Optimizing SLA-driven adaptive routing
Georgios Katsikogiannis, Sarandis Mitropoulos, Christos Douligeris · 2016
As network complexity and heterogeneity have increased dramatically in the last decade, modern policy-based management techniques for adaptive routing decisions regarding path selection optimization and intelligent path control are necessary. In this paper, we analyze the application performance requirements, and the event-based management options for adaptive routing. Then, we propose an innovative model to monitor the appropriate indicators with the suitable detectors, which improves the dynamic routing performance. We present the design and implementation of Performance-based and Event-Management Routing (PEMR) model and we validate the proposed model by enforcing the appropriate adaptive routing policies and decisions. PEMR can resolve failed or poorly performance paths into alternative routing paths automatically and it can adapt quickly to dynamic network conditions. Besides, we can avoid any conflicting policies (i.e. traffic load distribution versus cost minimization, path availability versus response time), and facilitate the optimum path selection.