Background:
This assignment contains two parts. First, you will research and explain the operation of DASH, TCP and AQMs in the context of Internet media streaming. Second, you will be evaluating a short Ethernet packet tracefile and evaluating the traffic contained within in the context of SIP/RTP-based Internet Telephony. At all times you will be researching available online literature to augment what has been presented in class.
A binary Ethernet packet tracefile in tcpdump format will have been uploaded to your lower numbered RULE host (the one assigned in your lab class), called broadband-trace-2017.dmp.
You will be expected to learn how to use Wireshark or similar packet trace analysis software in order to complete this assignment. You may also be required to (re)learn how to use Excel (or equivalent tool e.g. OpenOffice Calc) for some basic data analysis and graphing.
General Questions:
1. With the appropriate use of diagrams and clear descriptive text:
• Explain why an MPEG-DASH server will contain copies of a movie encoded at different quality levels and segmented into chunks
• Explain how the MPD is used by a MPEG-DASH client to adapt its behaviour when retrieving content for playback, and clearly describe the differences between chunks made up of separate files and chunks represented as byte ranges
• Explain the difference between a chunk's representation rate and the rate at which a chunk is retrieved
• Explain the role of a playout buffer inside an MPEG-DASH client and when they are prefilled
2. Imagine you have a server containing a movie segmented into chunks and encoded at constant bitrates of 200, 500 and 1000 Kbps. With the appropriate use of diagrams and clear descriptive text:
• How many bytes make up the chunks representing both 2 seconds and 8 seconds of content at the 200, 500 and 1000 Kbps representation rates? (Six different types of chunks in total.)
• How many seconds would each of these chunks take to send if they were encapsulated in UDP/IP frames and transmitted at precisely 1600Kbps (at the IP layer), with no other overheads and no packet losses?
3. Imagine you have a standard web server containing a movie segmented into chunks and encoded at constant bitrates of 200, 500 and 1000 Kbps.
Assume an additional 320 bytes of HTTP response overhead is required to encapsulate and send each chunk in response to GET requests from a DASH client. A single DASH client is connected to this server over a path exhibiting 350ms RTT and offering 16000Kbps in each direction at the IP layer. With the appropriate use of diagrams and clear descriptive text:
• What is the path's BDP?
• Assume the TCP layer's MSS is 1460 bytes, how long would it take for the DASH server's TCP layer cwnd to reach and exceed BDP during Slow Start from an initial window (IW) of 10 MSS?
• If the TCP stack at the DASH server begins in Slow Start with IW=10 MSS for the transmission of every chunk, how long does it take to transmit HTTP-encapsulated 4-second chunks encoded at the three representation rates? (Assume no packet losses or packet reordering along the path.)
4. With the appropriate use of diagrams and clear descriptive text:
• Explain briefly what is meant by tail-drop FIFO queuing.
• Explain what is meant by BufferBloat, and what are the consequences for a VoIP flow's latency and throughput when a VoIP flow and bulk TCP flow share and congest a FIFO bottleneck queue.
• Explain the role and operation of the active queue management (AQM) schemes known as PIE and FQ-CoDel. For both PIE and FQ-CoDel separately, explain how they impact on latency and bandwidth experienced by individual flows when multiple flows compete at a bottleneck router.
5. With the appropriate use of diagrams and clear descriptive text:
• Consider a home router that is forwarding packets out of the house over a link that carries 12Mbit/sec downstream and 1Mbit/sec upstream at the IP layer. How many two-way RTP-based VoIP calls, using standard G.711a codecs, could be active at the same time between the house and outside world? (Assume standard VoIP encapsulation)
Questions involving your tracefile:
5. The tracefile was created by monitoring packets at a certain point between a SIP Client and a SIP Proxy.
• What is the IP address of the SIP Proxy?
• What is the IP address of the SIP Client?
• How many hops away from the SIP Proxy was this tracefile captured?
• How many hops away from the SIP Client was this tracefile captured?
• Identify and explain each distinct Ethernet MAC addresses visible in this tracefile
6. A number of phone calls were made while capturing packets.
• How many phone calls are captured in this tracefile?
• How many SIP control packets were captured?
• How many RTP packets were captured?
• Describe the direction and time/duration of each phone call (to the nearest second).
• Who was called, and who was being called, in each case? (State their SIP identities)
7. Each phone call consists of two-way RTP media streams. With appropriate graphs, diagrams and clear descriptive text:
• As cumulative distribution functions, show the distribution of inter-packet arrival times of RTP frames in the caller to callee direction for each call. Describe possible reasons for the shape(s) you observe.
• As cumulative distribution functions, show the distribution of inter-packet arrival times of RTP frames in the callee to caller direction for each call. Describe possible reasons for the shape(s) you observe.
• Plot the instantaneous IP-level bitrate and packetrate over time for each RTP flow in each direction and for each call. Describe possible reasons for the shape(s) you observe.
8. What is the secret message encoded in your Ethernet tracefile? How did you find it?
Question ONLY for TNE80003 students:
9. NBN Co currently offers two analog voice ports (called UNI-V) on its Customer Premised Equipment (CPE) for fibre based services.
• What signaling protocol does the UNI-V use between CPE and access provider?
• What codec does the UNI-V use?
• What is the packetisation rate for the codec?
• What method(s) does NBNCo use to ensure QoS for UNI-V traffic traversing its network?
• What levels of QoS are there?