Toward Formalized Autonomic Networking

Abstract

Autonomic systems were first described in 2001. Over the last couple of years, autonomics has gained a lot attention in both academia and industry. After the initial interest, the basic concepts of autonomics remain extremely promising in the area of networking, in particular when considering the flexibility provided by the emerging technologies in virtualization and software defined networking. The fundamental goal of autonomics is self-management, including self-configuration, self-optimization, self-healing and self-protection. An autonomic network is defined when nodes have minimal dependencies on human administrators or centralized management systems. This paper formalizes autonomic networking.

Concurrency of self-* in autonomic systems

“…Autonomic computing has been set as an important requirement for systems devised to work in new generation global networked and distributed environments like wireless networks, P2P networks, Web systems, multi-agent systems, grids, and so on [44][45][46][47][48]. Such systems pose new challenges for the development and application of autonomic computing techniques, due to their special characteristics including: nondeterminism, context-awareness and goal-and inference-driven adaptability [1][2][3][4].…”

“…The homomorphic structure is dynamic in nature because self-*facets can be dynamically enabled or disabled in AS. For such each action of enabling or disabling, transformation [2,18,55] for the homomorphic structure on AS occurs.…”

“…Programming and developing concurrent applications in such ASs are raising challenges not tackled before, including nondeterminism, decentralization, context-awareness and goal-and inference-driven adaptability [1][2][3][4]. Hence, in future, we hope to enable further developments which are not exposed yet in this paper.…”

Products, Coproducts and Universal Properties of Autonomic Systems

“…We need such an action for which self-*action Given autonomic systems AS, AS ′ , and AS ′′ , and actions self-*action 3 : AS ′′ → AS and self-*action 4 : AS ′′ → AS ′ , there is a unique action AS ′′ → AS × AS ′ that commutes with self-*action 3 and self-*action 4 . We call it the induced action AS ′′ → AS × AS ′ , meaning the one that arises in light of self-*action 3…”

“…The term derives from the body's autonomic nervous system, which controls key functions without conscious awareness or involvement" [1]. AC in our recent investigations [2][3][4][5][6][7][8] is generally described as self-*. Formally, let self-* be the set of self-__'s.…”