Probe Station Selection in Non-deterministic and Dynamic Virtual Network Environment

摘要

The probe station selection algorithm as the main problem to address probing-based network monitoring solutions provides an approach to select nodes in the network where the probe stations should be placed. The locations of probe stations determine directly the probing efficiency, monitoring capability, and instrumentation cost of network monitoring. However, the non-determinacy and dynamicity of network bring to probe station selection new challenges in virtual network environment. To tackle with new challenges, the paper proposes a novel algorithm (called PSND) for probe station selection in the nondeterministic network environment through using a probabilistic dependency model. Then the algorithm is improved so that it can adapt to the circumstance of dynamicity (virtual nodes adding and deleting) of virtual network. Simulations and experimental studies show that PSND algorithm has a good balance between the size of probe station set and the execution time of algorithm comparing to optimal algorithm and random algorithm.