Computer Science

Biologically Inspired Networking

See allHide authors and affiliations

Science  31 Mar 2006:
Vol. 311, Issue 5769, pp. 1837
DOI: 10.1126/science.311.5769.1837b

Biological systems are typically better at adapting to new situations than computers because their design emphasizes robustness and sustainability even though the proximal response may not be the optimal one. In an information network such as the Internet, data are broken up into packets before being transmitted, and each packet can take a different path across the nodes of the network. How might a method for data transmission over multiple paths be redesigned whereby the network can itself adapt to an unpredictable and fluctuating environment?

Leibnitz et al. based their biologically inspired network routing scheme on a model developed to account for the response of Escherichia coli bacteria to variations in nutrient availability. The model uses stable attractors: equilibrium states into which the system settles until disrupted by a change in the environment, at which point the system converges to a new attractor. For network switching, information about the data paths (available bandwidth or transit time) is collected to find a stable attractor. When conditions change (for example, if a link breaks), a new attractor is selected, and the packets are switched to a new path. Because randomness is an intrinsic feature of the optimization method, the system is highly stable in noisy environments. — DV

Commun. Assoc. Comput. Mach. 49, 63 (2006).

Navigate This Article