Date of Award
Campus Access Dissertation
Doctor of Philosophy (PhD)
Large-scale network systems, such as Internet data centers, grid/cloud computing systems, are increasingly expected to be autonomous, scalable, adaptive to dynamic network environments, survivable against partial system failures and simple to implement and maintain. Based on the observation that various biological systems have overcome these requirements, the proposed architecture, SymbioticSphere, applies biological principles and mechanisms to design network systems (i.e., application services and middleware platforms). SymbioticSphere follows key biological principles such as decentralization, evolution, emergence, diversity and symbiosis. Each application service and middleware platform is modeled as a biological entity, analogous to an individual bee in a bee colony, and implements biological mechanisms such as energy exchange, migration, replication, reproduction and death. Each service/platform possesses behavior policies, as genes, each of which defines when to and how to invoke a particular behavior. SymbioticSphere allows services and platforms to autonomously adapt to dynamic network conditions by optimizing their behavior policies with a genetic algorithm. Through the evolutionary process, services and platforms also strive to satisfy given constraints for quality of service (QoS) such as response time, throughput and workload distribution. SymbioticSphere also allows services and platforms to autonomously seek stable adaptation decisions as equilibria between them and yield stable performance results that contain a very limited amount of fluctuations. This dissertation describes the design of SymbioticSphere and evaluates how the biologically-inspired mechanisms in SymbioticSphere impact the autonomy, adaptability, scalability, survivability, and simplicity of network systems.
Champrasert, Paskorn, "Leveraging Biologically-inspired Self-Adaptive Architecture in Network Systems" (2011). Computer Science Dissertations Collection. 1.