Addressing uncertainty during renaming using space-time contexts

This thesis addresses the problem of disruption during renaming in the Internet. Protocol names such as IP address change over time for a number of technical and non-technical reasons, and such changes result in disruption of communication. This disruption is acceptable in most circumstances but not all. In case of latter, additional approaches are required to deal with the fundamental problem underlying the disruption – uncertainty in terms of scope and cost of disruption. This work suggests an architectural mechanism, called space-time contexts, to deal with this uncertainty. The basic idea behind the mechanism is that namespaces such as IP do not have a notion of time, and that formally including a limited notion of time into the namespaces enables an alternative change model that can better deal with uncertainty. Further this change increases the expressiveness of the network as a whole. The approach has three basic parts. In the first part we discuss the challenges introduced by addition of temporal dimension to namespaces including feasibility, scalability, generality and usability. The second part presents a simple dependency graph-based model to evaluate the effectiveness of contexts. The third part presents a description of a simple prototype that realizes contexts. The prototype combines extensions to the host operating system and user-level coordination to demonstrate the feasibility and enable semi-automatic management of the contexts. Beyond addressing the risk associated with specific changes, contexts increase the expressiveness of the layered network in terms of control enabling a variety of advanced network management capabilities. This dissertation also presents potential applications of contexts.