摘要: While delay-critical applications typified by online multimedia communication are growing rapidly, the end-to-end (E2E) delay behavior in Internet remains poorly understood. This paper proposes a stochastic process model, which E2E alternations classified into two categories, jump and perturbation, according to whether statistical characterizations alter or not. As chief type of majority alternations, perturbations generally occur continuously period, can be analogous an ergodic stationary process. The authenticity model is verified with real life measurements collected all pairs pings (APP) project through months from PlanetLab nodes. Based on several estimation algorithms comparatively analyzed, experimental results demonstrate that terms minimizing mean squared error, most accurate prediction minimum recent measurements.