Heap Defend Signed Electronically with airSlate SignNow

Heap defend electronically signed

any questions all right sagia calls it an invite you thought I call it research report because it's really what it is is more like a good bunch of you know brain you know brain dumps from me so thanks for being here my enemies univer- lumberg as a security architect today my talk is about heat models for exploited seems so half of the talk will be pretty much about remaining you guys what we're talking about making sure we on the same page and the other half of the talk would be a bit more formal or tentatively formal with you know the kind of insights I think I have gained by analyzing a lot of different heap allocators over the years so how many of you have written a heap overflow exploits before many all right excellent so if there is anything wrong you know what I say please interrupt me right away I'll be glad to be proven wrong on anything and always eager to improve so interrupt me anytime if you have questions during the talk you feel free to actually you know ask them right away you don't have to end to the end so let's get started the big picture of this work is what I called one or two years ago the automated exportation grand challenge as you may all be familiar with the security exploit is a way to basically take advantage of security vulnerability in a program either to do untrusted code execution or to steal secret information for example private keys and a bigger challenge so a lot of these security experts are eaten manually generally or most ninety-nine point nine percent of the time and 4.01 percent of the time people claim that they can sometimes automatically generate his exploits I am highly skeptical of that in real world conditions so this work is trying to understand how to basically resolve one of the problems that I set up in the Grand Challenge which is a set of 11 problems that I believe are very hard to solve so some of the problems are about how to specify what an exploit should do how to generate input for the exploit how to represent state space for the programming under attack how to explore concurrency scheduling how to infer privilege separation and things like that so it's not it's not trivial at all one of the problem which is problem I of of the grand challenge is a heap prediction so but how can you how can you infer the state of the heap you know in a complex program is a really hard problem most people haven't really studied the problem at all a lot of the research on heap analysis has been mostly for things such as shape analysis shape now this is basically how to analyze the shape of data structures in the heat for example if those are linked list graphs trees IEC click etc...