Wednesday, August 21, 1995* - Volume 01, Issue 03
"God is in the details"
OUR SPONSOR: National Semiconductor and ACMA Computers
UPDATE INFORMATION
CONTACT INFORMATION
Rock bottom basics
If you want something to stay a secret, don't tell anyone, don't write it down. If you do have to send it to someone else, hide it in another message so that only the right recipient will understand. Many creative methods of hiding messages have been invented over the ages,. Cryptography in the computer age typically involves the translation of the original message into a new and incomprehensible one by a mathematical algorithm using a specific "key". Just like you need to open the lock on a door with a key, the algorithm and key protects your the contents of your message from unauthorized access.
Taking the lock and key model one step further, ease of getting into the crytographic system depends on the strength of the lock (the algorithm), and the possession of the key. For example, you may encrypt your message by changing every letter to the one n letters later in the alphabet, i.e., A is changed to B, B to C , and so on, for n=1. (This method is known as a substitution cipher and is reputed to have been used by Julius Caesar.) Knowing your algorithm, it's quite easy to reveal your true message by trying each of the 25 possible values of the key n.
If we are simply using a trial-and-error approach to discover the key, the difficulty of breaking the algorithm increases with the number of possible keys, and with the amount of time it takes to try each key. However, there may be other approaches that are less time-consuming. In substitution ciphers, knowlege of the usage frequency of letters in the language and in the message will indicate the probable substitutions. As every English-speaking parent of a child with a "Speak-and-Spell" toy knows, the "e" key usually wears out first.
The strength of the algorithm should be independent of the knowledge of how it works. Cryptographers also prefer to trust published algorithms over proprietary ones because this allows the community of experts to look for systematic weaknesses. A number of algorithms that appear impregnable at first have later yielded to specific techniques of attack. So the longer an algorithm has remained unbroken, the stronger it is likely to be.
The other part of the equation is the key itself. The longer a key is in use, the more likely someone will be able to discover it, and the larger the amount of information that would be compromised. Changing the key frequently, e.g., on a per message basis, therefore can strengthen the security of the system.
Breaking the code
Cryptographic algorithms come and go, and here are some of the latest.
One of the most widely used encryption algorithms is DES (Data Encryption Standard). It was developed by IBM and adopted as an official Federal Information Processing Standard (FIPS PUB 46) in 1976. DES uses a 56-bit key - considered to be just beyond the technology at the time to break with a brute force approach. However, by 1993 a custom computer costing $1 million could theoretically be built to crack DES by brute force in 7 hours. In the following year, a new technique called "linear cryptanalysis" enabled another researcher to break DES in 50 days on a single workstation, or about 14 MIPS-years assuming a 100 MIPS machine. Efforts to extend the life of DES include lengthening the key to 128 bits (still beyond the realm of computer horsepower today) and "Triple DES", which involves three passes of encryption and decryption using two different keys.
The RSA algorithm, used in public key cryptography and patented in the US, is based on the fact that it's easy to multiply two large prime numbers together, but hard to factor them out of the product. A challenge issued in 1977 to factor the key based on a 129 digit number (or 429 bits) was finally met in 1994 by an international group of over 600 volunteers and 1500 computers working for 8 months. The estimated effort was somewhere between 100,000 amd 1,000,000 MIPS-years. The decrypted message read, "THE WORDS ARE SQUEAMISH OSSIFRAGE", starting the tradition of using those words in key-breaking challenges.
IDEA, International Data Encryption Algorithm, is a new DES-like algorithm invented in 1991 in Switzerland. It uses a 128-bit key and was designed with more mathematical theory than DES. It is believed to be a strong algorithm, but no challenge has yet been mounted, and only time will tell whether it will withstand the attacks of cryptographers. IDEA is patented internationally.
RC2 qnd RC4 are two proprietary, DES-like algorithms that allow variable key lengths. They were developed by RSA Data Security, Inc, and are notable for having an exportable version with a limited, 40-bit key size. The 40-bit RC4 algorithm is used in the International versions of many software packages including the Netscape web browser. In the last week, a French programmer successfully broke a 40-bit key and decrypted a single secure Netscape transaction (created as a challenge using data from a hypothetical Mr. Cosmic Kumquat at 1234 Squeamish Ossifrage Road). Netscape estimates that the effort took 64 MIPS-years.
Where do keys come from?
For two people to communicate in encrypted mode, they would have to arrange the secret key before hand. Setting up keys presents all sorts of logistical problems when you are communicating with multiple people (a separate key may be needed for each), or when you cannot communicate the key information securely in the first place. Some systems get around this by establishing a centralized key authority, e.g., the US Government, or the key database in a Kerberos system. However, the invention of public key cryptography changed all that. Unlike private key or symmetric encryption where a single key is used for both encryption and decryption, public key (or asymmetric) encryption uses a public key-private key pair. The sender encrypts the message using the published public key, and the recipient reads the message by decrypting it with the private key. Public key directories could be published, just like telephone and other directories. However, the use of public key directories presents a different problem. Since the key is not prearranged through a trusted source, how can you be sure the public key belongs to a real and legitimate person or organization? All you might have is a disembodied e-mail address or URL.
There are currently two solutions to the certification issue - building what is commonly termed a "web of trust", and setting up official key issuing authorities. The web of trust works through personal recommendations. Your public key is "signed" by other people who can vouch for you. If I trust one or more of those people, I would believe that you are real and legitimate. The alternative is institutional, through a Certification Authority (CA). As states issue drivers licences or universities issue ID cards, public keys can be "issued" by an organization that is familiar with the individual involved, or Certification Authorities. These CA's in turn derive their "trust" from thecentral Internet Registration Policy Authority to form a hierachical organization.
Sometimes signed but never sealed
E-mail is arguably the most prevalent and important component of Internet traffic. We tend to think of mail mostly in terms of letters in sealed envelops. But mail messages being routed from computer system to computer system on their journey through the Internet are more like postcards than letters. Along the way, a message can be read, and even more, it can be modified. E-mail headers are notoriously easy to forge. Encrypting the message is one technique to prevent tampering. In addition, cryptographic techniques can also be used to securely "sign" a mail message without encrypting the message..
In signing a document, a one way hash function is used to reduce the message to a short, fixed length string of characters (a digest). The digest is then encrypted with the sender's private key to create a digital signature. To verify the signature, the recipient takes the message and converts it into digest form, and separately decrypts the sender's signature with the sender's public key. The results of these two operations should be identical if the message has not been tampered with. In addition, use of the sender's public key to decrypt the signature also confirms that the message actually came from the sender.
Common one way hash algorithms include MD2, MD4, and MD5 from RSA Data Systems, and the Secure Hash Algorithm (SHA) invented by the NSA as part of the US government Digital Signature Standard (DSS).
Sealed with a cipher
Public key algorithms such as RSA are used in mail encryption programs. However, because of the resources required to encrypt messages with RSA, only a randomly-generated session key is usually encrypted with RSA. This session key is used to encrypt the message with a private key algorithm such as DES. This approach provides the added advantage that a different key is used to encrypt each message that is sent.
PGP (Pretty Good Privacy) is the best-known secure mail program available today. It was written by Philip Zimmermann and first released in 1991. It uses RSA public key cryptography for session key exchange, and IDEA for message encryption. Because of US patents surrounding the RSA algorithm and US export regulations, PGP has had a colourful history. The result is different versions, not always compatible with each other, available for use within US/Canada and for the rest of the world.
PEM, the Privacy-Enhanced Mail standard, is developed as an Internet mail standard by many of the people who helped create PGP. The standard is described in RFC 1421-1423, and also includes the X.509 Key Certification Standard from CCITT. Originally, it only allowed a hierarchical key certification organization, however this has been modified to include the PGP-styled "web of trust" model. Reference implementations include RIPEM, Riordan's Internet Privacy Enhanced Mail developed by Mark Riordan, and a version from Trusted Information Systems, Inc. RIPEM currently does not support key certification.
The main difference between PGP and PEM appear to be in the area of sender privacy. PGP allows encryption without signing, i.e., anonymous, encrypted messages, and PEM does not. Additionally, the PEM message signature is outside the encryption envelop, allowing anyone to verify the sender of the message without reading the message.
Recently, RSA Data Security and a number of key messaging vendors have announced support for a new standard for secure MIME messaging (S/MIME).
Or is that the other way around?
There is usually a howl of laughter or outrage when cryptographers or privacy advocates mention that cryptography is considered a "munition" by the 1954 Munitions Control Act - just like nuclear missiles . What is often forgotten is that pivotal battles during World War II, such as the campaign against Rommel in North Africa, or the Battle of Midway, relied upon intelligence gleaned by breaking German and Japanese codes. In effect, cryptography played as important or perhaps even more important role compared with such plebian materiel as cannons or bombs. Although time has passed, export of cryptography is strictly regulatd.
But at the same time, the Internet has no natural political boundaries. PGP, placed on the Net by an enthusiastic user, quickly spread to the rest of the world. The same happened to RSA Data Security's RSAREF reference code library. The result is a federal criminal investigation of Philip Zimmermann. As the Internet community comes to grips with dealing with export control, World Wide Web and FTP sites are creating restrictions and posting advisories to discourage downloading non-Americans. More dramatic approaches include an effort to excise DES encryption out of the Kerberos network security system. This bare bones system would then be exportable and an encryption engine can be reinstalled abroad.
The software industry is acutely aware of the cost of export control, and is active in lobbying for its reduction or elimination. The recent quick breaking of the 40-bit key version of RC4 used for exportable versions of software has simply served to emphasize the limitations of existing export controls.
Interestingly enough, although computer code cannot be legally exported, source code in printed form within a book does not seem to face the same restrictions. Books containing sample DES implementations are widely available. Schneier's "Applied Cryptography" book can be exported, but not the diskette containing the same source code as is listed in the book. One might wonder whether this is a face-saving compromise along the lines of "Don't Ask, Don't Tell".
Of microdots, the USS Pueblo, and steganography
An encrypted message looks like gibberish, and alert people that there is something to hide. But what if the message is totally innocuous looking? This is an old trick that started centuries ago with writing in ink that is invisible until the paper has been heated. The microdot, a piece of film containing a very highly reduced image of the secret message and embedded in the punctuation marks of a normal document, was invented during World War II. In the 60's, a photograph of members of the USS Pueblo released by their captors to show their cooperation seemed ordinary enough until you looked carefully at the hanf positions. They were spelling out the word "SNOWJOB" in sign language. And we all remember the hysteria about finding satanic messages by playing certain recordings backwards.
What if you used the least significant bit of each pixel in a bitmap image to encode a message? The impact on the appearance of the image would not be noticeable. This is known as steganography, or covered writing. A 480 pixel wide by 100 pixel high image - smaller than many WWW home page banners, could theoretically contain a message of more than 5,000 characters. The ecoding is quite easy with a computer - and no complicated mathematics at all. And of course the same principles apply to audio and video files as well. Furthermore, the image can be used simply as a carrier, with the message being first encrypted. So an old trick has found a new, digital dog.
Applied cryptography for a little fun and mostly profit
While mail privacy is an important topic to many, cryptography also contributes to an area of far greater monetary value: online commerce. Public key cryptography enables digital signatures that verify the authenticity of a document, confirms the authorization of the sender, and non-repudiation of the transaction - all critical components to financial transactions. It contributes in other areas as well.
Timing is important to many transactions, e.g., tax returns must be filed by midnight, April 15, in the US, There needs to be a secure mechanism to timestamp the corresponding electronic documents. Surety Technologies provides a Digital Notary System that produces electronic timestamping. It is based on combining message digests (created as part of the digital signature process) and timestamp values from a tamper-proof clock to create "super" messages that can be uniquely identified. The contents of the original document need not be revealed to the timestamp service.
Current trends in electronic payment systems is firmly towards the development of a stored-value card. Money is "downloaded" and stored in an integrated circuit on the card. This can be for used for any purchase where cash can be used. Up to this point, cash has remained the last anonymous, untraceable method of payment. With the capabilities of a "smart" card such as the stored value card, all "cash" transaction information would also be captured, creating an associated loss of privacy. However, we need not lose our privacy to gain the use of electronic cash. Dr. David Chaum of Digicash has created algorithms for "blind signatures" that allow anonymous use of cash through a smart card mechanism. Whether this will be adopted by major card organizations, though, remains an open question.
Just sign me "a private netizen".
Digital signatures can help confirm that the sender of a message is really who he says he is, but what happens when he doesn't want someone to know? Anonymous letters - graffiti, ransom notes, whistle blowing, etc. have all been around for a long time and often serve a genuine purpose. Yet e-mail systems generally persist in recording and reporting information about the message originator. To preserve the anonymity of the sender, "anonymous remailers" have been set up on the Internet. Messages sent to these remailers for forwarding have the originating header stripped off and replaced with the address of the remailer. Replies to the remailer are then forwarded back to the originator without revealing his identity. While, the best known of anonymous remailers is Although most people think of anon.penet.fi, run by Penetic, a company in Finland, there are actually about 20 remailers on the Internet
In using an anonymous remailer, one must always remember that the remailer is but one part of the system, and that it too is subject to attacks. First of all, your mail may go through a number of hops before reaching the remailer. Mail can be monitored en route, and especially at the point of origin. Even if your message is encrypted, the knowlege that you are sending mail to an anonymous remailer may be cause for suspicion. Traffic analysis of the remailer - tracking incoming and outgoing mail and their sizes, can review the identity of seemingly anonymous messages. There is a variety of techniques to improve the certainty of anonymity and to defend remailers from attacks; e.g., by chaining - sending a message through a number of remailers to the final destination, or by holding mail a random amount of time to deter traffic analysis. However, not all attacks are technical. In February, the Church of Scientology was able to obtain the identity of one user of anon.penet.fi by a criminal complaint throu gh Interpol and the Finnish police authorities. Protecting privacy is a step removed from shielding criminal activity.
The gusher from the data strata
During the 1972 US Presidential Elections, the McGovern campaign found that one of its best sources of campaign contributions came from a list of Volvo owners. The discovery was serendipitous, but in the intervening years, computer technology has grown to the point of playing a vital role in identifying these kinds of trends and patterns. And increasing computerization means that far more information is being kept in electronic databases amenabale to efficient searches and correlations. Database marketing, or database drilling, is now a standard part of the marketer's repertoire. It has plenty of data to work on, from credit card transactions to subscriptions and mailing lists, and even the government gets in the act by selling information such as name, address, and driver's license number.
At the same time, the ability to link information in different databases - education and driving records, credit history, health, spending patterns, etc. - is no further away than a unique ID number. Most countries have some form of ID system for its citizens; the US has the social security number (SSN). Although the SSN was explicitly intended not to be a national ID number, practically it is used by many organizations for idenification purposes. The result of even partial linking of records would be an incredibly detailed history of our lives - after all, it's a mere matter of programming, and not complicated programming at that.
Cipper sightings on the Internet sea
Despite the mountains of data that accumulate about our highly computerised lives, the ability to (get a court order) and listen in to the conversations of Americans has always been a cherished tool by law enforcement agencies. The rise of digital telephony and telecommunications and cryptography threatened to put an end to that. The government's response was to propose the Capstone program. The key piece was the Clipper chip. Incorporated into all telephones and other such devices, this would permit encrypted communications through the use of the "Skipjack" algorithm. Except the device is built with a "Law Enforcement Access Field" (LEAF). Given a court order, the law enforcement agency could go to key escroll centers with the field information and obtain the keys necessary to decrypt the conversation.
The outcry from this announcement was loud and long. The Skipjack algorithm. classified and implemented in hardware, did not allow the cryptographic community to examine and test its strength. Worse, many were afraid that the algorithm contained a trapdoor that would allow decryption without the escrolled keys. Others thought that Clipper marked the beginnings of an attempt to outlaw cryptography. The irony was that upon closer examination, Dr. Matt Blaze of Bell Labs discovered a way to create a bogus LEAF and thus make decryption by the authorities impossible.
With that, the Clipper chip subsided into oblivion. However, in the last week, the Clinton Administration has indicated that it will soon propose a "son of Clipper" - and hinted at allowing non-government escroll centers. In the interval, the programming community had not been idle. Nautilus, an encrypted phone system running on PC's and communicating through the Internet was introduced in the spring. (The name Nautilus comes from the submarine that sank clipper ships in the Jules Verne classic "20,000 Leagues Under the Sea".) Philip Zimmermann has just announced PGPFone, a similar system running on Macintoshes.
It's eleven o'clock at night. Do you know what your software is doing?
First came the programs that helped us to write letters and spreadsheets on our personal computers. Then came the insidious little viruses - hiding inside other programs that did everything from wishing us "Have a nice day" to wiping out our hard drives. Well, whatever they did, they stayed within the confines of our computers. Then we got online services - our PC's started to communicate with the service mainframe at the other end of the phone line in a client-server fashion. And the client sometimes served the server...
As much as the client software can help you send mail and download interesting information from the server, it is just another program that has control over your computer and its resources. There are the usual urban legends about client software from major online services that rearranged the PC's file system - and some of this naturally happens when you download and install a software upgrade. What is less amusing is the report of a Registration Wizard within beta versions of Microsoft's Windows 95 system. This program gathers information about all the software that is being run on all the PC's in your network and then reports it to Microsoft when the you signs up for Microsoft Network Services.
Performance monitoring has been around for many companies, particularly in service organizations, e.g., the speed of the cashiers at busy supermarkets, or of directory assistance operators. As we depend more and more on personal computers and get increasingly interconnected, the potential for information gathering and reporting to outsiders rises. The early adopters are likely those online services who provide the communication link, since they can directly upload the information gathered. But what prevents another program from sending information via e-mail to its originators if the communication is possible? Visionaries speak of the much heralded "intelligent agent" technologies from the likes of General Magic. These would wander the net searching for information, purchasing hardware and software, making friends and influencing other agents. Given the above discussion, the possibilities for mischief - if unchecked - is mind-boggling.
The educated person knows where to find the information
The issues of privacy and cryptography are of vital interest to many of the denizens of the Internet. As a result, there are some excellent sites with extensive lists of resources available.
Curl up with a good book
While the online world is rich with resources, print is still a wonderful packaging technology, especially for larger documents. For those times when the screen has lost its phosphorescent allure, here is our selection of books and other printed materials for your consideration.
CRYPTOGRAPHY
PRIVACY
This issue of Netsurfer Focus is sponsored by National Semiconductor and ACMA Computers.
For more information about our sponsor and other advertisers,
please see the current issue of Netsurfer Focus Marketplace.
"http://www.netsurf.com/nsf/v01/03/nsfm.01.03.html"
Participation in the Focus Marketplace is unrelated to editorial coverage within Netsurfer Focus.
Last Updated 16 September, 1995.
Netsurfer Focus is currently a periodic supplement to Netsurfer Digest and Netsurfer Tools.
Netsurfer Focus Home Page: http://www.netsurf.com/nsf/index.html
Letters to the Editor: focus@netsurf.com
We appreciate hearing from you even if we do not manage to respond
to every message that is sent to us. Please note that we reserve the
right to publish letters to the editor in future issues of Netsurfer
Focus or on our website.
Back Issues: If you would like to obtain copies of back issues or resource files (in HTML format only) via e-mail, send mail to info-focus@netsurf.com with "send crypto-index" in the body of the message.
To subscribe to Netsurfer Digest or Netsurfer Tools:
By WWW form: http://www.netsurf.com/nsd/subscribe.html
By e-mail: nsdigest-request@netsurf.com
Body:
subscribe nsdigest-text
subscribe nsdigest-html
Netsurfer Focus
Publisher: S. M. Lieu
Production Manager: Bill Woodcock
Sales: John Pendleton
Netsurfer Communications, Inc.,
President: Arthur Bebak
Vice President: S. M. Lieu
NETSURFER FOCUS (c) S. M. Lieu. All rights reserved.
NETSURFER DIGEST is a trademark of Netsurfer Communications,
Inc. Other publication, product, and company names may be trademarks
of their companies.
"God is in the Details" is a quote from Mies van der Rohe.