Method for Securely Storing Electronic Records

A portion of the disclosure of this patent document contains material that is subject to copyright protection. The copyright owner has no objection to the facsimile reproduction by anyone of this patent disclosure, as it appears in the Patent and Trademark Office file or records, but otherwise reserves all copyright rights whatsoever.

The present invention relates generally to the storage of electronic files and, more particularly, to a more secure method of capturing, storing, retrieving, and presenting a stream of data. During the course of conducting business, managers are often required to maintain accurate records. To run a business smoothly and comply with regulatory and other legal requirements, business personnel must often be able to establish that a particular record was created or existed at a particular time. Also, they must be able to (1) confirm that the record has not been changed or (2) establish how the record has been changed over time.

For many years in the past, records were maintained, in the normal and ordinary course of business, on paper. Occasionally, a witness would sign and date a particular document to help establish that it existed as of the date of signing. The integrity of such files was generally assumed adequate, since it is often difficult to change paper documents without the paper showing signs of alteration.

Over the last several decades, however, computers and electronic files have become increasing ubiquitous. Business records are increasingly being maintained on electronic media, such as, for example, computer memories, floppy disks, magnetic tapes, and optical compact disks. Often, the dates and other data on such electronic media can be readily altered, without detection, even by operators who are not well-versed in digital technology. Many electronic documents can be copied and modified endlessly without obvious signs of the tampering, since it is often a trivial matter to change the date stamp on a computer file.

Thus, without precautions, stored digital files may often be easily tampered with. Various approaches have been taken to attempt to solve the problems associated with establishing the creation and existence of an electronic record at a particular point in time and confirming that the record has not been altered in the meantime. However, such procedures may prove cumbersome, expensive to implement, or unreliable.

One approach to maintaining electronic record integrity involves creating a "hash" of an electronic record: a check number representing the result performing computations on one or more digital representations of information fields in the document. To the extent someone alters a first digital document, the hash of the altered second document generally is different than the hash of the first document. Thus, after a document has been created, the hash of the document may, for example, be sent to witnesses or published in a newspaper. In this way, for example, the owner of the record may show that, since the hash of a particular record is the same as the hash published five years ago, the record must have been created and maintained in an unaltered state for five years.

Nonetheless, such a method still leaves open the possibility that two individuals may collude to falsely state the value of a hash. Also, with appropriate cryptographic techniques, undetected alterations may still be made. For example, one may alter a document as desired and then make other suppressed changes, such as a carriage return followed by a space-up command, such that the original and altered documents have the same hash value. See, for example, B. Schneier, Applied Cryptography, Chapter 3.8, Timestamping Services, pages 61-65 (John Wiley & Sons, Inc. 1994).

In a principal aspect, the present invention relates to an improved method of storing or retrieving an image data stream or electronic record. An image identification code and time data provided by a trusted source are combined to generate a key. The image identification code and time data are store in a public directory associated with the image data stream. Further, attributes of the image stream are determined and used to generated a verification code. The verification code is placed in a private area associated with the data image stream, and the private area is encrypted with the previously generated key.

In a related aspect of the invention, the key is generated using a user-supplied password as well as the time data and image identification code. The password is not stored in association with the data image stream, and, unless the password is again used, the key to decrypt the private area can not be generated. In another embodiment, at least a segment of the image data stream is encrypted in addition to the private area. In still another embodiment, the stored time data may be used to generate an audit trail of when the image data stream was modified.

Thus, an object of the present invention is an improved method for securely storing electronic records. A further object is a secure storage method for electronic records that is more reliable and easier to implement. Another object is a record storage method that is less prone to allow documents to be modified without detection. Still a further object is a storage method that provides an audit trail of changes that were made to the document since its creation. These other objects, features, and advantages of the present invention are apparent from the following detailed description.

FIG. 1 is an isometric view of a computer system employing a present invention which receives time data from a trusted source;

FIG. 2 is a representation of a document with a data image stream, image file header, image file directory, and private area that may be stored by the computer shown in FIG. 1;

FIGS. 3 a-e are block diagram of a methodology employed to store the document shown in FIG. 2; and

FIG. 4 a-c a block diagram of a methodology employed to retrieve the document shown in FIG. 2.

Background

Referring to FIGS. 1-4, a preferred embodiment of the present invention is shown as an improved method for securely capturing, storing, retrieving, and presenting a stream of data. As shown in FIG. 1, the method may be used with an electronic computer 10. The computer may be used to store an image data stream on a medium 14. The medium must be understood to encompass a wide variety of digital media, such as, for example, computer memories, floppy disks, magnetic tapes, and optical compact disks. An exemplary disk is shown in FIG. 1. In one embodiment, the computer 10 may be interconnect to an antenna 16, a communication network 18, and/or "smart" hardware, such as a dongle (not shown).

The computer 10 may include a processing unit 22, monitor 24, keyboard 26, and magnetic strip reader 28. See FIG. 1. Both the strip reader 28 and keyboard 26 allow a user (not shown) to supply data to the processing unit 22. The communication network 18 may include a cable 30 interconnected to a local or wide area network server computer 32. The network 18 may also include a telecommunication facility 34 that includes modems 36 and 38 and another computer 40. The computer 10 may also be interconnected to peripheral equipment, such as a paper printer 42.

For convenience, the server computer 32 and the remote computer 40 are both hereafter collectively referred to simply as the remote computer. In the most preferred embodiment, the remote computer is under the control of a third party, unassociated with the user who records an image data stream on the medium 14.

Protocol

The preferred embodiment utilizes an established tag-based file format for documents. The particular arrangement chosen is the Tagged Image Format File, or "TIFF," system described in the publicly available specification entitled: "TIFF Revision 6.0 (Jun. 3, 1992)." At present, the specification is distributed by Aldus Corporation, 411 First Avenue South, Seattle, Wash., 98104-2871.

TIFF is a useful and widely used format for storing, describing, and interchanging raster images. The TIFF specification states that TIFF describes image data that typically come from scanners, frame grabbers, and paint- and photo-retouching programs.

Overview

When the computer 10 is to store an image data stream on the medium 14, the computer 10 receives, from at least one trusted source, time data and an image identification code for the image data stream to be recorded. An image data stream may be, for example, a particular document, record, file, computer monitor image, or other collection of electronic information.

The present method depends upon receiving time data from a source that is trusted to be accurate. The computer 10 accepts the time data from a "trusted source" during the process of recording the image data stream on the medium 14. Of course, the more secure and trusted the source of the time data, the more assured the user may be that any recorded image data stream will be accurately recorded and retrieved, without alteration.

The data from a trusted source may be, for example, simply calendar and clock data, in terms of year, month, day, hour, and minute, received from the internal clock of the processing unit 22 of the computer 10. In a more preferred embodiment, such time data is provided by the clock of the remote computer.

Other, more precise time data may also be provided to the computer 10 via the communication network 18 or antenna 16. Such praise time data express time to the nearest tenth or one hundredth of a second. Such precise time data (measuring time in fractions of a second) may be noted and recorded by the computer 10. Such signals may be generated by, for example, global positioning system satellites.

In yet another embodiment, the global positioning system signals also provide to the antenna 16 a location signal that allows the computer 10 to determine the antenna's longitude and latitude position on the earth, as well as the height of the antenna 16 above sea level. At present, such positioning systems are accurate to within, for example, thirty meters.

The method described uses an image identification code, sometimes referred to the International Standard for Image Identification ("ISII"): a substantially unique number assigned to a particular image data stream to help protect its integrity. The unique number may be generated, for example, by a "hard lock" dongle or other intelligent hardware that generates a substantially unique number for each image recorded on the medium 14 by the computer 10.

In another embodiment the image identification code is provided by a trusted source, such as, for example, the processing unit 22 of the computer 10. In a more preferred embodiment, the image identification code is provided, upon the request of the computer 10, by the remote computer. In still another preferred embodiment, the antenna 16 receives the image identification code from a dedicated radio signal generator (not shown).

In such an environment, the computer 10 may issue a request to the remote computer, advising that it wishes to record an image data stream. After identifying itself and conducting the proper handshaking protocols to authenticate its request, the remote computer may then provide to the computer 10, via the communication network 18, a substantially unique code. The image data stream is then stored in conjunction with the image identification code.

When the remote computer issues the image identification code, it may also then record how the computer 10 identified itself, the time that the computer 10 made the request (as measured by the same or a different source than that used by the computer 10), and the unique image identification code that was provided to the computer 10. Of course, should any dispute over the integrity of an image data stream arise, the fact that the information stored by the remote computer is consistent with the information recorded by the computer 10 in association with the image data stream will help in establishing the integrity of the electronic record.

Upon receiving the necessary time data and image identification code, the computer 10 stores the image data stream on the medium 14, together with the image identification code and time data. A verification code that represents the image data stream is also stored in association with the image identification code. The verification code is generally stored in an encrypted, private area, but may also be broadcast back to the remote computer, so that the remote computer may record it in conjunction with the image identification code that the remote computer issued.

After being stored, a user may wish to review the stored image data stream on the monitor 24 or print it out with the printer 42. If the user wishes to alter the image data stream, the computer 10 may allow the changes to be made and record the changed document as a new image data stream. The computer 10 will also note the time of the change (and how the changes were made) and record such information in association with the new data image stream. In this way, the computer 10 may provide an audit trail of when (and how) the image data stream was altered.

A schematic representation of information that is stored by the computer 10 according to the present method is shown in FIG. 2. The process followed by the computer 10 in recording the data image stream on the medium 14 is shown in FIG. 3.

The computer 10 accepts the image identification code from the trusted source, as well as accepting the time data. Any precision time data is also noted.

The user may elect to provide a password to the computer 10 to store and later retrieve the image data stream. In such a case, the user may supply a password via the keyboard 26 or by swiping a magnetic strip, such as a magnetic-strip credit card device (not shown), through the reader 28.

In the preferred embodiment, the user may elect not to supply a password. In such a case, the computer 10 itself supplies a "standard" password that is maintained by the computer 10. The standard password remains the same for all the image data streams recorded on media by the computer 10.

Still referring to FIGS. 2 and 3, the computer 10 may also note the format or "source" of the image data stream provided. Thus, the computer 10 may record a source indicator that the source of a particular image data stream is, for example a facsimile machine or optical reader, if the format of the image data stream meets the requirements for facsimile machine records or optical reader records.

Upon accepting the image data stream, time data, image identification number, precise time data, location data, password, and source indicator, the computer 10 generates a record 50, such as that shown in FIG. 2. The record 50 includes a public image file header 52 and public image file directory 52, as well as an image file 56 and a private area 58. The image file header 52 is consistent with TIFF standards and includes a pointer 60 to the image file directory 54.

The image file header 52 is initially read by a computer when an image data stream recorded on a medium is to be reviewed. The image file header 52 allows the computer 10 to determine the format used to record the image data stream and where to look for other information necessary to interpret the digital data recorded on the medium.

The computer 10 stores the time data and image identification code received from the trusted source in the image file directory 54. If available, location data and the source indicator for the data image stream are also recorded in the directory 54. A pointer 66 to the private area 58 is inserted in the image file directory 54.

The recorded time data and image identification code are illustratively labeled 62 and 64 in FIG. 2. It must be understood that such information inserted into the public image file header 52 or information file directory 54 may, of course, be inserted into the private area 58, rather than, or in addition to, being inserted into a public area.

The password is used, together with the time data and image identification code, to generate a key. Since a user-supplied password is generally not recorded on the record or otherwise generally known by the public, if the password is provided by the user, the key may be considered "private." Otherwise, the key may be considered "public."

The private area 58 cannot be read by a reader or other computer without constructing the key. Thus, the use of the user-provided key allows for a higher level of assurance of the integrity of the stored image. As a consequence of using such a password, however, should the password be forgotten, the image data stream may not easily be constructed (since the password is one element in the seed value for generating the key).

The attributes of the image data stream are then determined. Such attributes can include, for example, the size of the image data stream. Also, the computer 10 can compute a hash for a particular segment 68 of the image data stream and/or for the entire image data stream. The size of the image data stream, as well as any other hash values derived from all or part of the image data stream, may be considered image data stream parameters. These parameters may be combined to generate a verification code. The verification code is then inserted into the private area 58. The recorded verification code is illustratively labeled 70 in FIG. 2.

In an alternative embodiment, the verification code may also be transmitted back to the remote computer. The remote computer may then store the verification code in association with the corresponding image identification code issued for the image data stream. Of course, the corresponding time data may also be recorded by the remote computer in association with the image identification code.

Further, if the user has requested a higher security level, the computer 10 may note the precise time data and location data. Such data may also be inserted into the private area (or transmitted to the remote computer or recorded in the directory 54).

In the preferred embodiment, the private area is encrypted using the key previously generated. If the user has selected an even higher level of security, the segment 68, or the entire image data file, may also be encrypted with the key.

Thus, various levels of security for the image data stream are available. For example, with "Level 1" security, the password is not supplied by the user. A standard password is simply supplied by the computer 10. The key is generated from the image identification code, time data, and standard password. With "Level 2" security, again, the computer-provided password is used, but information regarding each version of the image data file are recorded in the image file directory. In this way, an audit trail regarding changes to the image data stream may be generated. With "Level 3" security, such audit trail data is also recorded. However, a user-supplied password is used, together with the time data and image identification code, to generate a private key to encrypt the private area 58. If further security is desired, all, or a segment 68, of the image data stream may also be encrypted with the private key to create a ciphertext.

If the user requires still further security, after the private area 58 (and possibly at least the segment 68 of the image data stream) have been encrypted using the key, the image data stream (or ciphertext) may be still further encrypted using any one of a variety of publicly known encryption methods. The files associated with the image data stream, such as the image file header 52, image file directory 54, and private area 56 may also be further encrypted.

A currently popular data encryption method, and the one used with the preferred embodiment, is the Data Encryption Standard, or, DES," which was developed and promulgated by the U.S. National Bureau of Standards. With this technique, data is enciphered in 64-bit blocks using a single 56-bit key, as described in NBS FIPS PUB 46, "Data Encryption Standard," National Bureau of Standards (1977). The technique is symmetric, in that the same key is used for encryption and decryption. Alternative encryption techniques, include, for example, the RSA system (named for its developers, Messrs. Rivest, Shamir, and Adleman). The RSA system is described in U.S. Pat. No. 4,405,829 issued to Rivest et al.

After any such additional encryption, the resulting data stream, as well as the associated areas (such as the header 52, directory 54, and private area 58) are recorded on the medium. According to one embodiment, the recording is performed consistent with CCITT (Consultive Committee on International Telegraphy and Telephony) Group 4 standards for facsimile reproduction and transmission.

If a user should alter any image so recorded on the medium, the hash in the private area is unlikely to be the same as the hash or the revised image. In such a case, the present method will not certify that the parameters reflected in the verification code and the parameters of the retrieved document are the same and, thus, will not certify the retrieved image data stream as the original. In contrast, if the parameters of the recorded and retrieved image data streams do match, the present method will allow the computer 10 to certify the retrieved image data stream as an original: that it has not been changed since the time of its recordation on the medium.

Often, however, an author properly wishes to modify or alter a document. Thus, if an image data stream is altered, the entire process discussed above can be repeated, with the computer 10 obtaining a new image identification code and new time data from at least one trusted source. The new image identification code is then also recorded in an image file directory associated with the new image data stream. Further, the computer 10 adds a pointer associated with the new image data stream (preferably in the public directory associated with the new data image stream). The new pointer points to the older image identification code and time data.

Thus, when an older version exists, the directory 54 associated with every image data stream includes a pointer to time data for an earlier version. Consequently, the entire "lineage" of the image data stream may be shown by the computer 10 to the user, either on the screen or on a printout.

The method by which a computer may read a recorded image data stream and confirm that a particular image data stream has not been changed is shown in FIG. 4. The computer 10 initially reads the header 52 to determine the parameters that the computer 10 (or other reader) should use to interpret the document. Upon reading the pointer 60, the computer 10 may then read the image file directory 54. By reading the image file directory 54, the computer 10 knows the image identification code and time data, as well as the location of the private area associated with the image data stream.

If the user has supplied a password, it is accepted by the computer 10. Otherwise, the "standard" password is recalled. The password is then combined with the image identification code and time data to generate a key. The key may then be used to decrypt the private area 58.

The computer 10 also notes any recorded precise time and location data or source indicator. Upon deciphering the private area 58 with the key, the verification code is read. From the verification code, the computer 10 determines the parameters recorded for the previous, or "old," version of the image data stream. The recorded image is also decrypted using, as appropriate, the key, CCITT Group 4 standards, and DES.

After decryption, the image data stream parameters, such as its size and hash values for segments and/or the entire image data stream, may be computed. These parameters may then be compared with the parameters that are recorded in the verification code found in the private area 58. If all such parameters of the image data file stream and those parameters taken from the verification code in the private area are consistent, the computer 10 may then issue a certification that the parameters of the image data stream correspond with the recorded parameters.

Such certification of an image data stream may comprise simply flashing a notification on the monitor 24. In other embodiments, the certification may include printing a certification on the image data stream when it is printed by the printer 42, flashing a light for the operator to see, sounding an alarm, or activating another sensory transducer so as to communicate the certification to the user.

Further, the computer 10 may issue an audit report listing all the prior editions of the image data stream recorded. The report may show, for example, the time, location, or source of any change to the image data stream. As discussed earlier, dates and image identification codes are all linked to one another such that, by reading any one group of date data (and the corresponding image identification code), the computer 10 is pointed to earlier date data and an earlier image identification code. Consequently, the entire audit history may be conveyed to the user via, for example, the monitor 24 or printer 42.

While the present method utilizes a standard format, TIFF, the images are designed to be flexible and non-static entities. These images are an extension of TIFF style images. The image of the present method is designed such that any general Group 4 (facsimile) TIFF viewer should be able to decode the images when only "Level 1" security has been requested. The header, directory, and tags utilized to implement the TIFF format are described in greater detail below.

Image File Header

The header is identical to a TIFF 6.0 header, except that a "signature" added at the end of the header, as shown below.

Image File Directory

The layout of the Image File Directory ("IFD") entries (or "tags") is as follows.

Baseline Tags

All TIFF version 6.0 required tags must be present for an image. This helps ensure that a non-secure image is viewable via a product capable of displaying compressed CCITT Group 4 TIFF records. The required tags for all TIFF images are as follows: Photometric interpretation, compression, image length, image width, resolution unit, X-resolution, Y-resolution, rows per strip, strip offsets, and strip byte counts.

The following additional tags are defined using the described method (in a subroutine entitle Production Link 3.0): FillOrder and ImageDescription. The high level interface includes these tags dynamically when the interface structure has data. ImageDescription contains the image identification code assigned at scan time.

Private Tags

The preferred embodiment uses two registered private use TIFF tags, shown below:

The first tag identified above is used to describe the ISII certification number. The field shall is a TIFF type ASCII sequence. The second tag shown above is used as the entry point for an internal image format description area. The presence of this tag indicates the present method was used to record the data image stream. This tag is in addition to the signature discussed above, which is located past the TIFF header.

ImageDescriptor

ImageDescriptor is a private "sub" image file directory ("IFD") and follows most of the same rules as a general TIFF IFD. This private IFD may be stored in an encrypted format. The following tags are defined for Group 4 ("G4VZN") images produced by Production Link 3.0:

Each of the three tags is discussed further below.

This tag contains a pointer to the Image Identification checksum value block.

Checksum Generation (ISII Verification Data)

The verification data area is a checksum denoting the size of the image data stream, as well as a hash of a particular size segment of the image data stream and a hash of the entire image data stream.

Although the foregoing description will enable a person of ordinary skill in the art to utilize the present invention, a detailed C language source code listing is included as an Appendix. The listing provides detailed information concerning the operation of the overall system. Additionally, detailed features of the preferred embodiment will become apparent to those skilled in the art upon reviewing the source code listing and the detailed comments within it.

Preferred embodiments of the present invention have been described herein. It is to be understood, however, that changes and modifications can be made without departing from the true scope and spirit of the present invention. This true scope and spirit are defined by the following claims and their equivalents, to be interpreted in light of the foregoing specification. ##SPC1##