by Robert E. Kahn
This document proposes the development of a testbed for deposit, registration and recordation of copyright material in a computer network environment. The testbed will involve the Library of Congress and provide for electronic deposit of information in any of several standard formats, automated submission of claims to copyright, notification of registration and support for on-line clearance of rights in an interactive network. “Digital signatures” and “privacy enhanced mail” will be used for registration and transfer of exclusive rights and other copyright related documents. Electronic mail will be used for licensing of non-exclusive rights with or without recordation. Verification and authentication of deposits can be carried out within the testbed using the original digital signatures. A system of distributed redundant “Repositories” is assumed to hold user deposits of electronic information. The testbed provides an experimental platform for concept development and evaluation, a working prototype for system implementation and a basis for subsequent deployment, if desired.
INTRODUCTION AND BACKGROUND
Deposit, registration and recordation of copyright material and its associated claims to rights have generally been handled manually. Over the past two decades, the economics of information technology has enabled an electronic foundation for such material and claims. The key elements of this foundation are the personal computers, workstations, computer networks and peripheral devices such as scanners, printers and digital storage systems which have now become sufficiently powerful and cost effective to be put into widespread use. It is now essential that the underlying systems used to manage copyright be conformed to be compatible with the promise of this new computer networking environment. This paper addresses several essential steps that should now be taken to facilitate that process.
In the current manual system, claims to copyright are registered with the Copyright Office, Library of Congress. Deposits are accepted and stored in physical form including tapes and diskettes as well as paper and other substances. Notification of registration is also made in physical form. In addition, documents transferring copyright ownership and other documents pertaining to copyright may be submitted to the Copyright Office for recordation. While an on-line record of recent registrations and recordations may be accessed at the Copyright Office, there is only limited external dissemination of this information in electronic form for access at remote sites.
This approach requires considerable physical storage at the Library of Congress for deposited materials which can only increase over time. Materials stored in physical form will slowly degrade unless deposited in digital media in which case the contents may be reproduced subsequently without loss of information but at some cost for duplication. Even if it is available digitally, much, if not most, of this material will not generally be accessible on-line from any source. Rights to use the information in a computer network environment cannot usually be acquired easily or quickly, even if the identity of the rightsholder is accurately known. Fortunately, these limitations can also be overcome with the use of information technology and only minor modification to the current manual system.
COMPONENTS OF THE PROPOSED SYSTEM
This document proposes building a testbed to develop and evaluate key elements of an electronic copyright management system. These elements include:
- a. Automated copyright registration and recordation
b. Automated transactional framework for on-line clearance of rights
c. Privacy enhanced mail and digital signatures to facilitate on-line transactions
d. Methodology for deposit, registration, recordation and clearance
Current registration and recordation activities of the Library of Congress would be maintained and enhanced in the proposed testbed. It provides for repositories and recordation systems both within and without the Library of Congress, which would serve as agents for authors and other copyright owners which seek to register works with the library. In addition, the testbed provides for automated rights clearance, outside of but linked to the library, which would accelerate permissions and royalty transfers between users and rightsholders.
Electronic Copyright Management Testbed
A testbed is proposed to develop and evaluate these concepts and to obtain experience in the implementation and operation of an experimental system (see Figure 1). The proposed testbed consists of a Registration and Recording System (RRS), a Digital Library System (DLS) and a Rights Management System (RMS). The RRS will be operated by the Library of Congress and will permit automated registration of claims to copyright and recordation of transfer of ownership and other copyright related documents. The RRS would also provide evidence of “chain of title.” The DLS will be a distributed system involving authors, publishers, database providers, users, and numerous organizations both public and private. It will be a repository of network accessible digital information and contain a powerful network based method of deposit, search and retrieval. The RMS will be an interactive distributed system that grants certain rights on-line and permits the selective use of copyright material on the network.
The testbed would contain a digital storage system connected to an applications gateway (which is, in turn, connected to multiple communication systems including the Internet) to which documents would be submitted. The storage system would constitute an experimental repository for information. The applications gateway would be designed to support multiple access methods including direct login. The RRS and RMS would be servers connected to the Internet. Initially, they would be on a common machine, but they could later be easily separated. After development, the RRS would be relocated to the Library of Congress or its designated agent prior to being placed in operation. After initial implementation, the repository and the RMS would be replicable at other sites.
Electronic Bibliographic Records
An electronic bibliographic record (EBR) is created by the user for each digital document submission and supplied with the document for registration. The EBR is also suitable for use in cataloging and retrieval. The EBR may be supplied to other systems without the actual document but with a pointer to it. The EBR must contain a unique name for the document per author. If a name is provided that has already been used by the same author, it will be rejected with an explanation. An acknowledgment of deposit will be returned to the user along with a unique numerical identifier and a retrieval pointer to the document, and, in the event of a claim to copyright, a certificate of registration from the RRS.
When the EBR indicates a claim to copyright, the RRS will be supplied a copy of the EBR by the repository along with a digital signature (to be described shortly) that can be used to verify the accuracy of a deposit at a later time. The actual work would remain in the repository. The digital signature consists of a few hundred bytes of data and is approximately the size of the EBR. It should allow the authenticity of the retrieved document to be formally established at any time for legal and other purposes.
The RRS need not be collocated with a repository. It is expected that an operational RRS would be operated by the Library of Congress. The repositories would be operated by the Library of Congress as well as other organizations or individuals. Deposits in certain qualified repositories will constitute deposit for public record purposes. The Library of Congress will maintain its own repository of selected deposits.
Although a set of distributed repositories is envisioned for a widely deployed system, the proposed testbed will only have a single repository for experimentation. The repositories would be established in such a way as to insure the survival of the deposited information with perhaps different degrees of confidence (much like the treasury, banks and brokerage houses, for example). Certain information would probably not be deposited for purposes of registration and might be stored at the users local site or in a commercial repository. Highly valued information could be stored in rated repositories (5-star down to 1-star) with varying degrees of backup and corresponding costs. The most critical information, as determined by Copyright Office regulations, might be stored at the Library of Congress or the National Archives as a safeguard. The structure of such a system of repositories should be developed as part of the project.
The advantages of a distributed repository system are:
- Large amounts of physical storage is not required to be made available at the Library of Congress.
- Access to the original documentation is guaranteed by the DLS to the confidence level selected by the user’s choice of repository (again like the banks).
- Repositories serve as interfaces to the users, thus offloading and insulating any central servers and systems such as the RRS from potentially large user loadings and specialized customer service requests.
- Access to the RRS in transaction mode is available only to authorized repositories and RMSs that are qualified to use the RRS in that mode. An individual author, a collective licensing organization, a government or corporate entity or others may run an RMS. Authors and other copyright owners, as well as users may also connect directly to the RRS through a separate interactive user interface.
The Computer Network Environment
There are three specific actions of concern in a network environment. One is the movement of information already contained in a computer network environment thereby greatly facilitating the creation of multiple copies in multiple machines in fractions of a second. The second is the importation of external information, such as print material or isolated CD-ROM based material, which must first be scanned or read into the system before it can be used. The third is export of internal network based information to paper using digital printers or facsimile machines or copied to separable media such as tape or DAT for external transport to others. Some of these actions, such as local use on paper in very small quantities, may or may not be covered by fair use provisions. However, non fair use actions would require approval of rightsholders.
In addition to the above three actions, there is a fourth action that is facilitated by the computer network environment. Information in digital form has the property of being easily manipulated on a computer to produce derivative works. Such derivative works can also be easily moved about in a computer network environment and be subject to further manipulation by other parties. The technology makes it possible for parallel and concurrent manipulation of such information to result in an exponential proliferation of such derivative works.
Rights Management System
The four actions described above form a basis for a rights management system. In general, there will be many such systems operated by rightsholders or their agents for required permissions on either an exclusive or non-exclusive basis for a given type of action. To locate an RMS, a user requires the existence of a domain server that knows about the network names and addresses of all RMS servers. Transactions involving rights may be handled by direct exchange on-line between the user system and the corresponding RMS. Typically, this exchange would occur rapidly on-line, and we refer to this as the interactive clearance of rights. Privacy enhanced electronic mail would be available for exclusive licenses and other transfers of rights. Non-exclusive licenses might be handled by regular electronic mail.
Transfer of copyright ownership would usually involve recordation in the RRS and could conceivably be handled automatically by the RMS on behalf of the rightsholder and the user to facilitate matters. The confirmation from the RRS would also be passed back to the rightsholder and user directly or via the RMS using privacy enhanced mail. Various enabling mechanisms in the normal screen-based computer interface could be developed and invoked by a user to achieve rapid clearance. If included in the user interface, this capability would have the effect of creating an instant electronic marketplace for such information.
Recordation is defined to mean the official keeping of records of transfers of copyright ownership and other documents pertaining to copyright by the Copyright Office, Library of Congress. For legal purposes, proof of official registration of claims and recordations will be provided by the Copyright Office (via the RRS). Other registrations (at repositories) and non-exclusive licenses (via RMSs) will be certified by privacy enhanced mail. It will be up to the parties to such registrations and recordations to maintain electronic records of their transactions. These could also be stored within the DLS.
The electronic copyright management system actually requires several types of domain servers. First, documents can be easily retrieved via the DLS if the citation is accurately known or through one or more search and browsing processes otherwise. However, the mapping of a bibliographic pointer (to the designated repository) into its network name and address may require a separate server. Second, the above mentioned domain server for RMSs is needed. Third, the date and time that transactions have been requested and taken may need to be formally validated. An electronic notary and time server would provide such a capability as part of the privacy enhanced mail system.
Retrieval, Appearance and Submission of Documents
Retrieval of documents from the DLS is generally a two-step process. The initial step is to identify the document and to retrieve its EBR. This record will also identify the rightsholder and any terms and conditions on the use of the document or a pointer to a designated contact for rights and permissions. Rules would have to be formulated and posted to inform clearly what obligations a user incurs when accessing the system. For example, it may be specified that a submitted request with a valid EBR will then be taken to mean acceptance of the terms and conditions, including any implementation and usage restrictions or payment requirements. The rightsholder may also wish to place restrictions on the appearance of documents for certain uses.
As part of the process of document submission, a valid EBR will have been produced which can be used in the author’s system. Each author or other owner of copyright (or such owner’s successor in title or duly authorized agent) will maintain his or her own collection of EBRs. Searches and requests will typically be made to the user’s home system unless the rights have been transferred or delegated elsewhere (e.g. to a publisher, agent, or database provider). In applying for registration of claims to copyright at the Copyright Office, a user could be required to certify that he or she has the rights to the material and sign the submission digitally.
PRIVACY AND AUTHENTICATION TECHNOLOGY
This section briefly describes several key technologies to handle privacy and authentication in the digital network environment. Four such technologies are described below, namely: 1) Public Key Cryptography, 2) Digital Signatures, 3) Privacy Enhanced Mail, and 4) Notarization.
Public Key Cryptography
In conventional cryptography, a mathematical function and a “secret key” are shared by parties who wish to communicate confidentially. Each message to be sent is “encrypted” using the function and key and the recipient(s) “decrypt” it using the same function and key. This may be thought of as sharing a locked box in which several individuals have the key and any of them can lock or unlock the box at will.
In the late 1970’s, two Stanford University researchers, Martin Hellmann and Whitfield Diffie speculated that it might be possible to devise paired cryptographic functions which had the interesting property that one function would encrypt and the other would decrypt. In fact, the concept was slightly more sophisticated in that any message encrypted with either one of the functions could only be decrypted by the other. In other words, having access to the function which did the encrypting does not help when it is time to decrypt. Using the box analogy, the public key cryptography system would be like having a box with a two- key lock. If one of the keys is used to lock the box, the other must be used to unlock it. A person holding a key used for locking could not use it for unlocking.
One of the biggest problems with conventional cryptography is that the keys must be kept secret and must be distributed by secure means. The notions of Hellman and Diffie opened up a new way of thinking about key management. One key could be made public (e.g. the one to be used for encryption) and the other kept private. Anyone knowing the public part of a pair of keys could use it to prepare a message which would remain confidential until the person knowing the private key used it to decrypt the message. The public keys could be listed in public directories without any special protection since knowing them did not help anyone decrypt messages encrypted using the public key. This feature makes it far simpler to manage key distribution since the public part need not be protected.
Three researchers at MIT, Rivest, Shamir and Adelman developed a pair of functions meeting the requirements specified by Diffie and Hellman. These functions are now known as the RSA algorithms (from the last names of the inventors).
Since either key of a public key cryptography pair can be used to perform the initial encryption, an interesting effect can be achieved by using the secret key of the pair to encrypt messages to be sent. Anyone with access to the public key can decrypt the message and on doing so successfully, knows that the message must have been sent by the person holding the corresponding secret key. The use of the secret key acts like a “signature” since the decryption only works with the matching public key.
Buyers could send digitally signed messages to sellers and the sellers could verify the identity of the sender by looking up the public key of the sender in a public directory and using it to verify the source of the message by successfully decrypting it.
Privacy- Enhanced Mail (PEM)
Public key cryptography can be combined with electronic mail to provide a flexible way to send confidential messages or digitally signed messages or both. In actual practice, a combination of public key, conventional secret key and another special function called cryptographic hashing is used to implement the features of privacy- enhanced mail. The public key algorithms require a substantial amount of computing power compared to conventional secret key algorithms. The older secret key algorithms, such as the Data Encryption Standard (DES) developed by the National Institutes of Standards and Technology (NIST), are much more efficient. Consequently, confidential messages are typically encrypted using a conventional secret key which, itself, is sent, encrypted in the public key of the recipient. Thus, only the recipient can decrypt the conventional secret key and, eventually, decrypt the message.
To send digitally- signed messages, each message is run through a “hashing” algorithm which produces a compressed residue which is then encrypted in the private key of the sender. The message itself is left in plain text form. The recipient can apply the same hashing algorithm and compare the compressed residue against the one that was sent (after decrypting it with the sender’s public key).
One of the basic problems with this application of public key cryptography is knowing whether the public key found in the directory for a given correspondent is really that correspondent’s key or a bogus one inserted by a malicious person. The way this is dealt with in the Privacy- Enhanced Mail system is to create certificates containing the name of the owner of the public key and the public key itself, all of which are digitally signed by a well- known issuing authority. The public key of the issuing authority is widely publicized so it is possible to determine whether a given certificate is valid. The actual system is more complex because it has a hierarchy of certificate issuers, but the principles remain the same.
Using digital signatures, it is possible to establish an on- line notarization service which accepts messages, time- stamps them and digitally signs them, then returns them in that form. If the person desiring notarization digitally signs the message at the time it is sent to the notarizing service, then it will be possible, later, to establish that the person requesting the notarizing had the document/message in question at the time it was notarized. One can imagine that the originator of a message might have it notarized for the record and the recipient might independently do so. By this means, for instance, evidence of a contract’s existence in the hands of each party at particular times might be established.
VERIFICATION, AUTHENTICATION AND CERTIFICATION
The verification process uses stored digital signatures to ascertain whether a given copy is identical to the version which was originally deposited. If any portion of the copy differs from the original, the verification process will fail. Authentication or formal certification of deposits may be provided to a requesting party in traditional ways or via electronic mail. Privacy enhanced mail would be used to certify the authenticity of a deposit, as well as to certify registration and recordation records, for legal purposes.
The deployment of an electronic deposit, registration and recordation capability for use in a computer network environment would greatly facilitate and accelerate the move to a network base for information creation and dissemination. The system would be compatible with the current manual system and would support the ability of the Library of Congress to provide automated registration and recordation services. It would provide a foundation for straightforward and easy expansion and evolution and provide a direct linkage for the Library of Congress to the DLS. It would provide a prime working example for all other kinds of activities where claims registration and rights management come into play. Verification and authentication of copies of deposits may be performed electronically using digital signatures. Formal certification of deposits, as well as registration and recordation records, using privacy enhanced mail may be provided for legal purposes. A testbed which demonstrates the relevant concepts and ideas can be implemented within a two to three year period with initial limited use within a year.
Robert E. Kahn, Ph.D. President Corporation for National Research Initiatives Suite 100 1895 Preston White Drive Reston, VA 22091-5434