Why I Teach People How To Hack
Here I discuss my philosophy for why I focus on teaching hacking instead of traditional security. In short, I think that defense is best served if you adopt the hacker mind-set, the understanding about what can wrong and how it can be exploited. I adapt this approach in university courses, hacking competitions, and a security company I founded called Syndis. The approach is effective and engaging for students, but also raises important ethical questions.
Anatomy of a Phishing Attack
I was contacted by the Vocativ news network in 2013 and asked if I could do a live hack against a reporter. With the help of my colleagues at Syndis, here you can see results of a very simple but common attack in a somewhat dramatic portrayal...
How Alan Turing Cracked The Enigma
This general-audience lecture explais the details of the elusive and fearsome Enigma cryptographic machine that Germans used to encrypt message in WW2, and how the Allies systematically exploited the system. The talk gives an insight into some of the major contributions by Marian Rejewski and Alan Turing, and attempts to explain the ingenuity of their insights.
How Do I Learn How To Hack? (In Icelandic)
I gave a 90 minute brief introduction to the basics of hacking and how hackers operate at Reykjavik University a while back. The talk is in Icelandic. If you want to learn more, most of the material can be found in expanded form in the hacking courses I have been teaching.
Elastic Storage Via File Motifs
I gave this research talk at the USENIX HotStorage 2014 workshop, presenting an elastic storage system we have been working on. In short, cloud systems make it easy to elastically dial up CPU and memory resources for virtual machines, but it's hard to scale back storage since we're committed to not losing data. Harmonium allows you dynamically scale back storage resources by exposing a "motif" abstraction: a recipe for how files were created and how they could be recreated. Harmonium transparently removes old files, allowing them to be seamlessly recreated later when they are needed. You can read more in the accompanying paper.
Why It's a Small World
On Facebook, you know pretty much any person on the planet through a friend of a friend of a friend of a friend. In more traditional settings, the distance is about 6. Why is this case? In this Pearls of Computation talk given at Reykjavik University in 2014, I discuss some of underlying reasons and the models that scientists have come up with the explain our "small-world". (Most of the slides were borrowed from my colleage Jure Leskovec at Stanford University).
Ultra Scalable Messaging Systems
My PhD work included two systems that deal with scalability of data replication and messaging: Dr. Multicast and Kevlar. The former addresses scalability challenges of IP Multicast within data centers, arguing that the technology can be safely used. Kevlar is about enabling efficient messaging in wide-area networks, seamlessly bridging together different messaging systems. The talk is from 2011 so some of the context is a bit dated.