Assignment 6

The assignment is due Monday, April 28, 6.59 pm EDT, to be submitted via CourseWorks.

Some of the questions below are research questions, where you are asked to find information about a particular issue. You may use the Engineering Library, any text books you have, one of the paper from the class readings, or the web to come up with answers. Be sure to cite your sources. Generally, a paragraph or two should be sufficient to answer the question. There is no need to write a tutorial.

  1. As a next step towards building an interactive audio and video application, complete the following steps for your application:

    Your audio and video interoperate with the rat audio tool. Due to the JPEG encapsulation and vic stability issues, we will forego interoperability with vic. You can also use the rtp tools for debugging.

  2. Using Cisco or Juniper router specifications, determine what DiffServ modes (AF, EF?) and how many token bucket flows a modern router support. Choose a single router for your investigation.
  3. RTP inter-media synchronization: A system has two media streams, a G.711 audio stream and a video stream. The receiver receives two RTCP sender reports (SR). The first SR, for video, contains:
    NTP timestamp bf8c1623.7abc83a9
    RTP timestamp 18000
    The audio SR contains the following information:
    NTP timestamp bf8c1624.816fc9bc
    RTP timestamp 68880

    What audio sample timestamp should be played for the video frame labeled 20,000?

  4. A token bucket has an average rate of 30 Mb/s and a max. burst size of 100 Mb. Assume that the token counter is initially "full". The token bucket drops packets when there are no tokens. A host is connected via a 100 Mb/s network and starts transmitting at full line rate. How long can the host keep this up without having its packets dropped?
  5. We simplistically model an MPEG video source as sending I and P frames, with every 15th frame being an I frame. Assume that I frame has a size of 10,000 bytes, while the P frames are 4,000 bytes long. (Naturally, they are split into several packets.) (a) Describe a minimum-rate tocket bucket that fits this source. (b) Can you reduce either the token rate or bucket depth without changing the video source?