Outline of Slide Presentation at 2018 NFAIS Blockchain Conference By Patrice A. Lyons (Alexandria, VA; May 16, 2018)

Transcription

1 BLOCKS AS DIGITAL ENTITIES A STANDARDS PERSPECTIVE Outline of Slide Presentation at 2018 NFAIS Blockchain Conference By Patrice A. Lyons (Alexandria, VA; May 16, 2018) SLIDE 1: Introductory page. SLIDE 2: Discuss origins of Knowbot mobile program technology at CNRI in 1980s. An early use of this mobile program technology was in a CNRI project with the National Library of Medicine that enabled a remote user of the NLM databases, such as Medline and Toxnet, to communicate in the Internet using mobile programs. The information represented in digital form was incorporated in computer programs then written using object-oriented methods. The technology was also addressed at a meeting in 1989 at Rockefeller University, as well as at a workshop at CNRI with representatives of the scholarly publishing industry among others. Shortly after, the mobile programs were structured as Digital Objects ref Morris worm; and this later became known as the Digital Object Architecture (or DO Architecture). This led to the DARPA-sponsored Computer Science Technical Reports (CS-TR) project at CNRI in the early 1990s. Work in this period also involved the Library of Congress in experiments on implementations of the technology for registration and recordation of documents at the Copyright Office. Chaining of operations was an important aspect of this work, e.g., management of an author s manuscript, then possible submission to an academic publisher, and later dissemination of the digital object to the public. They were looking past the then standard print on paper, CD-ROM or other physical forms of expression to the challenges presented by the developing Internet. SLIDE 3: Like to step back a bit and put this conversation in the context of the Internet starting in its early days. Describe briefly my work with Bob Kahn & Vint Cerf (credited with creating TCP/IP while at DARPA) at Corporation for National Research Initiatives (CNRI) when the Internet was generally opened to commercial industry ref Rick Boucher bill in the early 1990s. In working on the establishment of the IETF Secretariat, and then the Internet Society, it became clear that a definition of what is meant by the Internet would be helpful. Group involved in developing the Internet gathered at Federal Networking Council in 1995; and the FNC agreed to the following Definition of Internet. Internet refers to the global information system that (i) is logically linked together by a globally unique address space based on the Internet Protocol (IP) or its subsequent extensions/follow-ons; (ii) is able to support communications using the Transmission Control Protocol/Internet Protocol (TCP/IP) suite or its subsequent extensions/follow-ons, and/or other IP-compatible protocols; and (iii) provides, uses or makes accessible, either publicly or privately, high level services layered on the communications and related infrastructure described herein.

2 Importance of recognition that Internet should be allowed to evolve. Mentioning this since, at a meeting of ITU Focus Group on Distributed Ledger Technology (FG DLT) at Bern, Switzerland, in February 2018, a proponent of a specific technical method of structuring a blockchain stated that this would replace TCP/IP. I noted that a better view was to consider this a logical extension of the Internet. It is important to build on the current technology, while integrating new methods of managing information represented as or converted to some digital form of expression. In the implementation of the DO Architecture, a system of proxy servers enables backwards compatibility if this is desired. The DO Architecture is based on the same important architectural ideas that underpin the Internet more generally, the three most important of which are: Open Architecture with defined protocols & interfaces; Independence from the underlying technology; and Minimized complexity for users. SLIDE 4: Present definition of Digital Object from slide. The notion of DO may be viewed as a logical extension of the notion of a packet in a network context. Important difference: an IP address is not persistent as an identifier. Once a packet is delivered there is no way to reference the packet anymore using the IP address. A DO identifier, sometimes called generically a digital object identifier, if properly maintained, is persistently associated with a given DO and many of you may be familiar with the trademark DOI of the International DOI Foundation used widely by the publishing, library, motion picture and other communities. The web and indeed the methods that preceded it uses identifiers that are tied to some aspect of the underlying technology. IP addresses identify specific machines and ports on those machines. The web generally ties its identifiers, called URLs, to specific files on specific machines rather than to the digital information in those files. In this context particular attention should be paid to the resolution mechanism that is deployed to resolve each DO identifier to state information about the DO, including its location(s), access controls, time of creation, public keys (if any), and verification information. Care should be taken to select a resolution system that is trustworthy. From a security perspective a DO can be signed and validated using PKI or other cryptographic methods intrinsic to the overall DO Architecture. At a meeting last year, a presenter, when discussing a particular blockchain implementation, stated that it consisted of SHA256 cryptography as a distinguishing factor: it should be noted that this technology is being used in many DO Architecture implementations as well; and is generally available. Each digital object, or parts of an object, may be encrypted or otherwise protected, apart from the transmission pathway. An observation on Internet technology as it has evolved may help clarify this point. Unlike web and other applications today, where end-to-end security is deployed for the most part, the so-called end/end principle is not a required aspect of the DO Architecture. In this respect, it is important to recall that end/end, or what is sometimes called virtual circuit technology, was not required in the early Internet, but was simply an 2

3 expedient to get started. The DO Architecture introduces a more flexible and efficient approach. SLIDE 5: In preparing my thoughts to share with you today, I recalled a book from my French literature studies called: Le Petit Prince by Antoine de Saint-Exupéry. The raconteur tells a story from his youth where as a six year old, he drew a picture of a boa constrictor digesting an elephant. When he shows the first drawing to adults, they viewed it as HAT. He tries again with an open view of the boa. My thought was that the picture of the boa constrictor digesting an elephant may be compared albeit in a very general way to a digital object. Looking at a digital object, in in essence a set of Os & 1s, a human may conceptually think there is specific information in it, but in reality, the sets of sequences of bits are intended to be parsed by a computer or other information system. In thinking about this, I recalled back in the 1990s when CNRI first developed the DO Architecture, it was presented to a group of about 45 companies meeting as part of CNRI s Cross-Industry Working Team (or XIWT). The group supported the implementation of the DO Architecture as part of the growing Internet (Slide 11, ENDNOTES, Managing Access to Digital Information report dated May 1997). The group addressed business, technical and legal aspects of containers, cryptolopes, packages, or, more generally, digital objects. As I recall, a current software implementation of blockchain technology, called Fabric (see Hyperledger project), continues to refer to containers. In the course of some recent work with CNRI, the implementation of Microcontrollers (MCUs) in the developing Internet of Things (IoT) was addressed from an information management perspective. Of course this is of interest to the groups interested in blockchain software deployments. Without going into the details of any specific project, as a general matter, in the context of MCUs, sometimes called computers on a chip, there is a tendency to talk about program v data in contrasting between what is termed the Von Neumann architecture and the Harvard architecture, or hybrids thereof. I will leave the technical discussion of this distinction to qualified system architects and software engineers; however, from a legal perspective, there is already a definition in the copyright law of computer program. Whether a so-called program is in the same area of an MCU as what is called data, or not, is arguably a distinction without a real difference, particularly where the assembled symbolic logic is executed to produce a specific result. As defined in the copyright law, a program consists of a set of instructions or statements (i.e., data) to be used directly or indirectly in a computer to bring about a certain result. Back to the drawings: when Copyright Office was first asked to register claims in computer videogames, while there were one or more works incorporated in the computer program, including pre-existing audio-visual and graphic art works, it was viewed as a derivative 3

4 computer program that based on or incorporated the pre-existing works. Much as a person viewing only the HAT wouldn t see the elephant being digested. In the late 1980s, the Office also considered the status of computer programs that, when executed resulted in a particular typeface. Today this would also include bar codes. In its initial rulemaking, the Office required the disclaimer of typeface elements embedded in a program. By 1992, however, the Office decided that this approach was not advisable and allowed for registration of a computer program as a literary work, even though typeface resulting from the execution of the program may be in the public domain. For further details on this matter, please see the Copyright Office Compendium of Office Practices, Ch SLIDE 6: Since time is limited, go more quickly through the remaining slides. From a standards perspective, CNRI worked with ITU-T, Study Group 17 (Security) for several years on preparation of what was approved in September 2013 as ITU-T Recommendation X The Digital Object Architecture formed the basis for the draft Recommendation when it went out for balloting in the summer of Only one government was opposed since it preferred a more neutral terminology such as virtual object. As approved, the concept digital object was replaced by digital entity and I will refer to this in the remainder of my talk. Ref again to meeting of FG DLT earlier at Bern. CNRI submitted a contribution (see ENDNOTES) requesting that a block be defined as a digital entity as described in X This was agreed and digital entity as described in X.1255 was noted in connection with the definition of block in the base terms that emerged from the meeting. The DO Architecture, as such, has also recently been the subject of negotiations at ITU-T, Study Group 20, Internet of Things and Smart Cities. This is an evolving situation. SLIDE 7 & 8: Let s step back a moment and consider generally the concept of block. Historically, a block, also called a data structure, was viewed essentially, as a sequence of bits, usually with a defined beginning and end and may or may not have been uniquely identified other than, perhaps, by its arrival sequence in time, and often involved encryption. Here I am viewing the picture of a box containing the drawing of a sheep as very roughly like a block very roughly. There may be information in digital form incorporated in this form of expression, but it is of a purely conceptual nature as is the drawing of the box itself until fixed in some tangible form, and there has been discussion over the years of recognizing a digital object as equivalent to a physical object insofar as the requirement of a copy is concerned, but that is a discussion of another day. SLIDE 9: A block may be considered a digital entity that is configured using specific methods; and there are many issued or pending patents on various ways to accomplish this. Bob Kahn and I wrote an article: Representing Value as Digital Objects, that was first published in D-Lib Magazine in 2001 (later republished in 2006 see citation in ENDNOTES). 4

5 It introduced the application of the DO Architecture in various then paper-based entities such as bills of lading, wills, and contracts. Show demo re one-dollar data body. Representation of value. The chaining of one digital entity to another was introduced. For example, if the Federal Reserve represents a one-dollar value in digital form, and structures this information as a digital object with an associated persistent identifier, and then a hash of this identifying information is incorporated in the digital entity when it is issued to a bank, and here what is termed consensus is achieved in various ways, including regulations and contract, it may be viewed as a type of blockchain; however, a more flexible approach would simply to view this as a one-dollar bill in commerce. There is often confusion here about tying the representation of value and chaining of operations to the latest jargon. Based generally on the above-mentioned article, CNRI filed a patent application in April 2003 that was later rejected by the USPTO on the basis that the claims were covered by CNRI s earlier issued 1993 patent (see reference in ENDNOTES). In light of the many, many patents issued, or applications pending related to blockchain technology, this prior art may be important to users, as it may serve to provide a public domain basis for developing their blockchain or other digital entity-based applications. The patent situation may be quite murky for many years. SLIDE 10: In conclusion, mobile program technology, whether structured as blocks, or, more generally, digital entities, may soon be needed in the context of implementations of the DO Architecture. While techniques like layering or end/end interaction will continue to have utility, a more integrated perspective is appropriate where autonomous devices are widely deployed that are capable of acting for users, or more likely, computer programs acting on their behalf, to produce desired results. The complexities and sheer volume of information represented as or converted to digital form that will be available requires a new paradigm for information management and the Digital Object Architecture should provide a sound basis for moving forward. 5