Improving Reliability through Smart Redundancy Yuriy Brun, George Edwards, Jae young Bang, Nenad Medvidovic Advisor: Nenad Medvidovic University of Southern California, Los Angeles

Abstract

Many software systems today, such as computational grids, include faulty and untrusted components. As faults are inevitable, these systems utilize redundancy to achieve fault tolerance. Typically, the redundancy schemes do not utilize runtime information to optimize the efficiency and reliability of the system. We present iterative redundancy and progressive redundancy, two new "smart" redundancy techniques that focus on making automated decisions based on runtime information to make the systems efficient and self-adaptive.

 

Bio

Yuriy received his Ph.D. from the University of Southern California in 2008 working on self-assembling Internet-sized computational systems. His interests are in the area of engineering self-adaptive systems. In particular, using mechanisms from nature to create engineering paradigms for robustness, fault and malice tolerance, scalability, and security. To that end, he does (1) theoretical work on design and complexity analysis of biologically inspired algorithms and (2) software engineering work on implementing these algorithms for Internet-sized distributed systems, grids, and clouds. His interests also include mathematically modeling biological and chemical systems and using such models to improve our ability to engineer complex systems. His work involves discreet math, mathematical modeling, theoretical computer science, software engineering, programming language design, and software architectures. Prior to his work at USC, Yuriy received his B.S. and M.Eng from MIT in 2003.