Interest in encryption arises from a history of technology transforming
commerce. The need for secure communications has propagated the development
of encryption schemes to protect the integrity of information as it is transferred
electronically. There are a number of security services involved in encryption
outside of the issue privacy: confidentiality--assurance that the parties
involved in a communication have exclusive knowledge of the message; authentication--assurance
that the parties involved are who they say they are; non-repudiation--proof
of which party sent the message or which party received it; and time-date
stamping--proof of when the message was created, sent and received. The
encryption scheme itself can be classified as single-key or public-key based
on its implementation.
The idea behind single key encryption is to encrypt a block
of data using an encryption algorithm and a number called the "key,"
which is known only to the sender and the recipient. The key controls how
the message is encrypted and every key does it differently. Thus, even it
is known that a message has been encrypted using a specific algorithm such
as DES, the Data Encryption Standard algorithm invented by IBM, it is impossible
to decrypt the message without knowing the key as well.
An example of single-key encryption is a password on a word processing document
or a PIN number on an ATM account. Single-key encryption is also called
symmetric encryption, since the same key is used to encrypt a message from
plaintext into ciphertext and to decrypt the ciphertext back into plaintext
again. The key itself is a randomly-generated stream of bits of a specified
No matter how secure the encryption algorithm used to generate the key to
encode the information, the single-key encryption system has two inherent
security weaknesses. First, the sender and the recipient of the message
need to share knowledge of the same secret key--meaning that each must trust
the other not to compromise knowledge that is exclusive to them. This also
makes anonymity almost impossible. Second, the communications channels used
to exchange keys must be independently secured to prevent fraud.
Public-key encryption, also known as asymmetric encryption,
solves both shared key risk and the distribution key problem. Instead of
a single key, the public-key encryption system uses a pair of keys which
are mathematically linked to each other--one of which is kept secret and
the other which is transmitted. Each user possesses a private and public
key pair which cannot practically be derived from each other.
With public-key encryption, only one key is needed to encrypt a message;
the other is used for decrypting the message. A sender with a recipient's
public-key can thus encrypt a message which only the recipient can decrypt
using his private key. Likewise, a sender can encrypt a message with his
private key which can be decrypted by anyone with the matching public key.
In this way, a recipient with the sender's public key can verify that it
was indeed the sender who originated the message. This is called a digital
signature. Public-key encryption thus provides for both confidentiality
The benefits of public-key encryption are clear. The sender and recipient
no longer need to communicate previously, nor do they need to exchange private
keys to send a communication that is signed and secure. They only need to
exchange public keys, which can be done over open communication lines. Public-key
encryption thus offers the possibility of a much more cost-efficient technology
for securing communications.
The idea behind public-key crytography was invented by Whitfield
Diffie, a mathematician and computer scientist from MIT, in 1975. His big
realization was understanding how to build a cryptographic system in which
each communication was controlled by two keys instead of one. In addition,
if the two keys were related, anything encrypted by one could then be decrypted
by the other. However, Diffie did not solve the problem of designing mathematically-related
keys that could not be derived from each other until he got together with
Stanford Professor Martin Hellman. From 1975 to 1978, Diffie and Hellman
co-authored a series of now-classic papers on public-key cryptography.
Several public-key encryption techniques now exist, subject to patents.
Among them is the method pioneered by Massachusetts Institute of Technology
professors Ronald L. Rivest, Adi Shamir and Leonard Adleman, known as RSA
encryption. The mathematics behind RSA's asymmetric encryption is based
on the notion of one-way functions. For instance, the product of two prime
numbers, 31 and 97, is simple to calculate (=3007), but difficult to factor
(=31 and 97). Likewise, raising numbers to powers is much simpler than taking
roots; it takes a few seconds to find 2^12 (=4096), but longer to calculate
the twelfth root of 4096 (=2).
RSA, which to date is the most successful public-key cryptographic system,
combines the fact that raising numbers to powers is a one-way function relative
to extracting roots and multiplying prime numbers is a one-way function
relative to factoring. With the latter, if the product is large enough and
no one but one person knows its factors, it constitutes a trapdoor that
allows only that person to decrypt messages. For example, using cryptographic
equipment, one can generate two large primes, each 300 digits long, whose
product is 600 digits. This number is huge; the number of particles in the
universe, for instance, is estimated to be less than 100 digits long. The
600-digit product is then made public, unfactorable by anyone except the
originator. In fact, the largest number of this type that has been factored--129-digits--was
initially proposed in the '70s and just solved last year.
The theoretical limits of public-key encryption rest with the
need to maintain the secrecy of the private key. Before transmitting the
public key, a sender must first generate his own public-private key pair
on a computer or other medium that can be kept secure from discovery. Thereafter,
the sender only needs to publish his public key to enable others to send
encrypted messages that only he can read.
Before using the key, however, any recipient needs assurance that the key
is indeed the sender's public key and has not been erroneously generated,
fraudulently generated or changed. To accomplish this, clients of public-key
encryption support a class of trusted third parties, called certification
authorities. These authorities certify the authenticity of a public key
as a match to the private key held by the sender. They are the mechanism
by which the identity of the parties involved in a communication is verified.
Certification authorities are divided into several levels, each certifying
a public key with a fixed level of assurance, resulting in a precise degree
of liability for error. At the highest level, the authority assumes all
risks associated with verifying the identity of both parties while certifying
the validity and enforceability of the transaction. Along these lines, a
cryptosystem can only be trusted as much the certification provided by the
Certification authorities may charge a fee for their services to cover their
costs and risks. The system actually requires trust rather than creating
it, however. For instance, suppose the sender meets with an authority or
else indirectly obtains all the information needed to issue a key-pair certificate.
The certificate consists of sender's public key encrypted with the authority's
own private key and thus, securely signed. The certificate is then stored
on the sender's own computer system.
To prove his identity to a recipient, the sender encrypts a message with
his private key as a digital signature and then appends his public key without
encryption along with the certificate. The recipient can then use the authority's
public key to verify that the appended public key is the same as the one
encrypted by the authority's private key. Either public key can then be
used to decrypt the original message.
Note that the requirement that the authority and the recipient must meet
or have a trusted exchange obviates one of the principal advantages of public-key
encryption--anonymity. In every encrypted transaction, the value places
on anonymity must be balanced relative to the level of trust required. In
addition, certification is only as trustworthy as the information given
to the authorities. Errors or false generation of authority keys also create
a trust problem. The solution to this is a hierarchical approach to key
management, which an authority's key is certified by a higher level and
Certification hierarchies tend to work well in structured organizations
such as corporate networks. But for this scheme to be used in communications
among different organizations, they must agree on a certification authority
that can be trusted by all. This is not necessarily easy--for example, corporate
users may not trust certificates signed by the certification authority of
a competing company. The alternative to certification hierarchies is a web
of trust where users decide for themselves which keys are valid. In other
words, users will trust keys which come from a trusted source and arrive
via a reasonably secure route--for instance, a worker may trust a key sent
by another co-worker arriving via a secure private network vs. a key from
purporting to be from an employee at another company arriving via the Internet.
The certification hierarchy model can be thought of as a continuous client-server
approach while the web of trust is a more distributed peer-to-peer approach.
The big advantage of the certification hierarchy model is that it is fairly
easy to enforce a consistent policy within a large organization. On the
downside, organizations that want to control who gets authenticated will
need to set up and maintain their own certification authority, a process
that requires specialized administration. The web of trust model, on the
other hand, tends to offload many administrative headaches to end-users.
Experienced users are likely to feel more in control since they are in control,
but the key management process can also be quite a challenge.
There is an economic risk that transaction costs will rise as
other types of assurance are required to provide new types of service such
as secure key exchange, secure time-stamping and archiving, key escrow and
secure networks for third-party certification as to a signer's level of
corporate signing authority. In addition, the system makes the need for
higher certification, with attendant security risks, greater in proportion
to the degree of security required.
Encryption scheme vendors are constantly looking for areas of innovation
for securing transaction systems. Such an area involves technological methods
for ensuring trust. Another is the creation of security measures using cryptosystems
in which security bears a relationship to the importance of the information
being handled and the anonymity desired by the parties involved. Given the
commercial interest in electronic commerce and the use of the Internet as
a mechanism for disseminating information, these problems should attract
a great deal of intellectual attention in the years to come.
Further systems are also needed to ensure that unwanted messages encrypted
with the recipient's one public key are managed, perhaps replacing the exchange
of encrypted messages with encrypted keys. If these fall into place, then
a reliable system for secure message transmission and private communication
can be made possible, and with a greater level of anonymity than is currently
practiced on most on-line systems. From this, one may conclude that encryption
techniques may beef up communal policies for managing communications and
identifying abuses in the system.
Encryption development could affect decisions concerning the liability of
on-line services or bulletin board systems (BBS) for disseminating unauthorized
or unlawful material such as pirated software or pornographic material.
For instance, a BBS that supports messages privately encrypted by its users
creates a shield against liability for itself since the BBS operator can
no longer exercise editorial control. Part of the problem in the past is
the uncertainty of local jurisdiction and federal authority over on-line
communications. The problem might be alleviated if all transmissions were
encoded so that content were truly personalized. In addition to encoding
a message in a form that enables its transmission, encryption also designates
the assent of the viewer and ensures that only the holder of an authorized
public key can read the information. In this way, private communication
would be managed by private control rather than federal law.
Standardization and increased availability of encryption may also raise
the standard for protecting proprietary information in the age of perfect
digital reproduction. The heavy demands placed on copyright law to prevent
unauthorized duplications may be eased dramatically if encryption is combined
with on-line publication. The reason is simple: the decryption of a message
encrypted with another person's key requires a computational act, which
would be a significant factor in establishing criminal intent. Many states
have already enacted computer crime laws to complement federal statutes
that make unauthorized interception of electronic communications illegal.
An unauthorized act of decryption is a sign of intent to seek illegal access
to private communications and most likely will prove a point of great significance
in criminal prosecution of interception and copying on-line in years to
As encryption becomes more widely available, lawyers have begun
to provide guidance on the implications of its use, from the formation of
contracts to criminal liability. Likewise, there has been recent litigation
and government interest in legal standards for network security. Public-key
encryption by itself has already spawned commercial and legal changes, including
redefining terms such as "signature," "original" and
California itself has recently announced a "digital signature"
statute to govern on-line transactions, defining legal requirements for
a digital signature "intended by the party using it to have the same
force and effect as the use of a manual signature." 2
The legal field must be involved in providing guidance on legal standards
for encrypted electronic communications. Many secured transactions will
shape security according to laws ranging from contract formation to criminal
liability. The converse is also true, as the establishment of a standard
fro commercial public key crytography will provide a foundation for addressing
related legal standards.
Encryption is entirely legal for use inside U.S. borders. Yet encryption
technologies, long used for protecting secret military and government communications,
are legally regarded as munitions-- the same classification as a truck full
of artillery shells--so are subject to strict export control. When PGP was
anonymously posted to the Internet in 1993, for instance, inventor Phil
Zimmermann became the target of a three-year criminal investigation that
ended only early this year. Government security specialists fear that the
widespread use of encryption could protect information involving criminal
activities such as terrorism or drug trafficking. The government this restricts
the export of all sophisticated cryptography in the interest of national
security. Industry also has a vested interest in limiting encryption. Companies
fear that disgruntled employees could encrypt corporate databases and throw
away the key or hold them hostage.
Adoption of encryption appears to be proceeding more quickly for authentication
than for privacy purposes, perhaps in part because products for authentication
are not subject to the same export restrictions as those for privacy. However,
the encryption industry is working hard to convince the federal government
of its importance to U.S. export growth. Representatives argue that the
government's policies are hampering the computing industry's global competitiveness.
A report commissioned by the NSA and administered by the Bureau of Export
Administration of the U.S. Department of Commerce determined that the U.S.
market share for encrypted products in 1994 declined significantly in Britain,
Denmark and Switzerland. In addition, sources in 14 countries told researches
that U.S. export controls have limited the availability of sophisticated
U.S.-made software for purchasing.
Currently, several industry coalitions are pressing for immediate action
by the Clinton Administration to ease restrictions. One such group, composed
of Apple Computing Corp., AT&T Corp., the Center for Democracy and Technology,
the Electronic Frontier Foundation, Lotus Development Corp., MCI Communication
Corp. and Tandem Computer, is asking the Administration for a policy that
balances the issues of security enhancement, international interoperability,
freedom for users to choose the method of encryption, marketplace acceptance,
privacy protection and respect for the legitimate needs of law enforcement.
Meanwhile, the government is moving ahead with plans to enhance its existing
Escrowed Encryption Standard (EES) to cover federal data communications.
The government also touted the development of a private-sector alternative
to EES, called the Clipper Chip, which would provide a standardized mechanism
for encryption but also a backdoor that permits the government to eavesdrop
on network and telecommunications.
Attracting consumers to the Internet is proving to be an uphill struggle,
in part because of security concerns. Potential merchants of the Internet
point out that 100 percent security cannot be guaranteed even in the real
world. The issue is not so much about commerce than it is about privacy,
however. Consumers have been especially wary of venturing into cyberspace
for fear the personal information transferred and stored electronically
will not be safe from hackers and cheats. In any case, paperless electronic
commerce will not succeed until the security of on-line open systems is
engineered to at least equal the level of security offered by traditional
kPmtApgAAAD05nOW7GK1t Auy6BGnGGwOZB2K2qVVCi - 4FEQOGRQxy TbyVGWQVT7nv5ae
That's "Hello. You look marvelous," spoken in PGP.
PGP or "Pretty Good Privacy," a program written by self-taught
cryptography expert Phil Zimmermann, is the best-known and most popular
encryption software available. In establishing PGP as an encryption standard
for the masses, Zimmermann has earned folk-hero status among the growing
cypherpunk subculture. PGP and Zimmermann's outspoken views on free speech
have in many ways come to symbolize the debate over cryptography for the
Phil Zimmermann, a product of 70s-era anti-nuke political action, has always
held a fascination with cryptography. When he first heard about the public-key
cryptography scheme developed by Diffie and Hellman, he toyed with the idea
of implementing such a scheme for the personal computer platform. Zimmermann's
philosophy about the public's right to cryptography is best related in his
own words, from the product documentation he wrote for PGP:
"You may be planning a political campaign, discussing your taxes, or
having an illicit affair. Or you may be doing something that you feel shouldn't
be illegal, but is. Whatever it is, you don't want your private electronic
mail or confidential documents read by anyone else. There's nothing wrong
with asserting your privacy. Privacy is as apple-pie as the Constitution...
If privacy is outlawed, only outlaws will have privacy. Intelligence agencies
have access to good cryptographic technology. So do the big arms and drug
traffickers... but ordinary people and grass-roots political organizations
mostly have not had access to affordable military grade public-key encryption
technology. Until now."
Zimmermann began work on the project in 1984, having never had
any formal training in cryptography. Instead of trying to come up with a
completely new encryption algorithm, he decided to leverage off RSA, a public-key
algorithm patented by RSA Data Security. The process was a slow one: Zimmermann
completed his RSA implementation in 1986, but it was 1991 before his first
version of Pretty Good Privacy was ready for release. It was a major accomplishment:
effective, yet faster than any other encryption scheme on the market.
Although he had dedicated seven years of his life and had emptied his savings
account in order to complete the project, Zimmermann made a surprising move:
he decided to give the program away. He decided against charging a fee (as
he had originally intended), because he was worried that the government
would soon outlaw encryption altogether. By allowing PGP to propagate as
widely as possible, Zimmermann hoped to ensure that the tools for privacy
were well in place by then. PGP did indeed spread like wildfire: within
hours of its release, it had been downloaded by people from all over the
Pretty Good Privacy's popularity has transformed Phil Zimmermann
into a poster boy for the "cryptography for the masses" cause.
Zimmermann's theology is passionate, interesting, and in the opinion of
many, paranoid. In support of his ideas, Zimmermann draws an analogy to
Americans' use of the US Postal Service. He poses the question, "What
if everyone believed that law-abiding citizens should use postcards for
their mail?" If this were the case, anyone sending their mail in an
envelope would be subject to suspicion that they were hiding something.
The reason using an envelope to send your mail through the postal system
doesn't attract suspicion is the fact that most people use them, whether
or not what they're sending is meant to be completely confidential. Envelopes,
then, are an accepted means of asserting a right to privacy which, according
to Zimmermann, is "as apple pie as the Constitution." Zimmermann
suggests that many people use envelopes because they consider what they
write to be nobody's business but their own, and they know that mail passes
through many hands before arriving at its destination. Envelopes make it
less feasible and more expensive for people to read your personal mail:
about the only way to read sealed letters without leaving much evidence
of the invasion is to painstakingly steam them open. Zimmermann argues that
cryptography is simply the extension of this same privacy to the growing
field of e-mail. In many ways, privacy is far less sure in e-mail than it
is in a sealed letter: it is easy to intercept e-mail, undetected, at any
of a number of points during its transmission. Once an e-mail transmission
path is compromised, very efficient snooping is possible: while a conventional
mail snooper might have to open hundreds of letters before finding one containing
valuable information, an e-mail snooper can easily create a program to scan
for and retrieve only those messages containing "key words" representative
of messages that might interest them -- like "bomb," "illegal,"
"VISA" or "confidential." This makes e-mail monitoring
feasible on a grand scale, in ways standard mail could never be. In fact,
the NSA already routinely scans international cablegrams in this way!
Zimmermann has also been a vocal opponent of US Governement attempts to
establish "key-escrow" data encryption standards like the so-called
Clipper chip. In "key-escrow" systems, a trusted third party is
effect given a copy of each chip's private key. This is done so that law-enforcement
agents can be ensured access to encrypted information, provided they can
procure if a warrant for the information. Zimmermann has a point when he
says that the next logical step, critical to making key-escrow systems effective,
would be to outlaw all other forms of cryptography. The implementation of
a key-escrow scheme for all encrypted communication would make such communications
only as secure as the third-party key holder is trustworthy... which was
the very reason why Diffie developed the public-key concept in the first
place. Electronic Frontier Foundation co-founder John Perry Barlow well-represented
the seriousness with which cypherpunks protect their private keys when he
said, "You can have my encryption algorithm... when you pry my cold
dead fingers from my private key." Diffie did not intend private keys
to be shared with anybody -- not collegues, not friends, not spouses, and
certainly not unknown, self-interested key registry employees who would
release keys at the first flash of a badge or warrant. Additionally, there
is nothing to guarantee that a key obtained from a key registry will be
used only for its intended pupose. A crooked law-enforcement agent could
make a great deal of money by making digital copies of private keys from
companies under investigation, carrying them out on a diskette, and selling
the keys to competitors eager for market advantages.
The legal issues surrounding PGP are numerous and complicated.
The RSA public key cryptosystem, which Zimmermann used as the basis for
the public-key technology in PGP, was developed at MIT, which holds a patent
on it dating from 1983. The exclusive commercial license for the sale and
sub-licensing of RSA is held by company named Public Key Partners (PKP).
Thus, when PGP was originally released, PKP cried foul, charging that Zimmermann
was blatantly violation U.S. law by using a patented, unlicensed algorithm
in his product. Zimmermann countered that the program was a "research
project," for which he was recieving no financial compensation, and
was therefore not a violation of commercial copyright laws.
In order to dodge further legal hassles, PGP 2.0 was released by an international
team of software engineers under Zimmermann's guidance. In a particularly
witty move, the new version was released by Peter Gutmann in New Zealand,
where RSA patents do not apply. Although it was released only in Europe
and New Zealand, "it spontaneously spread to the USA without help from
[Zimmermann] or the PGP development team," according to Zimmermann.
Zimmermann and his associates obviously planned the release with this result
in mind: although it is illegal to export powerful encryptioni technology,
there is no law against importing it.
The freeware version of PGP is distributed by MIT, under the terms of the
RSAREF license from RSA Data Security, Inc. (RSADSI). RSAREF is a subroutine
package from RSADSI that implements the RSA algorithm. The RSAREF license
allows for unlimited non-commercial distribution within the U.S. of programs
using the package.
In order to speed encryption and decryption, PGP only uses the
slower RSA technology for the public- and private-key components of the
program. The main bulk of the encryption is done using IDEATM, which is
conventional cipher covered by a European patent. Ascom-Tech AG, the patent-holder,
has granted permission for royalty-free, interational use of IDEA in all
non-commercial versions of PGP. Commercial and Government users must obtain
the $99 licensed version released by ViaCrypt.
Although PGP is the reigning standard among encryption programs, its existing
implementation is not very user friendly. Instead of operating in tandem
with the mail readers users are most familiar with, PGP commands are executed
by entering any of a number of command-line switches. Although there are
a number of front-end interfaces available to simplify the encryption/decryption
process, none of them hides the implementation to the point where a novice
computer user could use it confidently. PGP, therefore, circulates primarily
among more experienced computer users.
Although e-mail encryption protocols differ in particulars,
any such protocol must provide two things: confidentiality and authentication.
Confidentiality indicates that only the authorized recipient of a message
can read it. Authentication is the guarantee that the message is really
from its stated originator.
To provide these services, all primary protocols for encrypting e-mail rely
on single-key encryption to encrypt the message, public-key encryption to
encrypt the secret key, and a hash function to provide authentication.
With single-key encryption, both the sender and the recipient must know
the key ahead of time--a logistical challenge if the two have never exchanged
e-mail before. On the other hand, public-key encryption is not practical
for use in encrypting entire messages since it is computationally expensive.
Thus, typical e-mail protocols use single-key encryption to encrypt the
e-mail message and then public-key encryption to encrypt the key used in
encrypting the message.
Public-key encryption also works in conjunction with a hash function to
provide authentication. Essentially, a hash code, also called a message
digest, is a small chunk of data unique to a particular message but much
shorter than the message itself. The hash code is like a digital fingerprint--each
is so unique that it makes it possible to distinguish one message from another
even if the two differ by only one bit.
To provide authentication, an encryption protocol extracts a hash function
from the message, encrypts it with the sender's private key, and appends
the encrypted hash code to the message. Although the hash code can be decrypted
by anyone who knows the public key, a successful decryption verifies the
identity of the originator.
One of the key differences among the primary protocols is the
size of the key used in encryption. Key size is an extremely important factor
in ensuring cryptographic soundness. Since larger keys are computationally
more difficult to break, larger keys imply stronger encryption. Many encrypting
algorithms are considered cryptographically sound if and only if they are
used with fairly large keys.
Key management is another important issue in protocol design. The way e-mail
security schemes keep track of keys helps determine how much protection
they ultimately provide. With a public-key system, key management must fulfill
two parts: obtaining the public-key of a communications partner and verifying
that the key belongs to that party.
Consider the so-called man-in-the-middle attack, in which an intruder inserts
a relay process into the network between communications parties, pretending
to be both of them. Thus, when one party sends a message to the other asking
for his public key, the intruder intercepts the message and returns his
own public key instead. If he does the same with the recipient, the intruder
can intercept and decrypt encrypted messages from both sides and then re-encrypt
or substitute his own message. Meanwhile, each side continues to believe
it is communicating with the other.
Avoiding such attacks is one of the goals of key management. This is where
the certification authority comes in. Physically, the certification authority
can be many things, including dedicated hardware or software running on
a server. In practice, however, for most secure e-mail applications, the
certifications authority is likely to be software running on the mail server.
Regardless of the certification mechanism, the problem of bootstrapping
trust remains. At some point, whether the mechanism is a hierarchy or a
web of trust, the local e-mail client must be configured to specify which
certification authorities and users should be trusted and which should not
In the certification authority, a single root authority is hard-wired into
the client. The client software is then pre-configured to use a particular
certification authority or a set of users which cannot be changed by the
users. Thus, when a user first installs his e-mail package, the software
already "knows" who to trust. Recently, however, it has become
clear that a single root authority is not flexible enough. The answer is
configurable certification authorities which allow users to choose certification
authorities after verifying the authorities through some reliable channel.
The certification hierarchy most commonly used by secure e-mail schemes
is the X.509 spec from the International Telecommunication Union-Telecommunications
Standards Sector (ITU-TSS). The X.509 spec, which defines the relationship
of the certification authorities, is part of a X.500 series of recommendations
for a global directory structure using "distinguished names" to
track users and other objects stored in a directory.
Standards bodies and individual implementors have developed
five primary protocols for e-mail encryption. Although the purpose of the
schemes are the same--to protect e-mail as it travels over the Internet--they
different encryption algorithms and authentication techniques. They also
take different approaches to issues such as key management--issues which
are just as important as encryption in ensuring the integrity of data. Finally,
the design decisions vendors make to implement these protocols can determine
how easy they are to use and administer, especially on a large network.
Secure Multipurpose Internet Mail Extensions
Because it supports the encryption of multimedia data as well
as plain text, S/MIME is considered one of the most fully featured secure
e-mail protocols and is likely to become an Internet Engineering Task Force
(IETF) standard. S/MIME guarantees interoperability between any two implementations
while rival protocols only allow implementors to choose non-interoperable
encryption options. Many Internet mail vendors plan to support the scheme,
although commercial versions of S/MIME are not available yet. S/MIME is
being developed by a group of companies led by RSA Data Security, Inc.
One potential drawback with S/MIME is that it permits the use of keys that
are too small to ensure adequate security. All S/MIME implementations must
include RC2, a proprietary crytographic algorithm from RSA with a 40-bit
key. Because RC2 represents the lowest common denominator, any two S/MIME
implementations can communicate with each other using RC2 as the default
algorithm. The trade-off is security, however, since the 40-bit RC2 is not
unbreakable. In fact, in a presentation at a Business Software Alliance
conference last November, a panel of security experts maintained that 40-bit
keys offer virtually no protection. They claimed that a hacker could easily
crack a single 40-bit message in a weekend using borrowed computer time.
According to the panel, a corporate engineering department with a $300,000
budget could build a machine that could crack five messages a second. Thus,
any encryption scheme that operates with 40-bit keys should not be considered
completely secure. However, S/MIME also supports two different versions
of DES that use 56- and 168-bit keys.
Another cryptographic weakness of S/MIME, one that also affects MSP but
not other schemes, is that eavesdroppers can distinguish between encrypted
and signed-and-encrypted messages. More importantly, they can also tell
who signed the message. If, for example, Marc Andreesen and Bill Gates were
to exchange a number of signed messages, an eavesdropper might draw conclusions
that Netscape and Microsoft were forging a business relationship.
S/MIME's key management scheme users the hierarchical structure defined
in the latest update of X.509--Version 3, the most flexible up to date.
This explicitly allows certification paths to start in the local security
domain--the authorities controlled by an organization, for instance--of
the public-key user system. The other notable feature of X.509 Version 3
is its ability to bind keys directly to e-mail addresses as opposed to only
the X.500 distinguished names.
S/MIME's key management model is somewhat of a hybrid between certification
hierarchy and the web of trust. Network manages must configure their own
list of trusted keys--usually, the keys of trusted certification authorities.
Full certificates are then included with every S/MIME signed message. Although
this makes every message a few hundred bytes larger, this also makes key
management a little easier. Whenever a user receives a signed message, the
key of the signatory goes into the local key database where all keys are
stored. If the key is signed by a trusted authority, trust is then established
for that key and it can be used in the future without reverification.
Developed by Zimmermann in 1991, PGP has since become the de
facto standard for mail encryption on the Internet. PGP is already commercially
available via Viacrypt and other software companies. Version 3.0 will be
released this year.
PGP's underlying cryptography is quite sound. For message encryption, it
uses the International Data Encryption Algorithm (IDEA) with a 128-bit key.
In fact, there is no mode of operation with a smaller key size, which makes
PGP exceptionally safe.
Zimmermann designed PGP with the web of trust model for key management.
This tend to be one of the scheme's biggest drawbacks, especially in business
settings. Overall, PGP's key management is hard to learn, time-consuming,
and dependent on a great deal of manual intervention. Every time a new key
is needed either for checking a signature or encrypting e-mail, users must
perform several manual operations. First, they must obtain the key, which
maybe a remote operation. Next they must put the key into the local PGP
key database, called a keyring in PGP terminology. Finally, they need to
check if the key is valid. In PGP's current implementation, these steps
require fairly complex command lines. However, key management may eventually
become easier since this complexity is not inherent in PGP.
One of PGP's biggest drawbacks is in its inability to handle
multimedia data. PGP/MIME attempts to solve the problem by adding the ability
to handle MIME objects. The spec is currently wending its way through the
IETF standards process. There are currently two implementations of the PGP/MIME
draft: premail and the PGP/MIME reference implementation. Neither has been
MIME Object Security Services (MOSS)
Another attempt to encrypt multimedia data is MOSS, currently
defined in Requests for Comment 1847 and 1848. Although no full-fledged
commercial implementations have been available to date, there is broad support
for adapting some parts of MOSS for use with other crytographic protocols,
particularly the multipart/signed format which allows different parts of
a message to be individually encrypted and signed.
Message Security Protocol (MSP)
MOSS is generally cryptographically sound, but since it does not mandate
the choice of encryption algorithm nor key size, it could possibly operate
with a small (not secure) key size. Also, choosing a different encryption
algorithm can cause MOSS to fail to interoperate with other MOSS applications.
MOSS supports two modes of key management: X.509 and manual, using the web
of trust. Users get more options, but a missing feature is the crytographic
hash of the public key. Without it, users must either blindly trust the
mechanism delivering the key or examine the entire key for signs of tampering,
which can be cumbersome.
MSP attempts to revamp the secure e-mail protocols developed
by NSA for its Defense Secure Data Networking Systems for use with Internet
and MIME mail. The underlying protocol is standardized in the SDN.700 series
of documents from the National Institutes of Standards and Technology while
MIME integration is an IETF Internet draft. Although MSP has been around
for a while, especially in X.400 environments, specifications and implementations
for e-mail have emerged only recently.
A major issue with all of these schemes is how they will be integrated with
mail packages. In some cases, the encryption scheme is an additional application--for
example, PGP is packaged in a single application that can work with existing
mail packages, which is a large factor in its popularity. In general, however,
these schemes are best integrated directly into the mail package. One reason
for this approach is because integration makes it possible to integrate
key management and the users' address books. If the encryption protocol
is not integrated with the mail application, two address books are required:
one to select e-mail address, the other to map e-mail addresses to keys.
Because this requires users to keep track of two separate databases, this
approach can be cumbersome.
MSP contains two features that no other protocol does. The first is a cryptographically
strong signed-receipt capability, also known as a non-repudiation of receipt,
which offers proof that the recipient actually received a message and that
the message received is identical to the message sent. The second feature
is the ability to classify messages--as "top secret," for example.
An MSP client rejects messages if users do not have the appropriate authorization
to read them.
One problem with MSP is that different implementations are not guaranteed
to interoperate since two implementations may choose different and non-overlapping
sets of algorithms. Before communication is possible, users must agree on
both the MSP protocol and a specific algorithm set. The two main algorithm
suites implemented for MSP are the Mosaic suite--same as the encryption
used with the Clipper chip--and RSA with DES. A few implementations of MSP
are already shipping, with several more in progress. MSP is a good choice
for interactions with the government.
The key management of MSP, based on the X.509 certification hierarchy, closely
resembles S/MIME's. Thus, it raises similar issues about signed message
formats. At present it is still not clear whether MSP will be implementing
the X.509 extensions that make handling of certification more flexible.
4A of the Uniform Commercial Code ; regulations E and Z of the board
of governors of the Federal Reserve System, 12 C.F.R. 205, 225.
2 A.B. 1577, An Act to Add Sec. 16.5 to the Government
Code, Relating to Digital Signatures (1995).