A Systematic Approach to Constructing Families of Incremental Topology Control Algorithms Using Graph Transformation

TL;DR

This paper presents a systematic, model-driven methodology for designing and extending network topology control algorithms using graph transformation, enabling the construction of algorithm families with verified properties.

Contribution

It generalizes previous work to support the specification of families of TC algorithms and demonstrates this by re-engineering six existing algorithms and creating a new energy-efficient variant.

Findings

Successfully extended the methodology to support algorithm families.

Re-engineered six existing TC algorithms within the new framework.

Developed and evaluated a novel energy-efficient TC algorithm, e-kTC.

Abstract

In the communication systems domain, constructing and maintaining network topologies via topology control (TC) algorithms is an important cross-cutting research area. Network topologies are usually modeled using attributed graphs whose nodes and edges represent the network nodes and their interconnecting links. A key requirement of TC algorithms is to fulfill certain consistency and optimization properties to ensure a high quality of service. Still, few attempts have been made to constructively integrate these properties into the development process of TC algorithms. Furthermore, even though many TC algorithms share substantial parts (such as structural patterns or tie-breaking strategies), few works constructively leverage these commonalities and differences of TC algorithms systematically. In previous work, we addressed the constructive integration of consistency properties into the development process. We outlined a constructive, model-driven methodology for designing individual TC algorithms. Valid and high-quality topologies are characterized using declarative graph constraints; TC algorithms are specified using programmed graph transformation. We applied a well-known static analysis technique to refine a given TC algorithm in a way that the resulting algorithm preserves the specified graph constraints. In this paper, we extend our constructive methodology by generalizing it to support the specification of families of TC algorithms. To show the feasibility of our approach, we reneging six existing TC algorithms and develop e-kTC, a novel energy-efficient variant of the TC algorithm kTC. Finally, we evaluate a subset of the specified TC algorithms using a new tool integration of the graph transformation tool eMoflon and the Simonstrator network simulation framework.