The presence of time dependence on different types of connections has motivated new traffic engineering research. In this dissertation, we examined the impact of this time dependence on the video connections admission regions. We considered several video sequences with different dependence degrees or Hurst parameters. For Hurst estimation, we developed a software package named HEP (Hurst estiamtor package). The HEP package contains the R/S, Higuchi, AV (wavelet) and variance estimators. In our analysis, we studied two different call admission approched: a descriptor-based machanism (DBCAC) and a measurement-based mechanism (MBCAC). Both approaches are based on effective bandwidth criteria. For the DBCAC mechanism, we considered an effective bandwidth formulation using the Hurst parameter of the video connections. We obtained the admission regions based on analytical and simulation analysis. An important result showed that the admission regions were similar for the sources presenting long-range dependence on the admission regions. For the MBCAC mechanism, we proposed a new measurement scheme composed by a Kalman filter, a Hurst estimator and a effective bandwidth algorithm. In this proposal, the dependence degree was included in the measurement process o allow on-line Hurst parameter estimation. The mean, variance and Hurst estimatives were used to evaluate to the network condition and improved the effective bandwidth estimation. Moreover, the MBCAC mechanism was robust to the declaration errorsn resulting in moe accurate admission regions. The MBCAC results also confirmed the results obtained with the DBCAC mechanism, i.e; that time dependence had a small impact on admission regions.