Title
A Software Based Approach for Trusted Agent Execution on Malicious Host
Abstract
Absolute protection of mobile agents from attacks by malicious hosts is an open research problem. We propose a software based paradigm whereby an agent is protected from various static and dynamic attacks from a malicious host of an unknown hardware configuration, for a specific period of time. This time interval is computed by restricting the maximum resources that may be available to the adversary and the time complexity of the critical static and dynamic attacks that it may launch. We employ the technique of oblivious hashing (OH) using overlapped instructions [1], with multilevel pointer aliasing to thwart static analysis and instant code modifications. The host is required to obtain the aggregate OH value of the whole agent by executing it in an unobtrusive environment and to send it back to the agent originator within the specified time interval. To provide unobtrusive environment, we employ external timing analysis to detect major dynamic attack tools such as debuggers, virtual machines and emulators. Experimental results are presented that demonstrate the viability of the timing analysis mechanism in detecting dynamic attack tools on a range of Intel based machines.
Year
DOI
Venue
2009
10.1109/CSE.2009.398
Computational Science and Engineering, 2009. CSE '09. International Conference
Keywords
Field
DocType
computational complexity,cryptography,multi-agent systems,program debugging,program diagnostics,virtual machines,Intel based machines,debuggers,dynamic attacks,emulators,external timing analysis,instant code modifications,malicious host,mobile agents,multilevel pointer aliasing,oblivious hashing,overlapped instructions,software based approach,static analysis,static attacks,time complexity,trusted agent execution,virtual machines
Virtual machine,Computer science,Cryptography,Pointer aliasing,Static analysis,Computer network,Multi-agent system,Software,Static timing analysis,Hash function,Distributed computing
Conference
Volume
ISBN
Citations 
2
978-0-7695-3823-5
3
PageRank 
References 
Authors
0.40
10
2
Name
Order
Citations
PageRank
Sarosh Hashmi1182.27
John Brooke26216.01