Improving network robustness

We present a simple, fully decentralized approach to improving robustness of existing unstructured overlay networks against a selective deletion of nodes (e.g., attacks on network hubs). The approach is based on modifying the network by adding or rewiring links at random. We quantify the effectiveness of adding randomness to the network by computing the curve governing the tradeoff between the number of modifications and the increase in robustness. For certain networks, a relatively modest amount of randomization can significantly improve the average path length (quantifying performance degradation) and the size of the largest connected component (quantifying network availability) after an attack.

[1]  Kai-Yeung Siu,et al.  Distributed construction of random expander networks , 2003, IEEE INFOCOM 2003. Twenty-second Annual Joint Conference of the IEEE Computer and Communications Societies (IEEE Cat. No.03CH37428).

[2]  J M Carlson,et al.  Highly optimized tolerance: a mechanism for power laws in designed systems. , 1999, Physical review. E, Statistical physics, plasmas, fluids, and related interdisciplinary topics.

[3]  Christos Gkantsidis,et al.  Random walks in peer-to-peer networks , 2004, IEEE INFOCOM 2004.

[4]  Massimo Marchiori,et al.  Error and attacktolerance of complex network s , 2004 .

[5]  Sugih Jamin,et al.  Inet-3.0: Internet Topology Generator , 2002 .

[6]  A. Rbnyi ON THE EVOLUTION OF RANDOM GRAPHS , 2001 .

[7]  P. Erdos,et al.  On the evolution of random graphs , 1984 .

[8]  Albert-László Barabási,et al.  Error and attack tolerance of complex networks , 2000, Nature.

[9]  Hiroki Sayama,et al.  Optimization of robustness and connectivity in complex networks. , 2003, Physical review letters.

[10]  Mark E. J. Newman,et al.  The Structure and Function of Complex Networks , 2003, SIAM Rev..