Self-Assessment and Reconfiguration Methods for Autonomous Cloud-based Network Systems
DS-RT '13 Proceedings of the 2013 IEEE/ACM 17th International Symposium on Distributed Simulation and Real Time Applications
| Hi-index | 0.00 |
Certifying a network system S involves the assessment of how good $S$ meets its intended QoS objectives in a backdrop of uncontrolled external environment conditions incident on S. For complex network systems where a reasonably accurate and tractable computational model of S may not be known, hierarchical approaches based on "cyber-physical systems" (CPS) principles are attractive. Here, a piece-wise linearized simple model of S allows a controller to drive S with test inputs and evaluate the output behavior of S over a limited operating region. With model plug-in and controller algorithm switching, a management entity reasons about the behavior of S under different environment conditions and test inputs, to certify S with high confidence. The paper presents a case study of multi-source video congestion control over a bandwidth-limited network path to illustrate our CPS-based certification method.