Consider N users participating in an audio/video conference using IP multicasting. Each user is equipped with a microphone, a camera, and speakers. Assuming no RTP mixers are used: How many distinct SSRCs are needed?
1. Suppose that a sender sends a silence suppressed RTP stream of G.711 audio to a receiver (G.711 is the name of the standard for 64Kbps PCM we discussed in class). During a talk spurt 20 msecs chunks of quantized audio samples are encapsulated into packets and sent to the receiver (every 20 msecs). Justify your answers for the following:
a. How many samples are included in each packet?
b. If the RTP time stamp is initialized to zero, and if the first talk spurt consists of three or more packets, what are the RTP time stamps of the first three packets in the first talk spurt (Hint the time stamps are relative to the first sample)?
c. Consider the last packet in a talk spurt and the first packet in the immediately following talk spurt. Does the difference of the RTP timestamps of these packets depend on the length of the silence period between the two talk spurts?
d. If the last two packets received at the receiver have a timestamps of 0 and 480 and corresponding sequence numbers of 1 and 3 (i.e. packet 2 is lost), are these last two packets in the same talk spurt or in different talk spurts?
e. Outline an algorithm that uses sequence numbers and time stamps of RTP packets to determine the beginning of a new talk spurt?
2. Consider the adaptive playout delay algorithm. Illustrate through a simple example that adjusting the playout delay at beginning of each talk spurt results in compressing and elongating the silence period between talk spurts.
3. Consider an RTP session consisting of four users, all of which are sending and receiving packets to the same multicast address. Each user sends video at 100 kbps.
a. RTCP will limit its traffic to what rate?
b. How much RTCP bandwidth will each sender be allocated?
c. How much RTCP bandwidth will each receiver be allocated?
4. Consider N users participating in an audio/video conference using IP multicasting. Each user is equipped with a microphone, a camera, and speakers. Justify your answers for the following. Assuming no RTP mixers are used:
a. How many distinct SSRCs are needed?
b. How many distinct canonical names are needed? How will canonical names be used?