Ethics and Law on the Electronic Frontier
November 23, 1994
With every passing day, the Internet becomes increasingly commercialized. Vendors are being drawn to the Internet to capitalize on a market of 30 million potential customers. Today, most of these companies either provide information for free, or require their customers to telephone in their orders. Although these models of transaction work today, they are not conducive to the growth of the Internet. Neither is the rampant piracy of information. Nor is the lack of secure efficient methods to transfer money electronically. We must fix these problems before the birth of the National Information Infrastructure. The solution is a system of secure, private, electronic money transactions. Right now, techniques to provide electronic monetary transactions are in their infancy and software implementations are few and far between. For these techniques to mature, they must be tested, refined, and accepted by a critical mass. In this paper, I will discuss traditional techniques used for electronic monetary transactions, justifications for new the creation of a new system, and modern techniques for electronic monetary transactions.
Electronic monetary transactions have been available since the inception of email. People could buy goods by sending their credit card information in email. With the advent of network sniffers, and monitoring of email, this primitive technique has become obsolete.
This setback has not blocked the usage of electronic monetary transactions. Major commercial on-line services like CompuServe, Prodigy, and America On-line offer electronic shopping today[Apple]. These services allow customers to browse on-line catalogs, monthly magazines, and CD-ROMs, then place orders via the modem. They do so by use of propriety software that minimizes time when credit card information propagates through the telephone lines. When first joining these services, users register their address and credit card information for billing purposes. Next, they send this information over the telephone lines in encrypted form. In the future, when customers choose to purchase goods, they send a buy request over the modem. Since the on-line service already has the billing information, it does not need to be re-transmitted.
Although the former system works well over modems, it does not scale to use for general purpose transactions on the Internet. Today, it is possible to use data encryption to secure email from being corrupted or read by prying eyes. If two parties wished to securely transfer credit card information over the Internet, they can participate in Diffie-Hellman key exchange to get a session key. They can use this key to encrypt the credit card information. Work to automate this process is being done by Netscape and First Data [First] and Microsoft and Visa[Visa2].
Two problems with the former technique are that they require a secure repository for public keys, and use of cryptology faces stiff opposition from the NSA[Will]. Nonetheless, the efforts of Netscape and Microsoft are a step in the right directions. Their work will provide a mechanism for people to send credit card information over the Internet to large companies.
The ability to securely send credit card information over the Internet fills the need for transactions between individuals and large companies. But this is not the only party that needs to trade over the Internet. People should be able to trade money among themselves. The advent of secure electronic money can also alleviate problems of software piracy.
Software piracy is a huge problem. This problem has grown so large that companies like Microsoft and Describe Inc. offer awards totaling $30,000 for the capture of some pirates[Rewards]. And Microsoft routinely cooperates with Federal authorities to stop illegal software distribution. Apart from legal means, companies, most notably the people who make Canvas, Spectre VR, and X-Win, employ technology to combat software pirates. With their systems, each time a copy of their software loads, the software sends a message to the local network. If another copy of the software with the same serial number is currently loaded, the duplicate copy cannot run. Problems with net lag prevent this idea from being extended to include all computers on the Internet.
Microsoft and Novell are debating plans to introduce software with new pricing scheme. Instead of paying for software when you receive the disks, they will charge a nominal fee each time interval the software is in use[Foley]. Of course there is a ceiling for maximum price. The main problems these two companies are ironing out are the time logging measures and payment capacities. If these two problems are solved with electronic payment schemes over the Internet, software piracy on Internet machines will vanish. When people run pirated software on machines attached to the Internet they will still have to pay the per usage fees. Secure, fast, electronic transactions will enable software vendors to distribute their goods over the Internet without fear of piracy.
There is no doubt that large companies will flock to the Internet. Since these companies can establish accounts with credit card companies, credit card numbers are a valid method of payment. For small companies and individuals this is not the case. The Internet provides opportunities for anyone to compete by offering an unp! recedented gateway for distributio n. Entrepreneurs will make upshot services. Secure electronic money will allow people to establish enough trust to use their services.
Ziff-Davis, Wired, and many other publications freely distribute their articles on the Internet. If this practice is reduces their profits (from people who would no longer buy the physical magazines), this service will cease. Record labels, especially those who sue on-line services complain over the ever increasing trend of digitally pirating records, and will stay away from the Internet until solutions form. Unlike the Rolling Stones, who chose to give a free concert[Globe], other groups will be hesitant to mimic this act without a pay-per-view scheme. What is the solution to this problem? Once again, the answer is fast, secure electronic money.
The solution to the former problems is to use electronic money with
digital transactions. Good implementations of electronic money are
completely software based, provide security over hostile networks,
unforgeable, convenient, have short transaction time, and usable be
any group. In the following section, I will explore the basic
premises of electronic money, potential problems, implementation
features, and the pathway for acceptance of electronic money.
Digital signature technology is the foundation for electronic money. Digital signatures are an implementation of public key cryptography and hashing functions that enable the creation of tamper proof documents, which unequivocably associate their owners. As long as there are registries that store the public keys, everyone can authenticate a digitally signed document. Digital signatures play a chief role in electronic money because they allow people to create guaranteed IOU notes.
The following is an example of the basic steps used in the creation, trade, and deposit of electronic money. In this, and all subsequent examples, Bob will be the customer and Alice will be the vendor. Bob contacts his bank and asks for $X to be transferred from his savings account to digital money. The bank digitally signs a message that contains the text $X, a serial number, and a bank code. Bob sends this message to Alice. Alice verifies the bank's signature, accepts the money, and sends Bob his products. At her convenience, Alice signs the money, sends this new message to the bank, which records the serial number and credits Alice's account[Chaum].
Speed and privacy are two sets of tradeoffs one must weigh when implementing electronic money. The first tradeoff one can make with digital money occurs during the verification process. Sometimes people need to make a tradeoff between speed and security. On-line electronic money is very secure, but take time to process. Off-line electronic money is less secure, which effectively allows for forgery, but takes virtually n! o time to process[F inney]. The second set of tradeoffs pits privacy against logging. Implementations of electronic money allow the logging of transactions in a manner similar to that of credit cards. Electronic money can also act like cash, hence the name digital cash or blind money. This implementation makes transaction logging impossible[Sullivan]. Current implementations can mix and match features from the two sets of tradeoffs.
Since electronic money is digital, it can be copied without defects. This means there must be safeguards to ensure that people don't transfer the same money to multiple people, double-spending.
The first way to do this, on-line electronic money, is to check the validity of the buyer's money up front. In addition to checking the bank's signature, Alice would need to submit the serial number to ensure that the money had not been spent. If not, the transaction continues as usual. Otherwise Alice knows that Bob double spent, and she would terminate the sale[Finney]. The drawback from on-line electronic money is speed. The bank must verify every monetary transaction regardless of size. Depending on factors like net-lag and the size of the queue at the bank, this could be unacceptable for some applications.
The second way is by using off-line electronic money. In this process, after verifying the bank's signature, Alice accepts the electronic money. Regardless of whether or not the note had been spent, when Alice returns the note to her bank, it will credit her account. The bank will then sue Bob for damages[Finney]. This system sacrifices security for speed. To minimize monetary loss through double spending, many proposals suggest limiting off-line electronic money to purchases of small value. Of course, the banks must still catch Bob, and while double spending on small purchases is not allot of money, 2^1024th spending is.
The second set of tradeoffs compares blind digital money with t! raceable money. With traceable electronic money, when Bob initially purchases the money, another field is encoded in the document. This contains publicly accessible information that permanently associates Bob with the money. As with credit cards, people will be able to make databases that associate people with the goods they buy.
With digital cash, when Bob asks the bank to give him notes, he can do
so in a manner that the bank will not be able to track the serial
number. Now when Alice deposits Bob's money, the bank will not be
able to generate full transaction records. When used with off-line
electronic money systems, a new set of procedures is added to this
basic scheme. With this, when the original note is being encoded,
extra verification data is left with the bank and on the document.
After Bob presents his money to Alice, Alice issues a set of random
challenges to Bob. Alice sends both her inquiries and Bob's responses
to the bank. As long as Bob does not double spend his money, his
identity ! remains secret. Otherwise, the bank will have enough
information to piece together the identification of Bob[Finney]. There main drawback to blind digital money
is the processing power required to issue and verify the challenges.
Although there have been no announcements, I am sure the government
will object to blind electronic money since it allows money laundering
As with any technology, nothing is perfect. The integrity of electronic money systems can be damaged with attacks on double-spending mechanisms, electronic money can be intercepted and stolen, users can make mistakes, and viruses can attack systems. If electronic money is to become the standard form of monetary transactions on the Internet, these problems must be addressed.
The present system for off-line electronic money safeguards against double spending by tracking serial numbers. If the money is spent multiple times, the bank will have enough information to find! the identify of the original owner. In most cases this will suffice, but imagine the following scenario. Suppose Bob gives Alice some electronic money. Following normal procedures, Alice verifies the bank's signature, then gives Bob his product. Instead of transferring the electronic money to her account, Alice decides to spend the electronic money with Bob's serial number. Eventually this act will attract attention. The bank will learn Bob's identity, then incorrectly sue Bob. The only fix to this problem is to prevent Alice from doing anything other than storing or depositing Bob's money. There are no technical means to do this in software.
Although on-line electronic money eliminates the problem of double spending, it does not prevent money from being stolen. Suppose Frank is using a network sniffer while Bob sends his electronic money to Alice. Milliseconds before Alice sends the serial number to the bank, Frank sends his copy. He gets an O.K. and deposits the money to his account. When Alice receives a response saying Bob's money is invalid she breaks the transaction with Bob. This leaves Bob with spent electronic money, and no merchandise. I see two techniques to fix this problem. The first, is to log information on who deposited the money along with serial numbers, and to have electronic receipts. This way, when problems ensue, there will be records which that Frank to the crime. Of course, this leads to more record keeping and processing time. The other method is to use secure communication lines, or use Diffie-Hellman key exchange and encrypt the transactions. The problems here are export regulations, and processing time.
People are stupid and accidents will happen. When electronic money resides on personal computers, mishaps could result in financial disaster. Files are deleted, hard drives are accidentally formatted, people forget pass phrases, and the occasional natural disaster occurs. Are these valid reasons to avoid electronic money? No. Equivalent events happen with physical cash every day.
Unfortunately, there are more serious threats to electronic money than
mishaps caused by its owner. These are viruses and trojan horses.
Today, creators of these programs do so for "fun and games," but when
money comes into the picture, one can easily imagine a new set of
hackers who will create programs to steal or damage electronic money[Thomas]. Although current electronic money schemes
are provided by third party companies, it is easy to see a migration
of these features into operating systems[Sullivan].
Only then, will firewalls be added such that applications using
electronic money libraries cannot corrupt the data, and network
firewalls so that electronic transactions cannot occur without users
An idea without an implementation is no good. David Chaum, the
chairman of DigiCash and "father of digital money," thought the same.
Today, his Dutch based company provides the most complete!
implementation of electronic mone y. Currently, client software for
DigiCash that operates under MS Windows, Macintosh, and UNIX is being
tested. DigiCash has involved banks for withdrawal and the deposit of
money. They also offer help on how to setup shops that accept
DigiCash. Right now, DigiCash implements blind, on-line electronic
money[DigiCash]. Right now, DigiCash has a huge
head start over Visa[Visa1], and a slew of others
who are entering the scene. Infrastructure that can be used to
support electronic money is also growing. Proposals for secure html
and secure newsgroups[Ringuette] are being
evaluated, and public key servers are popping up on the Internet.
Even though a technology may be up for the task, the public still may reject it. Electronic money is no exception. Companies may create implementations of electronic money that is secure, fast, and easy to use, but fail if they aren't accepted by the public. In order for electronic money to flourish, not only must there be a good implementation, but compatibility, international standards, and insurance must be present too.
Today, DigiCash is the only major company that has a working implementation of digital money, but soon there will be others. Visa and Microsoft[Visa2], NexusBucks, GhostMarks, and even the United States Postal Service are readying new products[Sullivan]. These competing systems will give rise to compatibility problems. Will there be an easy system to exchange money types? Or will there be advertising campaigns similar to "but they don't take American Express?" My feelings are that, as with ATM cards, there will be multiple standards. Later, as the designers of the money systems change their proprietary applications into libraries, operating system will transparently resolve incompatibilities.
Other than to provide protection for its users, security exists so that the designers! of a system will be less accountable if problems occur. Certified systems provide better defense than unproved systems if their designers go to court [Anderson]. To bring confidence to its customers, electronic money code should be submitted for review. Furthermore, legitimate systems should have large insurance policies, preferably issued by the US Government. Insurance will give customers confidence to invest their money with this new technology. If a particular electronic money scheme fails, or the organization that issues it goes belly-up, the holder of the electronic money will be safe.
The foundation for electronic money is digital signature technology. To implement these signatures, one must employ forms of cryptography. Due to ITAR, most methods for cryptography require strict permits for exportation[Will]. These export regulations will prohibit to use of some electronic money schemes to the United States. In turn, efforts to create international sta! ndards will suffer. The combinati on of ITAR and the lack of international standards could cause problems for international commerce on the Internet.
The future is bright for commerce on the Internet. Initially clients and vendors will transact by sending encrypted credit card information over the Internet. When electronic money technology matures, new opportunities will rise. People will be able to buy and sell information, small companies and entrepreneurs will be able to compete against large companies, and copy protection schemes will mature. Things will start slowly, as pioneering companies like DigiCash test their products. When flaws with double spending are solved, thievery via packet sniffing is stopped, compatibility rises, standards emerge, and products receive insurance, electronic money will take off.