FINAL - EVERYTHING Flashcards
what is multimedia?
two or more continuous media
what are the three broad categories of audio/video services?
streaming stored audio/video, streaming live audio/video, interactive audio/video
how does a client stream stored audio/video service?
downloads through the internet
how are audio/video services stored?
files are compressed and stored on server
what is digitization?
conversion of data into a stream of numbers
what is compression?
a coding process that will reduce the overall number of bits needed to represent information
what is sampling?
measuring analog signal at regular time intervals
what is it called when you sample amplitude (voltage)?
quantization
what is the nyquist theorum?
the idea that we need to sample the signal 2f times per second given the highest signal frequency ‘f’
what is pulse code modulation?
converting analog audio to compressed binary from sampling, quantization and encoding
what 3 questions do you need to decide when digitizing audio data?
sampling rate? how finely is the data quantized and is it uniform? file format?
what is the difference between predictive encoding and perceptual encoding?
predictive: difference in samples are encoded instead of the actual sampled values (typically used for speech)
perceptual: tries to mimic how people perceive sound (mp3)
what is the difference between predictive encoding and perceptual encoding?
predictive: difference in samples are encoded instead of the actual sampled values (typically used for speech)
perceptual: tries to mimic how people perceive sound (mp3)
what is the value of DC value in jpeg compression?
the average pixel value multiplied by a constant
what are AC values in jpeg compression?
changes in the pixel values - no change means AC = 0
why is jpeg called lossy compression? and what phase is responsible for this?
because some data is irreversibly lost, happens during quantization phase
what do you call a time ordered sequence of frames?
a video
what is spatial compression?
using JPEG or something close to compress each frame of a video
what is temporal compression?
removal of redundant frames
MPEG divides frames into what 3 categories?
i-frames, p-frames and b-frames
what is an i-frame?
independent frame not related to the previous frame or following frame, meant to represent an interval
what is the point of iframes?
they can stop cross origin domain policy, you can also load lots of different resources in an i-frame
what is a p-frame?
predicted frame - related to the previous i-frame or p-frame
what is a b-frame?
bidirectional frame - relates to previous or following frame
does the first approach for streaming stored audio/video involve streaming? whynot?
no because the entire file must be downloaded before the client sees any of it
what is the problem with the 2nd approach (web server + metafile)
both the browser and media player use HTTP services which run over TCP, this is totally fine for retrieving the metafile but TCP is not designed for audio/video file retrieval
why is TCP bad for audio/video files?
it re transmits lost or damaged segments which fucks up your stream
what protocol does a media server use and why?
UDP so lost and damaged segments are not retransmit
what is RTSP?
an out-of-band control protocol that can control the playing of audio/server
what is RTSP similar to?
second connection in FTP
what does RTSP not do?
define compression schemes, define encapsulation for transmission, restrict transportation method (udp tcp), restrict how media player buffers
what does RTSP do?
allows media player to control transmission, play pause fast forward rewind, messages use a different port number than media stream
what is an out of band protocol?
sends control data from a different connection then main data
what are the similarities between streaming stored audio/video and streaming live audio/video?
both sensitive to delay, both do not accept retransmission
what are the differences between streaming live and stored?
stored communication is unicast and on demand, live is multicast and live
what is the “jitter” phenomenon?
when packets of audio/video are transmitted and have different network delays –> you start playing packet 1 and it finished before packet 2 arrives
what is the purpose of a timestamp in real-time audio/video?
to solve the jitter problem by showing how a packet delay is relative to the previous packet
why do you need a playback buffer in real time interactive audio/video?
to store the data until they are played back based on the what the timestamps dictate
what else is needed other than a timestamp for real time traffic?
packet numbering for ordering purposes
real time traffic needs the support of what due to the fact that it used multimedia?
multi casting
what is “mixing”?
combining data from multiple streams into one
why is TCP bad for real time interactive traffic?
because it re transmits lost or damaged packets
is UDP suitable for real time interactive traffic?
not really because it lacks the ability to timestamp and order packets or mix them
what is RTP for?
real time traffic
what does RTP provide?
timestamping, sequencing, and mixing
what does RTP need to be used with?
UDP
how does encapsulation of packets work with RTP?
first encapsulated by RTP packet and then these packets are encapsulated in a UDP segment
what network layer does RTP operate in?
can be transport layer OR application layer
how is RTP used in the transport layer?
RTP libraries are used with languages in that layer to make RTP into a sublayer of the transport layer
how is RTP used in the application layer?
RTP packets are sent into a UDP socket interface
what is included in a RTP packet header to be encapsulated by a UDP segment?
data encoding, sequence number, timestamp
what does it mean if the P. and X. header fields are set in an RTP packet?
P - there is padding
X - extra extension header
what is the point of the CC and M field in RTP packets?
CC - how many contributors (mixing)
M - end of data
how many bits is the PT header field in RTP and what is it for?
8 bits - defines payload type
for the following RTP header fields, how many bits and how is it set?
- sequence number
- timestamp
- ssrc
- csrc
sequence number - 16 bits, first is random then incremented by 1 each time
timestamp - 32 bits, first is random then sum of past timestamps plus clock time for first bye produced
ssrc - 32 bits, defines identity of source (random # chosen by source)
csrc - 32 bits, defines # of sources (max 15)
what is RTCP
control protocol for RTP
when is RTCP used
when networking applications multicast to multiple receivers from one or more senders
how are RTP and RTCP packets distinguished from each other?
distinct port numbers
what do RTCP packets include?
sender and receiver ports
what are in RTCP sender report packets and what are they used for?
timestamp, allows receiver to synchronize RTP messages –> think audio and video synchronization
what are RTCP receiver reception packets for and what do they do?
passive participants, informs sender and other receivers of service quality
what are in RTCP source description packets?
email address, name of sender and application that generates RTP stream
what is the point of RTCP bye message?
shut down stream
what does an RTCP application specific message allow?
new message type definitions –> for when packet wants to use new applications
what is the restriction on RTP port numbers?
they must be an even number
where must the UDP port be that corresponds to an RTP packet?
immediately after the RTP port
what are the flow characteristics?
reliability, delay, jitter and bandwidth
what is the consequence of low reliability?
losing a packet or ack
what is responsible for lost packets in multimedia network apps?
network loss (congestion), end to end delay loss (packets arrive too late)
for multimedia application what range of delay makes a packet effectively useless?
100ms to 5 seconds
what causes jitter?
differing delay time between successive packets
what determines if jitter is high or low?
the difference between delays –> larger difference means higher jitter
what is packet marking need ed for?
for routers to distinguish between different classes; and treat packets accordingly
what are the tradeoffs when you change the size of the chunks in a group (n) for FEC simple scheme?
increase: less bandwidth waste, higher probability that 2 or more chunks are lost, longer playout delay
decrease: more lost packets can be recovered, more bandwidth waster
what is the disadvantage of interleaving for packet loss?
increases latency (need to sort out all the chunks in each packet)
what are the 3 different types of queuing?
FIFO, priority, weighted fair
In FIFO scheduling what happens if the arrival rate at the queue is higher than the processing rate?
the queue will fill up and new packets discarded
what is the condition of starvation for Priority scheduling?
when low priority queues never have the chance to be processed
how does the system process packet in weighted average scheduling?
round robin fashion with higher priority queues getting more processing relative to their priority weight
what is traffic shaping?
mechanism to control the amount and the rate of the traffic sent to the network
on what scheduling queue is the leaky bucket approach implemented on?
FIFO
in a token bucket approach if n = 100 and the host is idle for 100 tickets how many tokens are collected?
10000
what is the problem with the leaky bucket approach to traffic shaping
it limits the output rate to the average of bursty traffic
what is the difference between an access link and a trunk link?
access links are assigned to one specific vlan whereas trunk links carry multiple vlans
what are some reasons for grouping devices into vlans?
service need, protocol, physical proximity
each switch port can be assinged to _______ vlan(s), each switch port can be assigned to _____ vlan(s)
multiple, only one
where are dynamically assigned Vlan ports stored?
in virtual membership policy servers
what is the area between switches called?
switch fabric
what is the area between switches called?
switch fabric
what is frame tagging in Vlans?
when the switch appended a Vlan ID to each frame
which switch appends the Vlan ID? which switch removes it?
first switch in the path, last switch in the path
what happens is a tag is not removed from a frame before it is sent to non-vlan-capable devices?
the device won’t be able to understand the frame
what is cisco’s proprietary protocol called for tag formatting?
inter switch link (ISL) protocol
what are the benefits of using switches to create Vlans over using routers to create distinct networks?
easier to administer, less expensive, higher performance (less latency)
what are the disadvantages of using switches to create Vlans over using routers to create distinct networks?
Vlan may be tied to a specific vendor
what are routes needed for despite advances in switch technology?
filter WAN traffic, route traffic between distinct networks, route packets between Vlans
what is trunking?
when you connect two switches together
when is trunking useful?
when trying to configure Vlans that span multiple switches