Autonomous Management for Pervasive Computing

Network complexity will increase dramatically over the next 5 years as will the amount of devices inhabiting

these networks. Ad-hoc and active paradigms will make the already onerous task of network management

increasingly problematic. An approach to managing such networks based on bacterial colony behaviour is

discussed, offering innate abilities for essential tasks such as software proliferation, load balancing and

differing but distinct qualities of service. Robustness to fractal request streams is also demonstrated using

real world requests as a source of simulated network load. The ‘hands off’ element of the adaptive algorithm

is a major asset for any configuration and optimisation task. This biologically inspired adaptive

management solution could be the ideal approach to managing the behaviour of complex data networks of

the future.