Cryptographer from Carolina: Starting OFF on ONLamp

by Justin Troutman

I suppose an introductory post would fit, given that this is my initial blog entry. Without any ado, I'm Justin Troutman, from the greater Charlotte metropolitan area of North Carolina. I'm a maturing cryptographer, which essentially equates to the fact that while I practice it, I'm still immersed in the incessantly extensive theoretical nature of academic cryptography. I tend to focus primarily on the cryptanalysis of symmetric primitives, such as block ciphers, as well as MACs, or message authentication codes, built from block ciphers. In fact, I'm a MAC-zealot, you might say; it [authentication] is an aspect that I advocate quite often.

For the past couple of years, I've taken an exceptionally passionate interest in cryptovirology, which is an intriguing sub-genre of cryptography that involves the offensive, malicious application of cryptographic primitives in adversarial attacks. (It is, in itself, an infantile field, pioneered by Dr. Adam L. Young and Dr. Moti M. Yung - two cryptographic luminaries, respectively.) I'm in the midst of conducting some original research concerning cryptovirological information extortion, and the implications of game-theoretic, arbitrated protocols for shifting trust, ensuring fairness, and establishing formal notions of security, such as IND-CCA2 and INT-CTXT. The most practical facet of this research is an efficient, MAC-based cryptovirus.

Recently, I gave a lecture on cryptovirology and game theory, at Duke University, for the TIP (Talent Identification Program), which is basically an opportunity for young folks, in the middle school to high school range, to be exposed to a plethora of subject matter. I was incredibly pleased with the response of the students, and it sparked my enthusiasm for preparing cryptographic material with an aim towards that audience, as well as the layman-oriented population in general. As such, I've been experimenting with an informal style of writing about cryptography, in such a way that casually introduces readers to aspects of cryptography that aren't covered in the general media; it's intended to inform the layman, while containing material that even a seasoned cryptographer would appreciate. There's a large gap between the academic community and the layman community; it's this gap that I'm trying to close, even if just a little.

Departing from all the crypto-jive, I suppose I could say a little about the non-cryptographer in me. I'm a native of Gastonia, North Carolina, which is strongly evidenced by my unmistakably deep southern Appalachian accent and dialect, and I spend a considerable amount of time exploring this state that I love most, whether it be fly fishing, hiking, skiing, snowboarding, or letting my Sony capture what one's eyes capture the best. Oh, and my master-of-19-instruments best friend is teaching me the mandolin. It will complete the Southern me. I can already tear up the Andy Griffith theme song.

Perhaps I shouldn't admit that I'm a hardcore UNC basketball fan, since I lectured at Duke recently, but I can't help it; if you're raised to love the lighter shade of blue, you'll understand what I mean. I like my orange too, though - Tennessee football, that is. To cap off my sports interests, I'm becoming more and more proud to admit I'm a Cincinnati Reds fan, now that they're having a decent season. However, Barry Larkin's departure was the end of the Reds generation I grew up with. I'd like to see them build a team that can do what they did in '90.

Anyhow, that about sums it up for now. If you're so inclined to see what I'm up to, you can peruse http://www.justintroutman.org, which essentially points to http://www.extorque.com - Extorque, from which I consult on matters cryptographic and perform cryptanalytical research. It also houses links to articles and academic papers pertaining to cryptovirology and various other cryptographically-focused things. My blog entries will cover things of this nature, and I'm certainly interested in what you folks would like to see discussed. I'm relentlessly strict about good cryptography, and prefer a no-nonsense treatment of security, so I welcome the most critical of feedback.

Until then - cheers, from the humid Carolinas!

2 Comments

will
2006-08-15 06:28:33
are viruses that use rc4 to "guard" themselves covered by cryptovirology, or is it just malware that uses crypto as a weapon?
Justin Troutman
2006-08-15 16:04:47
Will,


Traditionally, it wouldn't be covered by cryptovirology; that is, the polymorphic viruses you read about in virus-related literature aren't of a cryptoviral nature, as described by Dr. Young and Dr. Yung. They coined the definition of a "cryptovirus," which states that it employs a public key; this public key (asymmetric cryptography) allows the cryptovirus to perform one-way operations that only the adversarial cryptovirus designer can reverse. Polymorphic viruses have integrated symmetric cryptography for quite some time, for the purpose of obfuscating their code and masking their presence. Dr. Young states that if a polymorphic virus contains a public key, and uses it in the manner described above, then it is a cryptovirus as well - thus, being covered by cryptovirology.


However, I've researched methodologies for constructing a cryptovirus that rely solely on symmetric cryptography, which I've given a less specific definition. In my research, I make some considerable changes and security trade-offs, for the sake of simplicity and the formalization of security proofs. It conflicts with the existing definition of a cryptovirus, but states that the cryptovirus itself may use symmetric or asymmetric cryptography, or both. As for polymorphism, the same situation applies; it is possible to realize a cryptovirus that both "uses crypto as a weapon" and also uses a primitive like RC4 to "guard" itself, as you say. Nowadays, however, we would probably go with a more conservative primitive, such as the AES. But, RC4 suffices for this explanation.


Cheers!