Classes of Optimal Network Topologies Under Multiple Efficiency and Robustness Constraints.
ABSTRACT We address the problem of designing optimal network topologies under arbitrary optimality requirements. Using three critical system parameters, efficiency, robustness and cost, we evolve optimal topologies under different environmental conditions. Two prominent classes of topologies emerge as optimal: (1) Star-like topologies, with high efficiency, high resilience to random failures and low cost, and (2) Â¿Circular Skip ListsÂ¿ (CSL), with high robustness to random failures as well as targeted attacks, and high efficiency at moderate cost. We analyze CSLs to observe that they show several structural motifs that are optimal with respect to a variety of metrics.
- SourceAvailable from: cs.cmu.edu
Conference Proceeding: The impact of DHT routing geometry on resilience and proximity.Proceedings of the ACM SIGCOMM 2003 Conference on Applications, Technologies, Architectures, and Protocols for Computer Communication, August 25-29, 2003, Karlsruhe, Germany; 01/2003
Conference Proceeding: Cayley DHTs - A Group-Theoretic Framework for Analyzing DHTs Based on Cayley Graphs.[show abstract] [hide abstract]
ABSTRACT: Static DHT topologies influence important features of such DHTs such as scalability, communication load balanc- ing, routing efficiency and fault tolerance. Nevertheless, it is commonly recognized that the primary difficulty in design- ing DHT is not in static DHT topologies, but in the dynamic DHT algorithm which adapts various static DHT topolo- gies to a dynamic network at Internet. As a direct conse- quence, the DHT community has been paying more atten- tion to the dynamic DHT algorithm design, resulting in a variety of DHT systems lacking of a common view for anal- ysis and interoperation.In this paper we reiterate the im- portance of static DHT topologies in the DHT system de- sign by analyzing and classifying current DHTs in terms of their static topologies based on a grouptheoretic model: Cayley graphs. We show that most of current DHT propos- als use Cayley graphs as static DHT topologies, thus tak- ing advantage of several important Cayley graph proper- ties such as vertex/edge symmetry, decomposability, opti- mal fault tolerance and hamiltonicity. We observe that sev- eral non-Cayley-graph based DHT proposals such as Ko- orde/D2B/Distance Halving and Pastry/Tapestry also rely on techniques in their dynamic DHT algorithm design try- ing to imitate desirable Cayley graph properties. Based on Cayley graphs, we propose the class of Cayley DHTs as a unified group-theoretic model for investigating DHTs from a graph theoretic perspective. The significance of Cayley DHTs is in their explicit inspiration to a uniform dynamic DHT algorithm design, which can directly leverage alge- braic design methods thus is able to generate sets of high- performance DHTs adopting various Cayley graph based static DHT topologies but still sharing the same dynamic DHT algorithm.Parallel and Distributed Processing and Applications, Second InternationalSymposium, ISPA 2004, Hong Kong, China, December 13-15, 2004, Proceedings; 01/2004
- [show abstract] [hide abstract]
ABSTRACT: An important feature of many complex systems, both natural and artificial, is the structure and organization of their interaction networks with interesting properties. Here we present a theory of self-organization by evolutionary adaptation in which we show how the structure and organization of a network is related to the survival, or in general the performance, objectives of the system. We propose that a complex system optimizes its network structure in order to maximize its overall survival fitness which is composed of short-term and long-term survival components. These in turn depend on three critical measures of the network, namely, efficiency, robustness and cost, and the environmental selection pressure. Using a graph theoretical case study, we show that when efficiency is paramount the "Star" topology emerges and when robustness is important the "Circle" topology is found. When efficiency and robustness requirements are both important to varying degrees, other classes of networks such as the "Hub" emerge. Our assumptions and results are consistent with observations across a wide variety of applications.03/2004;