Method and apparatus for predicting video traffic
View Patent ↗A method and apparatus for predicting video traffic are provided. The method includes extracting the statistical properties of previous Moving Picture Experts Group (MPEG)-4 video traffic and predicting the size of a subsequent frame based on the statistical properties of the previous MPEG-4 video traffic. Therefore, it is possible to address such problems associated with traffic bursts as transmission delays or low throughputs and thus to improve the performance of communication and network systems that transmit video traffic.
1. A method of predicting video traffic, the method comprising:
classifying input video traffic into a number of types of frames;
estimating a probability density function (PDF) regarding a size of the frames; and
predicting a size of a subsequent frame using the PDF,
wherein the predicting the size of the subsequent frame, comprises calculating the size of the subsequent frame based on a difference between a predicted size of a previous frame and an actual size of the previous frame and a standard deviation of a size of a reference frame,
wherein the reference frames are classified into a plurality of sections based on the size of the reference frames to define control points, and the PDF regarding the size of the reference frame is estimated using the control points,
wherein the types of frames comprise an I-frame, a P-frame and a B-frame,
wherein step of estimating the PDF comprises:
estimating the PDF regarding a size of the I frame;
estimating the PDF regarding a size of the P frame; and
estimating the PDF regarding a size of the B frame.
2. The method of claim 1 , wherein the estimating the PDF, comprises estimating the PDF using a cubic-spline interpolation (CSI) method.
3. The method of claim 1 , wherein the video traffic is real-time MPEG-4 video traffic having variable bit rate (VBR) properties.
4. A non-transitory computer-readable recording medium storing computer-readable code for predicting video traffic, the computer-readable code comprising code for:
a monitoring unit which classifies input video traffic into a number of types of frames;
a statistical property estimation unit which estimates a probability density function (PDF) regarding a size of the frames; and
a frame prediction unit which predicts a size of a subsequent frame using the PDF,
wherein the frame prediction unit comprises:
a first frame predictor which predicts the size of the subsequent frame using the PDF provided by a first statistical property estimator;
a second frame predictor which predicts the size of the subsequent frame using the PDF provided by a second statistical property estimator; and
a third frame predictor which predicts the size of the subsequent frame using the PDF provided by a third statistical property estimator,
wherein reference frames are classified into a plurality of sections based on a size of the reference frames to define control points, and the PDF regarding the size of the reference frames is estimated using the control points,
wherein the types of frames comprise an I-frame, a P-frame and a B-frame,
wherein the statistical property estimation unit comprises:
the first statistical property estimator which estimates the PDF regarding a size of the I frame:
the second statistical property estimator which estimates the PDF regarding a size of the P frame; and
the third statistical property estimator which estimates the PDF regarding a size of the B frame.
5. The medium of claim 4 , wherein the video traffic is real-time MPEG-4 video traffic having VBR properties.