Many of the technical vulnerabilities which make mail bombing a serious threat
could be significantly mitigated by enhancing the security infrastructure of the Internet.
Published by Linn in 1988, RFC 1040 [13] discussed
SMTP message encryption and authentication procedures.
His continuing work in RFC 1115 [14] specified the
cryptographic algorithms to support Privacy Enhanced Mail (PEM),
including the use of public-key encryption algorithms.
Presently, the PEM specification is contained in four RFCs which
are all a part of the ``Privacy Enhancement for Internet Mail'' series
(RFCs 1421-1424):
Part 1:
Message and Authentication Procedures
Part 2:
Certificate-Based Key Management
Part 3:
Algorithms, Modes, and Identifiers
Part 4:
Key Certification and Related Services
The core security problem with e-mail bombs is the authentication
of the originator. PEM (and similar cryptographic e-mail services)
offer both symmetric and assymetric authentication, supporting a
wide variety of cryptographic algorithms [16]. Methods for processing
mail addressed to mailing lists are also provided; however,
cryptographic authentication remains problematic for e-mail transport [17]. The reader
is referred to many numerous references on the subject [16]
[17] for a more detailed technical discussion.
Unfortunately, PEM provides integrity protection only on the body
of a message. The header fields of an SMTP message are not protected
because MTAs need to modify many of the header fields during
e-mail transport [17]. As pointed out earlier, the entire SMTP
infrastructure relies on a complex, heterogeneous internetwork of
MTAs and MUAs. Therefore, cryptographic solutions which work robustly
with intermediate systems are very difficult to design. Scalability
and interoperability become complex technical issues which are very
expensive to design, implement and sustain.
Scalability and interoperability concerns are also
major obstacles in the global management of crytographic keys, known as
the emerging Public Key Infrastructure, PKI.
Combining a heterogeneous public-key infrastructure with a robust
global MTA infrastructure will provide the developer community with cryptographic
tools to address e-mail sender authentication. However, by no means will
a PKI solve the e-mail bomb threat without MTA and MUA integration.
Finally, PKI-MTA integration may not significantly mitigate
the e-mail bomb problem because PKI was
designed to address confidentaility, authentication, integrity,
and non-repudiation and not denial of service attacks.
All technologies that present new opportunities also propagate new vulnerabilities
and risks.
Emerging public-key cryptosystems are traditionallly viewed
as defensive mechanisms; strenthening the integrity, confidentialility, and authentication
of our electronic infrastructure. However, the widespread availability of
cryptosystems create potential offensive threats to the infrastructure
which are normally considered after design and deployment. The interested
reader is referred to an excellent paper by Young and Yung [18] which
discusses cryptovirology and cryptoextortion; emerging topics which are
out of the scope of this paper.