richter; Jürg BIRCHER
Original assignee: Uptime Products Ag
Priority date: 2006-11-24
Filing date: 2007-11-23
Publication date: 2008-05-29

Abstract

The invention relates to a document management method and device, particularly for filing documents in an audit-proof manner. The device comprises a server (5) for generating a time stamp and/or a time stamp signature for each document (D1', Cn) that is to be saved as well as a journalizing unit (11) for recording the time stamps and/or time stamp signatures in a uniform journal file (Jn, 7). In the disclosed method, a fingerprint which is journalized in the journal file (Jn, 7) along with a time stamp signature ({DM1}ks

Description

Device and method for document management

The present invention relates to a document management apparatus and method according to the preambles of independent claims 1 and 11.

Today, there is a regular need for structured collection and management of printed and / or electronic documents, such as correspondence, receipts or other written documents or electronic mail, transaction data, picture documents, etc. The invention relates specifically to an audit-proof document management system (DMS), ie the capture, storage, management, making available and retrieving of such documents, so that in particular the integrity and storage time of the stored documents can be determined in a binding manner. Documents are understood here to be systemically closed units of data which form a contextual unit. As far as paper documents are concerned, these can be letters, forms, administrative files, receipts or even entire books. In addition, multimedia files (such as a photograph) can represent documents in a broader sense. By contrast, dynamic data groups are not to be understood as "documents", as they are typically managed in database systems which map continuously changing contents which do not form a uniform context on the screen or ephemeral data sets, in particular if they are purely technical functions and not intended for a single person In general, a document is characterized by the following typological features: formal unity (structure, design), document order (order, logical affiliation, Version etc.), communicative content, document character (legal properties, editing options, archival eligibility, etc.), time (creation, modification date, last access, archiving time, etc.), personal data (creator, distributor, reader, person responsible, etc.). Often, the document implicitly or explicitly contains this information or it is associated with the document as attributes or linked to it.

A digital document can be present as a directly readable string, which is coded according to standardized character sets (so-called Coded Information Documents (Cl-D)) or as an image file or as a non-coded data set (so-called NCI-D, ie Non Coded Information Documents). Thus, document management systems differ fundamentally from the so-called database management system, which in turn can provide supporting functions for a document management system, but which concern a different problem level.

The present invention serves equally the management of digitized paper-based documents as well as the management of electronic documents, which under the term of electronic document management (EDM) represent a subset in the field of document management systems. Frequently, there is also talk about document management in the narrower sense, which is primarily intended to cover the administrative functions of large document databases. In the core area, document management systems in the narrower sense are described in ISO standard 10166 for Document Filing & Retrieval (which, however, is not significant in practice). The information managed by DMS in the narrower sense is document-oriented. Their management, especially access and presentation, are determined by document features. For example, to enable search queries, keywords are assigned to the documents by the DMS ("keyword") or the documents are completely indexed.Typically for DMS in the narrower sense, the formation of so-called containers, ie document groups, a version management and so-called. DMS in the narrower sense of the term does not include workflow aspects, document processing and other process integration, including archiving functions by DMS in a broader sense, also referred to as Information Management Systems (IMS) or Enterprise Content Management Systems (ECM) addressed. The latter also deal with digitizing (especially scanning) and archiving documents.

Various document management systems are known from the prior art, which serve the paper handling and the electronic management of documents and data.

No. 5,742,807 discloses a system and a method in which electronically stored documents are identified by means of document attributes by means of a one-way hash function and are managed via a hash storage location intermediate index for the document management system. This intermediate index is used for allocating and finding document attributes in memory. Although this document deals with document content and its modification because of allocation problems, it offers no solution for ensuring the integrity and authenticity of stored documents in the sense of a revision-proof system.

Document US 6,810,404 shows a method to process documents of various types (paper based and electronic documents) in a computerized system. The system is used to record, manage and archive such documents, with their document attributes being analyzed and checked for preconditions with a view to archiving the documents, in order to check the archiving capability of the document. Also in this system, an archive index is formed and the document may be moved from a first to a second memory area. This system deals with the archivability of a document, but also does not provide a solution for a revision-proof DMS. From JP 2001 2431 19A2 it is known to check the data integrity when transferring between two memory areas for document groups. However, that solution does not allow for document integrity or authenticity for each document in an archive store.

From US 6,188,766 an authentication system is known which uses fax services to authenticate individual documents and provide them with a time stamp. Archiving takes place at the fax recipient or at the publisher of the time stamp. This solution is technically and practically limited in many ways. On the one hand, the authentication and archiving takes place at an external service provider, which leads to high transaction costs, on the other hand, the system is limited to fax systems, which is impractical for high transaction capacity or large amount of documents to be authenticated per time.

From EP 0 892 521 a method and a device is known, which serves the long-term authentication of digitally signed documents. The local solution is based on the fact that the necessary PKI data is archived in order to be able to understand the verification process with regard to the signed document at a later date. For this purpose, a long-term signature verification server is used, which can replicate long past signature verifications. This solution is not designed for fast processing while ensuring the authenticity and integrity of individual documents at acceptable cost.

Overall, it can be seen that these various known solutions are each very specific to a specific problem solution. All solutions have the disadvantage that they lead to high transaction costs or considerable infrastructure costs, in particular when processing a very large number of documents. For this reason, it is often necessary today, special proprietary hardware solutions in the form of so-called. Using HSM modules (High Security Modules), which store documents and signatures securely, causing undesirable infrastructure costs and causing known disadvantages and restrictions of proprietary solutions. In addition, the systems described do not address a reliable archiving solution for a DMS that reliably ensures the timely authenticity and integrity of each individual document.

It is therefore an object of the invention to provide a method and a device for document management, which reliably and cost-effectively enables the proof of the temporal authenticity and the integrity of documents and in particular allows a high performance and a fast archiving process when large processing capacities are incurred.

This object is achieved by the characterizing part of the independent patent claims 1 and 11.

The invention uses for this purpose a method for document management for audit-proof archiving with the following method steps:

1. Create a first set of data from a digital document containing a hash of the digital document and an internal / local (not a certified timestamp (TSA) timestamp);

2. Encrypting this first set of data with a private key of a local server to a timestamp signature (this step (subject matter of claim

2) can be omitted in special embodiments);

3. journal this timestamp signature as a journal entry in a journal file; 4. repeating steps 1 to 3 above for at least one more or a plurality of documents and forming a journal file with journal entries of the time stamp signatures of the corresponding documents;

5. Completing the journal file according to predetermined completion criteria and transmitting the hash value of the journal to a certified time stamp company (TSA);

6. Receiving a (certified) timestamp signature from the timestamp company containing at least the hash value of the journal, a certified timestamp from the timestamp company, and encrypted with a timestamp (TSA) key using an asymmetric encryption method;

7. Save the journal and its timestamp signature in a repository and repeating these steps appropriately.

In particularly preferred embodiments of the method, in method step 1, the data volume containing the hash value of the digital document and the internal time stamp, with a signature of a server (internal timestamp signature), information about the hash function or the relevant algorithm and Document information (file ID, document or document component information and version number) provided. In method step 3, the corresponding journal entry is provided with an identifier about the respective journal and / or journal segment (journal ID and journal segment ID). In addition, for the purpose of search options, the document meta data is preferably also stored.

In addition to the time stamp signature, a journal entry ID can be stored with each journal entry, assigned to the document and stored in a memory area In addition, this Jourπaleintrags-ID is preferably returned to the document (mapping). In preferred methods, additional attributes and / or properties for the document V are stored with each journal entry, such as a document ID, which allows a fast and efficient assignment of journal entry and document. This document ID is also preferably returned to the document or its document container.

In order to increase the collision resistance of the system, the first data set is preferably assigned a unique number or a unique document ID in addition to the hash value of the digital document and the internal time stamp. It is also possible to add a serial number and / or special metadata and / or additional time information to the first data set.

A preferred embodiment of the method provides that not only the hash value of the respective journal is transmitted to the certified time stamp company, but alternatively or cumulatively the complete journal.

The inventive device for document management for audit-proof archiving has a computer / server for creating timestamps and / or timestamp signatures for each document to be stored, a journalization unit for recording the timestamps and / or timestamp signatures to a uniform journal file and a communication connection a certified time stamp company (TSA). In this case, this communication connection can be made by the server itself or by a separate computing / communication unit. To improve performance, it is preferable that the journaling unit be one with the server.

A variant of the invention provides that at intervals in addition to the time stamp signatures for the documents or instead of these interval time stamp signatures. stored for the entire journal file in the journal file. For this purpose, the server generates a timestamp signature for the entire journal file at intervals after several documents to be stored or even after each document, up to the respective time / content status of the journal file.

Furthermore, the invention offers the particular advantage that the journal entries can be journalized within a journal within a journalization unit, and only then stored in a final journal, the timestamp signature being obtained only for the final journal. This allows a particularly high performance for searches within large or a large number of journal files. For this purpose, a searchable index is preferably additionally stored in or together with the journal file.

Reference to the figures, embodiments of the invention will be explained in more detail. Show it

1 shows a conventional solution for the temporal signing of documents

2a shows a document management system with a time stamp signature method according to the invention

2b shows a second embodiment of a document management system with a time stamp signature method according to the invention

2c shows a third exemplary embodiment of a document management system with a time stamp signature method according to the invention

3 shows the procedural sequence of the time stamp signature method according to the invention 4 shows an overview of the verification process with regard to inventive audit-proof archived documents

Fig. 5 is a schematic diagram for explaining the verification process with respect to the embodiment of Figure 2c

More particularly, the invention relates to the portion of a DMS that relates to the storage and archiving of documents. Functionality, terminology and system structure of such DMS are assumed to be known in the following. The invention is, for example, in connection with the document management system ARTS® of UPTIME products AC, CH-8004 Zurich, (http://www.uptime.ch), which covers the areas of document capture, processing, archiving and document retrieval includes, usable. Under Archiving, the permanent storage of documents, i.d.R. scanned documents, in particular paper-based documents, such as letters, incoming invoices, delivery notes, signatures, policies, etc., or outgoing documents, such as customer letters, receipts, etc., which are stored in a known manner (still commonly referred to as "COLD (Computer Output on Laserdisk) method).

A challenge with such archiving processes is the guarantee of the so-called revision security as well as the assurance that the archived documents are legible and retrievable even after a long time, ie after ten years or more. In particular, the problem of audit security, which as a rule presupposes the fulfillment of defined criteria such as data integrity, temporal assignability, definiteness of the archiving procedure according to national legislation, presents problems if the archiving costs per document should be affordable. In order to be able to assign a document to a defined time (reception or creation time), a temporal verification is required. Due to the legal requirements for digitally archived documents, time stamps and digital signature procedures, which are issued by a certified independent authority (usually called electronic signature procedures), are required for a large number of documents to be archived Of more importance, however, is the technical problem that the collection of time stamp signature from a digital certificates issuing institution (so-called TSA ("Time Stamp Authonty") and PKI) per transaction data retrieval via computer mumcation and / or computer networks, which consumes temporal resources and the performance of the storage or archiving process massively affected in a short time a large number of archiving processes (eg hundreds of thousands of transactions per hour e in the case of document archiving or banking transactions, etc.), substantial bandwidths are required and the processing capacity of the TSA and the local network massively claimed. Assuming that a large number of customers request certificates of a particular TSA, a reliable archiving process for the individual customer can scarcely be guaranteed, since bottlenecks, collisions and Pnoritatspro- cause problems, which in turn had to be handled with expensive and complex technical precautions Here now the idea of invention begins by reducing the transaction volume with the TSA while maintaining auditing security and a very high-performance solution creates

FIG. 1 shows a conventional archival process, obtaining a digital signature from a certified publisher of timestamp signatures ("Timestamp Company"), here designated TSA (Time Stamp Authonty). Such a digital signature contains a timestamp for verifying the documents Further data, which is explained in more detail below Em Paper-bound document 1 is digitized in a first step to a document V and stored in a Repositoπum 10 This digital document V is a so-called Fingerpπnt 2 assigned by in known By means of a hash function (eg SHA-I) a hash value H is calculated for the document. In order to increase the collision resistance of the overall system, ie to avoid two identical fingerprints because of identical hash value, the fingerprint is also supplemented by a random number, a nonce N. The corresponding unique value of the fingerprint 2 represents the content of the document V (and thus of the document 1). In a further step, this fingerprint 2 is supplemented by an internal time stamp T. The first data group 3 (T + H + N) is now transmitted to the TSA. The TSA adds to this data group 3 an accurate time, ie a verified time stamp VT of the TSA, and, if necessary, additional authentication information AI. These data are finally encrypted with a TSA key, usually using an asymmetric encryption method. This signed data set formed from the time stamp T, the hash value H and the nonce N and the verified time stamp VT and the possible authentication information AI, encrypted by the private key ks ^" ^{] of} the TSA, thus form a time stamp signature 4 for the document T or the original document 1. The document V and the time stamp signature 4 are archived individually or jointly in the DMS in a data container V, 4. In this way, the time and integrity of the document V can be checked at a later time by means of the time stamp signature 4 It will be appreciated by those skilled in the art that, in the context of this per se known process, especially the inclusion of the TSA for each individual document brings with it the aforementioned disadvantages and problems.

In Figure 2a, a first embodiment of a DMS according to the invention will now be explained. A document 1 is, if not already present as a digital document, scanned or digitized to a digital document V, in the present example, for example, as a PDF file. In the manner described above, a fingerprint 2 is formed for the document V from a hash value H and a nonce N. The fingerprint is assigned an internal time stamp Ts. Internal timestamp Ts is understood to mean a time stamp which is generated by the company using / operating the DMS. Preferably this time stamp from a central, in-house or independent UTC synchronized server 5 awarded. However, this server 5 does not have the quality of a TSA because its operation is not done by a publicly-certified institute. Furthermore, the invention makes it possible to associate with each document T or its fingerprint, in addition to the internal time stamp T _s, a sequential number (serial number / ID), special metadata or additional time information (cf., for example, the embodiment described below according to FIG.

In a next step, the server 5 creates a preferably encrypted by means of a private key ks ^{1 of} the server 5 certificate representing a signed internal time stamp IT _n - corresponding to an internal timestamp signature - which is typically a maximum of 100-200 bytes. This signed, but not TSA-certified, time stamp IT _n does not meet the auditability criteria by itself. By means of a public key k _s (not shown in the figure) of the server 5, the hash value, and the internal time stamp Ts can be extracted from a repository 10 or by individual clients at a later time.

For special systems, the signing can also be omitted by means of the private key ks ^"1 , for example if an immediate certification of the document is not required (eg very short chronological sequence of journal files) or not (eg for very large transaction volumes) signed internal timestamp IT _n only the data H, N and Ts, and is used in this form for the subsequent journaling.

In order to ensure the requirement of revision reliability, a further step according to the invention is provided. The respective signed internal time stamps IT _n are chronologically journalized, preferably in a uniform journal file 7. Each journal entry comprises per document V the data of the internal time stamp IT _n . In addition, each journal entry can contain additional information, such as a journal entry ID JEID _n , an identifier for the respective document class (dependent on the respective DMS), as well as other attributes and properties for the document T (depending on the DMS). The journal entry ID JEID _n and / or a journal ID JID _n can be added to the container file of the document V so as to make it possible at a later date to easily assign the document to its internal timestamp signature IT _n . However, the assignment can also take place via an index via the decrypted hash values H and Nonce N, as will be explained below. Preferably, each joool file in turn has a journal ID JID _n for uniquely identifying each journal. In particular, this journal ID JID _n makes it possible to identify the relevant journal J _n at a later point in time if a specific document needs to be verified and whose journal entry data is required for this purpose. If required, the journal ID JID _n can also be used to record the chronology of the journals or other classification criteria and information.

After expiration of a certain time interval or after detecting a predefined number of internal timestamp signatures IT _n , the respective journal 7 is completed. Such a time interval may be, for example, an hour, a day or a month. Immediately after completing the journal 7, it will be saved as a single file, unless it has already been entered as a journal file during the journal entry. This is necessary, for example, if the journal entries are recorded individually on a database basis. This journal file 7 is in turn provided with a time stamp signature 9 via a TSA (or a TSA server). Preferably, a fingerprint method is here also used in which a hash function and optionally a nonce or the Journal ID JID is uniquely identified and this amount of data 8 from hash value (and nonce or journal-ID) of TSA in order timestamped _n the journal file and signing is transmitted. In particular embodiments, the journal file itself may be assigned additional time stamps by the server 5, which represent, for example, the beginning and the end of the journal, or further metadata. The respective journal files 7 as well as the The associated time stamp signatures 9 of each journal file 7 and the journal ID JID _n are stored here in a common repository 10. It can be seen from FIG. 2 a that here also the digital documents T are stored in the same repository 10. Of course, it is not necessary that the logical or physical memory areas for these data are identical, and in the context of the invention, the documents T and the journal files or the data necessary for the management of these (esp. Journal IDs JID _n ) in different Be stored in repositories. The assignment between documents T and the respective journal is ensured via hash value + nonce or via respective indexes, journal entries or journal entry IDs.

The foregoing procedures will enable one skilled in the art to appreciate that each individual document T (and hence a paper bound document 1 identified thereby) is verifiable as to its integrity via a signature associated therewith, and at the same time by means of the authenticatable journal file 7 provided by a TSA with a Timestamp signature 9 is provided, even its authenticity is established (the term authenticity is here not to be understood in the sense of a personal verification). Overall, the system is thus able to ensure audit security with high performance, while at the same time the transaction volume with the TSA is massively reduced in accordance with the task and the data traffic between the local server and the TSA is correspondingly low. During the archiving of a specific journal file 7, an index is created which makes it possible to quickly find and verify the documents via the fingerprints, serial numbers of the documents or time ranges with regard to the document retrieval.

In order to check the temporal authenticity and integrity of a single document V, the corresponding journal file 7 is retrieved via the fingerprint of the corresponding document and the associated internal timestamp signature IT _{n of} the document V is determined. Depending on the DMS, the journal file concerned for the timestamp IT _{n in} question is sent via a suitable query, eg a time identification or metadata of the document, such as journal ID, document type, etc., called and the timestamp signature IT _n read out. In the case of variant embodiments in which a journal entry ID JEID _{n is} allocated for each journal entry and returned to the document and recorded in the respective document container, the respective timestamp signature IT _n can be sent directly via the journal entry ID JEID _n (if necessary under Incl the journal ID reference). In other cases, the assignment of the document V to the associated time stamp signature IT _n preferably takes place via its fingerprint. From the time stamp signature IT _n , the associated hash value and the time stamp can be read out by means of the public key of the TSA and of the server 5 and thus verify verified time and the integrity of the document. The integrity check is performed by comparing the hash value read with the hash value determined via the document to be checked by means of the corresponding hash function.

From the point of view of audit security, it can be seen that if a correctly signed journal file 7 is available, its temporal authenticity and integrity are fixed, and thus the specifically affected timestamp signature IT _{n is also} considered authenticated. The result of this method is that the affected document T is verified to be temporally authentic and with integrity.

The method according to the invention can be well integrated into DMS in which several representations of the same document exist. It is selectable in these cases, if only the main document, several or all representations of the same document V are provided with a time stamp signature.

FIG. 2b shows a further variant of the invention which differs in several ways from the above-described exemplary embodiment. In the present case, a digital document T should be archived in a revision-proof manner. The digital document T is associated with metadata meta-data (eg keywords, author, creation date, etc.), whereby document V and metadata together in the present case are to be understood as document containers 1 ". a unique document ID DocID is also assigned, similar to the embodiment according to Figure 2a, a hash value is now formed via the document container 1 "and, together with the document ID DocID, from an internal server 5 with an internal time stamp T _s so that a timestamp signature IT _n is created. In addition, the document ID is stored in a database 12. The document V or the document container 1 "is stored in a repository 10, which may possibly coincide with the database 1 2. A nonce is no longer provided in this method or a corresponding device If required, a hash value can alternatively be formed only via the actual document V (without metadata and possibly further data).

At least one, but usually a plurality of such time stamp signatures IT _n are now stored in a journal 7 analogously to the above-described embodiment. 2b may be of the figure, however, be clearly seen that each journal entry IT _n instead of having a journal entry identifier with the corresponding document ID DoclD _n is indexed, which allows each entry IT _n directly to a document container 1 assign ". The locking and signing each Journals 7 within the journalizer 1 1 by means of a timestamp signature can be done in an analogous manner as in the above-described embodiment or according to the options described below: The journals 7, journal ID JID _n , the time stamp signatures 9 and other administrative data (such as Public keys, etc.) are stored in the repository 10. Preferably, the journal IDs JID _{n are} also stored in the database 12 and managed and indexed together with the respective document IDs, so that at a later time, using journal ID and document ID, very fast to a specific document 1 'or to a documentcont 1 "the associated journal entry IT _{n be} found and the corresponding verification of the document are made (see. in detail the description below). Preferably, the inventive method and the corresponding device in this variant is designed so that for each document V, the journal ID JID _{n is} stored not only in the database 12 or in the repository 10, but each document V is returned (mapping). This not only simplifies finding the respective verification data (time stamp signature IT _n ) of the document, but also has the particular advantage that the document container can be directly retrieved whether the corresponding document has ever been archived in a revision-proof manner, since only documents archived in a revision-proof manner according to the invention have a journal ID JID _n .

FIG. 2c shows a further embodiment variant of the invention. This variant is distinguished by the possibility of being able to export or verify the document container 1 "or parts thereof detached from the entire journal in a revision-proof manner, as in the above-described examples per document container 1" in the sense of the following statements Journal entry generated. In the example according to Figure 2c contains the document container 1 "not only a singular document V, but several representations of the document V, here exemplified as each a file in PDF and DOC format of the document shown The document containers 1 "may contain any number of document representations Ci, C ₂ or additional information, such as meta-data C ₃ , for example. In general terms, the document container 1 "can contain components C ₁ to C _n - or in other words: a document to be journaled in an audit-proof manner can only be formed by a component of the document container 1". In this case, various representations of a document each represent one of the components Ci to C _{n of} the document container 1 ". The meta-data C ₃ or other document information (such as special file attributes), associated auxiliary documents, etc. can be understood as a special case of a document component C _n become. In view of the purpose in this embodiment of being able to verify one or more components of a document container 1 "quickly and in isolation (detached from the entire journal), a journal entry JE _{n is} created for each document or document component C _n contains the following data:

1) Journal ID JID _n and journal segment ID SID _m

2) Internal timestamp signature IT _n

3) Algorithm Information Al _n

4) Component Information Cl _n

5) Component data CD _n (not required for revision security)

The internal time stamp signature IT _n is created similar to the above example, as shown in FIG 2b. Typically, a hash value h [C _x] or h [C _π] is a document component C _x calculated and this _x in the respective associated time stamp signature IT (not shown in Figure 2c) committed. However, it is also possible to use a hash value H (C _x )... H (Cy) for each of several document components C _x ... C _y or, in special cases, a common hash value H (C _x ... Cy) (the latter not shown and also not described further below) and to store them in the internal timestamp signature IT _n . The respective time stamp T ₅ and the associated certificate (RSA signature) are in turn supplied by a server 5. For security reasons, the private key k _s ^{"] of} the server 5 is neither permanently stored internally in the system nor is it available outside the respective process; rather, it is held only in the process memory for the duration of the process.

As algorithm information Al _n the required data on the system used in the digest algorithm are stored so that at a later date directly from the Journal entry JE _{n is} traceable with which function the aforementioned hash values were calculated. This makes it possible to replace the hash function at a later point in time in a strain gage according to the invention (eg from SHA-I to SHA-2). It should be noted here the completeness that, if desired, the algorithm information Al _{n may} also be stored only at the level of journal segments or even only the full journal and only with the export of the respective journal entry JE _n can be exported (see. see below).

The component information Cl _n serves to identify the respective document component C _x . Preferably, the file ID (FiIeID) of the document container 1 ", component name and component version are used here for unique identification of a file component, but it is also possible to use a standalone document ID as shown in the above examples.

In addition to the component information Cl _n-component data CD _s are stored in the Journal Entry JE _n. This information is used here preferably for later searches directly in a stored journal or journal segment. The component data CD _n here include, for example, the meta data C ₃ . Scope and type of component data CD _{n to} be stored can be defined system-dependent. For the actual verification process of respective document components, the component data CD _n are not necessarily required or not relevant.

Finally, the journal ID JID _n and the journal segment ID SID _{n have} to be addressed. This information is fixed at the latest during the actual storage process of the journal entry JE _n in the journalization device 1 1. In the case of more complex systems, according to the invention, several archiving or working processes Pi to P _n , parallel, sequential or overlapping, can take place within the framework of the DMS. For every work process Pi to P _m , here is a nes Journal segment JS JS ₁ to _m provided, which are stored in total in a pan-adhering Journal 7 within the Journalisierungseinrichtung 1. 1

As already explained above, the respective journals 7 are completed and stored at periodic intervals. Each journal segment JS _m is identified by a journal segment ID SID _n . By its affiliation with a particular journal 7, which in turn has a journal ID JID _n , each journal segment is identified. Since each journal entry JE _{n is} stored in a journal segment, the journal entry JE _{n can} each be assigned a journal ID JID _n and a journal segment ID SID _n during the storage process. In this way each journal entry JE contains _n-mentioned five information packs JID _n, _n SID, IT _n Al _n Cl _n and CD _n. The timestamp signature IT _n per journal entry JE _n contains, as stated above, both the hash value H and an internal time stamp Ts and thus corresponds to the information IT _n = {HJsJks ^{' 1} . Alternatively, the journal entry JE _{n contains} an RSA signature plus a time stamp Ts. Journal ID JID _n and journal segment ID SID _n permit the unambiguous assignment of the respective journal entry JE _n to a journal segment JS _n .

Several of these journal entries JE _n are stored in this manner within a segment Journal JS _m. In addition, a public key kp _{m of} each work process is stored in each journal segment JS _m . In order to be able to prevent the unauthorized manipulation or erroneous alteration of an entire joumal segment JS _n within a journal 7, the storage of the respective public key k _{Pm is made} encrypted, symmetrical encryption being preferred.

The signing of the journal segments JS _n and the journal 7 can be done in different ways now: either each journal segment JS _n is signed at its conclusion by a certified time stamp 9 a certified timestamp company TSA. In this case, analogously to the preceding embodiments, a data set 8 to this TSA transmitted. This data set 8 contains at least one hash value h [k _Pm ] of the respective public key kp _m . Alternatively, only at the conclusion of the entire journal 7, a certified time stamp 9 can be obtained and either the journal 7 can be signed as a whole or the relevant time stamp 9 can be stored multiple times, ie per journal segment JS _n . In most cases, obtaining a single certified time stamp 9 and storing it once per entire journal is preferable.

An embodiment which is preferred according to the invention in the context of the embodiment just described is that the journal 7 shown in FIG. 2c is understood as a temporary journal. In this case, it is possible to sort all the journal entries JEi ₁ to JE _m , _n before the obtained time stamp 9 for the entire journal 7 'or to optimize them for a binary or indexed search. It can be seen that for a temporary journal 7, instead of the above-described symmetric encryption of the respective public keys kpi to kp _m, these can also be signed (temporarily for the duration of the process).

Since each journal entry JEi, i to JE _{m, n} has journal ID JID _n and journal segment ID SID _m , the public keys kpi to kp _m can be stored separately in the final journal 7 ¹ , since each journal entry is again stored via its journal segment Index can be assigned. The final journal 7 'thus contains the following data areas: All journal entries JEi _1I to JE _{m, n} (sorted or search optimized), certified timestamp 9 (see below) and all public keys kpi to kp _m . These public keys kp _m are stored in the final journal 7 'unencrypted, ie the above-mentioned symmetric encryption of the respective public key kp _m is optionally resolved. In addition, the DMS own document classes are preferably stored in the final journal 7 ', which - together with the data contained in the journal entries (especially meta-data) - wide search options directly in the respective journals 7' allows, which accordingly a self-contained verification and search structure (self- without the actual document or component content itself). The skilled artisan recognizes that the final journal 7 ', unlike a temporary journal 7, the public key kpi to kp _m - as briefly mentioned above - need not be additionally encrypted or signed, since the temporal authenticity of the journal 7' secured by a certified timestamp, as explained below.

The final journal 7 'thus specified is now provided with a verified time stamp 9'. This contains a hash value over all public keys k _P1 to k _Pm a certified time stamp and thus serves the temporal verification and integrity check of the journal. The idea of the invention does not rule out that, instead of a TSA, an external HSM device (high security module) carries out the authentication, since this is required only per journal and not per document component, and thus largely avoids the aforementioned disadvantages. Finally, it should be noted that the journal (s) 7 are again stored in a repository 10, which in the present embodiment also forms the storage area for the document containers or document components.

FIG. 3 now shows the method sequence of the archiving process according to the invention in a flow chart for the exemplary embodiment according to FIG. 2a.

In a first method step a), as far as a paper-bound document Di is to be stored, a digitization step is undertaken and a digital file Di 'is created. This is done for example by a scan. In a second step b), a first data set DMi is created which contains a hash value h [Di '], a nonce N and an internal time stamp T _s , which is obtained from a server S (not from the TSA).

In a third step, the first data set DMi is signed or encrypted with a private key k / ^{1 of} the server 5 to form a timestamp signature ITi = (DMi which is usable for a subsequent journalization. This process step can be omitted in special systems (cf also see above or the dependent claim 2). Embodiments in which, after the first method step, the first data set DMi (per document) is signed with a private key of a server 5 to a time stamp signature {DMiJks ¹ for the subsequent journalization, however, always have the advantage of a particularly high security against inadmissible mutations ,

This time stamp signature {DMiJks ^"1 is journaled in a next step d) as a journal entry Jei and complemented in preferred variants of the invention to other data such as a journal entry ID JEIDi or document attributes _attributes! (Such as a document ID DocIDi). The step d) is repeated for a plurality of documents D _n , so that in the course of the method step e) a journal file JA is created which comprises a plurality of journal entries in the form JA = JEAI, JEA ₂ , JEA3, .... J EA _Π - The next procedural step f) is that a journal is completed according to predetermined termination criteria (eg time interval or number of journal entries JE _An ) and its hash value h [J _A ] (possibly supplemented by a nonce or journal ID JID _n ) is transmitted to a TSA, where the hash value of the journal J _{A is provided} with a certified time stamp T _A and signed by means of a private key krsA ^{1 of} the TSA to a time t-stamp signature {h [J _A ], TA] I ⁽ TSA ¹ - This timestamp signature {Π [YES], TAJICTSA ^"1 is returned to the respective enterprise (usually server 5). The journal J _A and the timestamp signature {h [J _A ] are now stored in a suitable repository 10 at the respective company. In a method step g), the method steps a) to f) are repeated analogously so that a journal file JB is created, which is correspondingly temporally signed via the TSA and also together with its time stamp signature {h [Jß],

in a method step h) is stored in a repository 10. These process steps g) and h) can be repeated as needed. It may be desirable according to the invention that the respective journal files 7 not only after their completion with the external time stamp signature {h [J _A ],

but additionally by internal timestamp signatures, which are created at intervals by the server 5 or by a separate server and can verify the integrity of the journal itself. These interval timestamp signatures can be created periodically (eg after m journal entries IT _m or every 30 minutes) and integrated into the journal file, so that these eg a structure ITi, IT ₂ , ..., IT _m , ITZi, IT _{m +} i, ..., ITZ ₂ , ... IT _n ^ ₁ , IT _n , where ITZi stands for a first and ITZ ₂ for a second interval timestamp signature. In particular, the use of a separate server for the creation of such interval timestamp signatures is advantageous in order to additionally increase the manipulation security of the overall system and may also be advantageous in particular if not every single journal entry IT _{n is} signed. In these cases, it is ensured that the interval timestamp signatures ensure (internal) temporal authenticity at regular intervals before the journal is completed altogether.

A corresponding method step (see claim 9) would preferably be integrated at intervals according to step d) described above. The interval time stamp signatures ITZ _n are preferably stored in the journal file itself, but can also be stored independently and assigned to the respective journal J _n via the journal ID JID _n .

In special embodiments with the above-described interval time stamp signature method can also, especially if contextually related document groups are to be stored, instead of a timestamp signature per document D / created an interval timestamp signature on the entire journal after storage of each document become. In other words, in this case, only the non-timestamped hash value of each document (combined with any data additionally to be stored in the journal per document) is attached to the respective journal. thereafter provided in total with an interval time stamp signature. Such a journal file accordingly has the typological structure H [D ₁ '], ITZi, Ji [D ₂ ¹ ], ITZ ₂ , h [D ₃ '], ITZ ₃ , h [D ₄ '], ITZ ₄ ,. ..., the interval time stamp signatures ITZ _n here each representing a timestamp signature over all temporally preceding journal entries, so that one could also speak of a cumulative time stamp signature method.

It will now be explained with reference to FIG. 4 how, at a later point in time, a particular document Di can be verified from the point of view of revision. In a first method step a), the digital document Di 'is designated as representative of the document Di (insofar as it is a paper-bound original). The corresponding journal file JA is determined via its hash value + nonce, via its document ID, which in this case must be contained in the document container, or according to the example in Figure 2a via a journal entry ID JEIDi (see above). Subsequently, in a step b), the corresponding journal entry JEAi is extracted from the associated journal file J _A via the journal entry ID JEIDi or the document ID. In the next step c) the verification of the journal entry JEAi and the journal file JA itself takes place. As described above, the journal entry JEAi contains both the hash value h [Di] and the internal time stamp Ts. Both values can be extracted from the journal entry JEAi by using the public key ks. The temporal authenticity of the time stamp Ts can be checked by comparing the read hash value Pi [D ₁ ] with the hash value that can be calculated for the file Di 'to be checked. In addition, the integrity and temporal authenticity of the journal JA itself must be checked. For this purpose, the two information ΓI [YES] and TA are determined by means of the public key kjsA and again the corresponding extracted hash value h [J _A ] is compared with the hash value of the journal file J _A to be checked. If the values match, the temporal authenticity of the time stamp Ts is verified and the document Di 'is verified to be temporally authentic and integer. With reference to FIG. 5, a later verification process for a document component according to the exemplary embodiment according to FIG. 2c is described by way of example with reference to an example. In this exemplary embodiment, according to the invention, it is achieved that the temporal authenticity and integrity of a document component C _n (or also of component data CD _n ) can take place outside the original DMS. For this purpose, the desired document component C _n and the signature for this component C _n , which is assigned to it via journal ID JID _n and journal segment ID SID _n , together with the public keys kpi _m and the certified time stamp 9 or 9 ' (or in the alternative embodiments with the HSM signature) exported. The document components can be verified by means of the time stamp 9 or 9 'and the information Al _n via the hash function used and via the public key k _P _m . The journal segment ID SID _n defines the authoritative public key kp _π , which permits the verification of the document via its internal timestamp signature IT _n or its hash value analogously to the above explanations regarding FIG. Unlike what is described there, however, the integrity and temporal authenticity of the entire journal need not be checked, but these are determined directly for the corresponding document or document component C _n .

The device according to the invention, which allows the method to be carried out, comprises the usual components of a DMS, namely at least one client and a DMS server 5, which performs the DMS function (digital storage of documents, retrieval of documents , Business logic, archiving functions), and a repository 10 for storing the data and the journals or timestamp signatures and other data according to the invention. According to the invention, the DMS server 5 has a timestamp function and preferably allows the creation of timestamp signatures that are created with a suitable encryption method (eg RSA). Preferably, the server has a UTC synchronization. The device also has a journaling device 11, which has a time stamp for every document archiving or receives a timestamp signature from the DMS server 5 and stores them in groups in a timestamp journal. This journalization device can be formed by a separate device, such as a powerful computer, or integrated directly in the DMS server 5.

In particular embodiments, it may be desirable for the journals to be completely archived in a TSA. In these cases, after completion of the journal file (YES) according to the termination criteria, not only the hash value h [J _n ] of the respective journal J _{n is} transmitted to the certified timestamp company, but alternatively or cumulatively the complete journal YES.

In order to be able to ensure audit-proof long-term archiving of the documents, it is provided according to the invention that the public keys used (public key) of the server 5 (and any further internal server) and the external certified time stamp company TSA are stored in the document management system. Either these public keys are stored and managed separately with each journal file or, since they can be uniquely assigned via the journal IDs.

A further inventive step in connection with a long-term archiving is that a time stamp signature is obtained per journal or preferably groups of journals from a second (or further) certified time stamp company TSA _{2 that is} independent of the certified time stamp company TSA. This measure has significance if the (very unlikely) case should occur that the key pair of the first timestamp company compromises, ie "cracked", so that no audit-proof information would be guaranteed by the key pair k _s / k _s ° The time stamp signature of the second certified time stamp company would serve to further verify the integrity of the journals or journal groups. Appropriate procedure makes it clear that this can achieve a particularly high level of security for a revision-proof document management system with comparatively little effort and costs. The acquisition of a timestamp signature in conventional procedures per document from two certified companies would be neither practical nor cost reasons practical.

Claims

claims

1. Method for document management for audit-proof archiving, characterized by the following method steps:

a) generating a first data set (DMi, (h [C _n ], Ts)) from a digital document (Di ', C _n ) which has a hash value of the digital document (h [Di'], h [C _n ] ) and an internal timestamp (Ts); b) journaling the amount of data (DMi, (h [C _n ], Ts)) as journal entry (JEi); c) repeating steps a) to b) for at least one further document (D _n ) and forming a journal file (JA) with at least two journal entries (JEA ₁ ,

d) Completing the journal file (J _A ) according to predetermined termination criteria and transmitting the hash value of the journal (II [J _A ]) or the hash value of a public key (h [kp _m ]) with which the journal entries (JE _n ) are signed to a certified time stamp company (TSA); e) Receiving a time stamp signature ({h [J _A ], TA ^ _S A ¹ , 9, 9 ') from the time stamp company (TSA), calculated via the hash value of the journal (h [J _A ]) or via the hash value of the public key (h [kp _m ]) and via a certified time stamp (TA) from the time stamp company (TSA) and encrypted with a key (kγs _A ) of the time stamp company (TSA); f) storing the journal (J _A ) and its time stamp signature ({h [J _A ],

, ⁹ - ⁹⁾ ^m (a repository 10).

2. The method according to claim I _1, characterized in that after the first method step, the first data set (DMi, (h [C _n ], Ts)) with a private key of a server (5) to a time stamp signature (DMl Jks ^{" 1} ) which is used for subsequent journaling.

3. The method according to claim 1 and 2, characterized in that in the method step a) algorithm information (Al _n ) and document information (Cl _n ) and at Veifahrensschritt b) each journal entry (JEi, JE _n ) with an identifier (JID _n , SID _n ) are provided.

4. The method according to any one of claims 1 to 2, characterized in that in addition to the time stamp signature ({DMiJks ^ or the first data set (DMi) with each journal entry (JEi) a journal entry ID (JEID _n ) is stored and this

Journal entry ID (JEID _n ) is assigned to the document (Di ').

5. The method according to any one of claims 1 to 4, characterized in that with each journal entry (JEi) further attributes and properties for the document V are stored.

6. The method according to any one of claims 1 to 5, characterized in that the first data set (DMi) in addition to the hash value of the digital document (h [Di ']) and the internal time stamp (T ₅ ) a unique number (N) or a contains unambiguous docu- ment l D (Docl D);

7. The method according to any one of claims 1 to 6, characterized in that the first amount of data (DMi) additionally a serial number and / or special

Contains metadata and / or additional time information.

8. The method according to any one of claims 1 to 7, characterized in that in addition to the hash value (h [J _π ]) of the respective journal (J _n ) alternatively or cumulatively the complete journal (J _A ) is transmitted to the certified time stamp company ,

Method according to one of claims 1 to 8, characterized in that in the journal file (J _A ) interval time stamp signatures (ITZ _n ) at intervals in addition to the time stamp signatures ({DMiJks ^"1 ) or to the first data sets (DMi) or instead of those for the entire journal file (YES) at intervals.

10. The method according to any one of claims 1 to 9, characterized in that the journal entries (JEη _, i .. _{m π} ) within a temporary journal (7) within a journalization unit (1 1) are journalized, and then the journal entries (JEi, i ... m _r π) are stored in a final journal (7 ') and the timestamp signature ({II [YES],

{h [kp _m ], TA) kjs _A ¹ ) is only obtained for the final journal (7 ').

1 device for document management for audit-proof archiving according to one of the claims 1 to 9, characterized by a server (5) for creating time stamps and / or time stamp signatures for each document to be stored (Di '), a journalization unit ( 1 1) for acquiring the timestamp and / or timestamp signatures into a uniform journal file (J _n ) and a communication link to a certified timestamp company (TSA).

1 2. Device according to claim 1 1, characterized in that the journalization unit (1 1) with the server (5) forms a unit.

Device according to one of the claims 1 1 or 1 2, characterized in that the server (5) at intervals generates timestamp signatures for the entire journal file (J _n ) and at least one document (D /) to be stored.