Novel Mathematics-Inspired Algorithms for Self-Adaptive Peer-to-Peer Computing

This paper describes, and evaluates benefits of, a design methodology to translate certain mathematical models into the design of novel, self-adaptive, peer-to-peer (p2p) distributed computing algorithms (“protocols”). This methodology is potentially a good vehicle for translating natural phenomena, representable via mathematical models, into practical p2p protocols. Concretely, our first contribution is a set of techniques to translate certain discrete “sequence equations” rigorously into new p2p protocols called “sequence protocols”. Sequence protocols are self-adaptive, scalable, and fault-tolerant, with applicability in p2p settings like Grids. A sequence protocol is a set of probabilistic local and message-passing actions for each process. These actions are translated from terms in a set of source sequence equations. Without having each process simulate the source sequence equations, the emergent behavior of a sequence protocol in a p2p system is equivalent to, and ...

• Type: Conference
• Year: 2007
• Where: SASO
• Authors: Steven Y. Ko, Indranil Gupta, Yookyung Jo
• Keywords: SASO 2007, Self-Organizing Systems, Sequence Equations, Sequence Protocol, Source Sequence Equations