What is Self-Organisation?

This is the second post in a series of posts about CloudLighting. In this post, we discuss self-organisation, one of the principles behind the CloudLighting system. The self-organising behaviour built into, and exhibited by, the cloud infrastructure will address the complexity of cloud ecosystem.


Self-organisation is a powerful technique for addressing complexity and borrows heavily from the natural sciences and the study of natural systems.

Alan Turing once observed that global order arises from local interactions. Components in a self-organising system are mutually dependent and typically only interact with nearby components. However, the system is dynamic and therefore the components can change state to meet mutually preferable states. As they meet these states, they adapt and achieve fit and this propagation of fit results in system growth.

As more and more components adapt and become assimilated within the system, complexity increases to incorporate the individual characteristics of components. Growth only stops when resources have been exhausted and self-maintenance is the de facto purpose of the system.




The application of self-organising and self-management principles to cloud computing is at an early stage. Research shows that cloud computing systems are inherently self-organising and while they exhibit autonomic features are not self-managing as they do not have reducing complexity as a goal.

However, cloud computing represents a complex system and therefore self-organisation is an appropriate technique to address this complexity. Zhang and Marinescu propose an auction-driven self-organising cloud delivery model based on the tenets of autonomy of individual components, self-awareness, and intelligent behaviour of individual components. They simulate a new model of a complex heterogeneous system with a very large number of components and with many interaction channels among them. Preliminary results suggest that the self-organising architecture was scalable and the bidding mechanisms and coalition formation algorithms are feasible at scale.

Leave a Reply