|
The design of computer network configurations needs to account for evolution. Inspired by natural
complex systems, we proposed composing the computer network out of Network
Cells (NC) capable of evolving through continuous learning and adaptation. We also
proposed the CellNet framework within which the NC functions and interactions are
defined. CellNet framework is derived based on the concept of separation of concerns
that tackle complexity by dividing the problem space into smaller manageable,
loosely-coupled, and easier to solve sub-problems allowing for better understanding
and design. The CellNet framework abstracts network functions into three main core
concerns and one crosscutting concern. These identified abstractions will be the
reference model for specializing the NC. Our future work will encompass formalizing
a new design methodology that we refer to as Cell Oriented Architecture on which the
NC and the CellNet framework are based. We intend to construct a network prototype
composed of NC using the CellNet framework to study the performance gains
achieved when introducing learning and adaptation to networks.
|
