Coalition for Electronic Markets
Brad Cox
George Mason University
bcox@gmu.edu; 703 968 8229 Voice; 703 968 8798 Fax
A 21st century infrastructure would address many practical problems. For example, the government can serve as a catalyst for the private sector development of an advanced national communications network, which would help companies collaborate on research and design for advanced manufacturing; allow doctors across the country to access leading medical expertise; put immense educational resources at the fingertips of American teachers and students; open new avenues for disabled people to do things they can't do today; provide technical information to small businesses; and make telecommuting much easier. Such a network could do for the productivity of individuals at their places of work and learning what the interstate highway of the 1950s did for the productivity of the nation's travel and distribution system.A TECHNOLOGY POLICY FOR AMERICA Six Broad Initiatives Bill Clinton; September, 1992
"We believe the creation of a national information infrastructure must be a national priority, and we are willing to work in partnership with the government to see that it gets done. The development of an information infrastructure will raise the standard of living for all Americans and enable our country to prosper in a competitive global economy."
|
| Figure 1: The top two circles depict two kinds of property that have existed since antiquity. The bottom circle depicts a new kind of property that never existed until the emergence of interconnected computers within the last few decades. |
However, these quotations do not address a crucial difference between the manufacturing age infrastructures of the past and the emerging information age infrastructures of the future. The highway, canal and railroad infrastructures of the past carried the tangible goods of the manufacturing age. Tangible goods consist of atoms that are governed by physical laws of conservation of mass, energy and spin. These laws provide a physical basis for commercial transactions, even such fundamentals as what it means to `buy', to `sell', and to `own'.
However the electronic goods of the information age are made of bits, not atoms. Bits are electronic ephemera that can be copied in nanoseconds and transported at literally the speed of light. Since physical conservation laws do not apply for bits, this undercuts the market processes that underlie the specialized and distributed economy upon which the phenomenal achievements of the manufacturing and engineering community have been based.
By `electronic goods', we mean the intangible electronic properties that reside on computer networks such as Internet, Prodigy, America Online and the NII (see Figure 1). Widely recognized examples of electronic goods are computer software, clip art, font libraries, electronic documents, reusable software components, and so forth; the intangible goods of the information age. Such goods are traditionally categorized as `intellectual properties'. However they are fundamentally different from the abstract properties addressed by patent, copyright and trade secret laws. Electronic goods reside externally, outside the human mind on computer networks. They are accessible to technological monitoring and metering in a way that true `intellectual property' is not. Such metering could provide a basis for commerce in electronic goods analogous to the physical conservation laws that underlie commerce in tangible goods.
We propose that the Department of Commerce establish an ATP program to build and deploy a market infrastructure that would address the ownership and revenue collection issues of information age commerce. By integrating computers and communications technology with financial institutions as outlined in this document, the Department of Commerce could remove the most crucial obstacle to the often-predicted transition to an information economy becoming more than a trendy figure of speech.
This document proposes that the U.S. Department of Commerce establish an ATP program to build and deploy a commercially robust infrastructure for commerce in information age goods. This infrastructure would address information age ownership issues like these:
 |
| Figure 2: The integration of computers and communications technology with financial instutions creates the possibility of invocation-based market mechanisms, in which revenue is collected when goods are invoked instead of when replicas are acquired. |
The key insight in this proposal is that copy protection is not a robust basis for commerce in goods that can be copied and transported without cost. For easily copied goods, it is technically far easier to base ownership on copy projection, in which users are encouraged to acquire software without charge. The owners acquire revenue as the end-users use their goods, not when they acquire them.
The tangible goods of the manufacturing age are made of atoms that abide by physical laws of conservation of mass, energy and spin. This means that when a customer acquires a copy of a tangible hardware device (for example, a memory chip or a computer), they are unable to replicate additional copies. Thus scarcity has emerged as the traditional basis for commerce.
This traditional basis is robust for tangible hard-to-copy goods. By robust, we mean that this mechanism supports the broad and deep structure of production characteristic of advanced engineering societies. Producers at every level of the structure of production know that they will benefit whether their goods are consumed directly or indirectly. For example, a chip producer knows that revenue will accrue not only from those who use chips directly, but also indirectly as end users buy goods that use chips as internal components.
But since no comparably robust revenue collection mechanism exists for easily-copied computer-based information age goods, a comparable structure of production has never emerged (see Figure 3). The electronic goods of the information age are made of bits that can be copied in nanoseconds and transported at literally the speed of light. Since conservation laws do not apply for bits, this undercuts the market processes that support the commercial structure so characteristic of hardware engineering today; the very structures upon which hardware engineering has achieved the maturity to which software engineering and computer science can only aspire.
Electronic objects differ from tangible objects by being unable to monitor their acquisition but trivially able to monitor their invocation. For example, it is easy to make software count how many times it has been invoked, but hard to make it count how many times it has been copied. Rather than continually fighting this intrinsic characteristic with legal/moral restrictions and copy protection technologies, we propose that revenue collection be based on the invocation of those copies within a computer. We call this approach an invocation-based market mechanism to contrast it with the acquisition-based mechanism upon which commerce in tangible goods is based. By basing revenue collection on invocation instead of replication, the easy-to-copy nature of information age goods becomes the basis of a solution, as distinct from endlessly fighting it with technical, legal and moral restrictions.
Figure 3 shows a multi-level structure of production for electronic goods that have been instrumented for distribution via this approach. Each level of goods in this figure are coded to invoke a pair of invocation monitoring instructions called query and commit. Each instruction bears a variable-length argument list with certain mandatory arguments to identify the goods that are enclosed by this pair of instructions. Query is a predicate that governs whether this invocation will be allowed. Commit indicates that the current invocation has been successfully completed and should be recorded as a basis for subsequent billing.
When such goods are subsequently distributed and executed on the end-user's computer, the time sequences of the query and commit instructions (with their arguments) are recorded in a computer-resident tamper-resistant cache. Periodically (say monthly) the cache is uploaded to a financial organization for billing. The financial institution uses Terms and Conditions Algorithms (TCAs), supplied to them as executable programs by the goods' owners, to convert the raw invocation data monetary amounts due.
A settlement process then ensures in which these amounts are electronically debited from the end-users' account and credited to the account of the top-level component's owner (the electronic bookstore in Figure 3). The settlement process continues by debiting the account of the bookstore owner by the amount owed to the owner of the multimedia document, and so forth throughout the producer-consumer hierarchy. Account imbalances beyond some threshold are periodically reconciled in the familiar fashion; by mailing bills or checks as appropriate.
|
| Figure 3: Information age goods are assemblies of smaller components, just as manufacturing age goods like automobiles are assembled from pre-fabricated engines and crankshafts, and these in turn from raw materials. |
Invocation-based charging does not mean that charges are closely proportional to usage; often known as pay-per-use or pay-per-view . It only means that invocation records are used as the basis for whatever terms and conditions the vendor of a component chooses to provide. Each component vendor expresses its business conditions as a Terms and Conditions Algorithm (TCA). It provides this algorithm to the billing agency before releasing the product to the market. The billing agency runs these Terms and Conditions Algorithms in order to convert invocation records into monetary amounts.
These algorithms can easily support the familiar pay-to-own terms and conditions that dominate today's shrink-wrap market. For example, pay-to-own involves charging a fixed fee the first time the product is used by a new customer. But different algorithms can as easily support terms and conditions that are not viable options today. For example, pay-if-you-like-it involves a different algorithm that delays the pay-to-own fee from the first to the n'th use. If the vendor so chooses, he can also provide algorithms that implement radical incremental charging options such as pay-per-use, pay-per-minute, pay-per-keystroke. Such terms and conditions might be popular for customers who prefer that charges be closely related to the software's measured utility.
Although terms and conditions can remain the same or be radically different, software distribution becomes much simpler. Since revenue is now based on usage instead of on acquisition of replicas, there is no longer a need to restrict copying. With invocation-based charging, software vendors actually gain new paying customers regardless of how the copy was acquired.
Although this was not anticipated as a benefit by the originators of this proposal, the proposed infrastructure's ability to disable further access to software if account balances are not within threshold could also be used to disable further invocations of dual-use software on a client-specific or even geographic basis. This could be useful during wartime, for example.
The best indication of industry's commitment to a solution to this important but controversial issue is the history of the Coalition for Electronic Markets. Paul Kozemchak of the ARPA policy office called Dr. Cox last spring in response to an editorial Dr. Cox had published regarding this issue[1]. He asked that we organize a coalition of government, academia and industry to deploy a solution with the aid of matching funding from the ARPA Technology Reinvestment Program. Although there was very limited time available for organizing such a coalition, the charter members listed below committed to almost $1.5M in in-kind contributions to be matched by ARPA for a project total of $3M. (ARPA has not yet announced whether CEM's proposal will be funded.)
This metering per use idea makes a great deal of sense in a manufacturing environment. Charging by the CPU is extremely inefficient, and leads to curious behaviors on the part of the customers. Charging by the application makes people run their factories with only part of the functionality available. Charging by use makes more sense. It would be desirable to ship the same software, but charge the larger company more for its use than the smaller company. Since the software model knows the size of the factory it should derive the licensing fee. Let the customer with 100 machines pay more than the customer with 10 machines. Let the customer who runs many applications pay more than the customer who runs fewer. This will create a healthy tension where the users are trying to minimize the software use, and the developers are trying to tempt them with features they need. Every marginal user's needs will be met, because the developers want to add them to the regular user base. I believe that this will lead to better service, and better products.
The Society of Information Management is the professional organization for more than 2,700 senior information technology executives around the world. Its 1992 member survey identified software licensing as a critical information technology issue. The following quotes are from the Software Licensing Position Statement of April 1993 in which this society summed up the view of software consumers within large corporations of means for resolving their difficulties with software licensing.
The ideal software vendor and customer partnership of the future will be based on an understanding of mutual issues, trust and the need to balance revenue stream generation for vendors with cost control for customers. As organizations move toward multi-vendor, multi-platform, open systems environments, the pricing models of today must give way to more flexible and equitable licensing models and allow for more regional and global focus on the part of customers.
The evolution toward rapid application development technology further adds to licensing fees. Licensing models have not kept pace with this evolution.
The evolution of desktop software towards the use of object-oriented technologies will transform the "Application-Centric" environment to one which is "Document-Centric." This will further add to pressures for licensing which addresses software integration trends.
Vendors must have the ability to monitor customer usage. Fair and equitable licensing models should ensure vendors' revenue streams are providing cost advantages to customers.
Their recommended approach is consistent with that of this proposal. The difference is that this proposal addresses this issue on an industry-wide basis and not exclusively from the perspective of a particular constituency. SIM's president, Steve Coppell, responded to our request for a statement to enclose with this proposal as follows:
"The Society for Information Management (SIM) has addressed changes and issues regarding software technology and recognizes the software monitoring issue as one of the obstacles preventing this technology from moving forward. SIM is closely watching with interest the proposal put forth by the Coalition for Electronic Markets."
MountainNet is engaged in the AdaNET Research Project, funded by NASA, administered through the Johnson Space Center and the University of Houston Clear Lake, under the Repository-Based Software Engineering (UHCL/RBSE) program. The purpose of the RBSE Program is to promote widespread adoption of modern software engineering, improve software quality, and reduce system life-cycle costs, across the public and private sectors of the economy. Bob Terry's responded to our request for a statement to enclose with this proposal as follows:
"I spoke with my company's President (Mike Digman) and received his approval to endorse the approach the Coalition for Electronic Markets is proposing to ARPA. As I stated, we are currently attempting to monitor the access and usage of the objects in the AdaNET repository in order to provide feedback for authors, clients, and project management. We believe that your proposed work in these areas could provide us valuable technical assistance. We are interested in being involved with your proposed project in any possible manner."
This means that the Department of Commerce has a unique opportunity to lead the way into an entirely new form of commerce. The major components of this approach (computers, communications and financial institutions), already exist. So do the lower level technologies that would be involved in a robust deployment of this approach (public key encryption, tamper-resistant silicon and/or software, communication protocols, graphical user interfaces).
The ATP program could play a key role, not only in government's traditional role as a source of funding, but also by providing the seed crystal into the super-saturated solution around which an entirely new form of commerce might crystallize.
The challenge of transitioning to this nation to an information age economy is similar to the challenge that such innovators as Eli Whitney, Roswell Lee and John Hall surmounted almost two hundred years ago at the dawn of the manufacturing age with the determined support and assistance of the U.S. Congress and the Bureau of Ordnance.
If the software industry is ever to achieve the long-awaited breakthrough in software engineering productivity via a robust market in off-the-shelf reusable components, a crucial first step is to deploy a revenue collection mechanism capable of supporting the evolution of a multi-level structure of production for computer-based information age goods. A solution to the revenue collection issues as proposed here is arguably the most crucial step towards the information age becoming more than a figure of speech.