Quality-adaptive Prefetching for Interactive Branched Video using - - PowerPoint PPT Presentation
Quality-adaptive Prefetching for Interactive Branched Video using HTTP-based Adaptive Streaming Vengatanathan Krishnamoorthi 1 , Niklas Carlsson 1 , Derek Eager 2 , Anirban Mahanti 3 , Nahid Shahmehri 1 1 Linkping university, Sweden 2 University
1 Linköping university, Sweden 2 University of Saskatchewan, Canada 3 NICTA, Australia
Vengatanathan Krishnamoorthi1, Niklas Carlsson1, Derek Eager2, Anirban Mahanti3, Nahid Shahmehri1
We have all seen a movie that (in our taste) is...
2
We have all seen a movie that (in our taste) is...
3
We have all seen a movie that (in our taste) is...
4
We have all seen a movie that (in our taste) is...
5
We have all seen a movie that (in our taste) is...
6
… or where we may have wanted our favorite character to make a different choice...
We have all seen a movie that (in our taste) is...
7
… or where we may have wanted our favorite character to make a different choice...
We have all seen a movie that (in our taste) is...
8
… or where we may have wanted our favorite character to make a different choice...
We have all seen a movie that (in our taste) is...
9
… or where we may have wanted our favorite character to make a different choice...
What if we can personalize the storyline based on the users preferences or path choices?
10
What if we can personalize the storyline based on the users preferences or path choices?
11
What if we can personalize the storyline based on the users preferences or path choices?
12
What if we can personalize the storyline based on the users preferences or path choices?
13
What if we can personalize the storyline based on the users preferences or path choices?
14
… already many examples how creative content creators provide interactive experiences and story lines …
What if we can personalize the storyline based on the users preferences or path choices?
15
YouTube Interlude
… already many examples how creative content creators provide interactive experiences and story lines …
… and even books!
What if we can personalize the storyline based on the users preferences or path choices?
16
YouTube Interlude
… already many examples how creative content creators provide interactive experiences and story lines …
… and even books!
What if we can personalize the storyline based on the users preferences or path choices?
17
YouTube Interlude
… already many examples how creative content creators provide interactive experiences and story lines …
… and even books!
What if we can personalize the storyline based on the users preferences or path choices?
18
YouTube Interlude
… already many examples how creative content creators provide interactive experiences and story lines …
… and even books!
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
satisfaction is greatly influenced by:
– Playback stalls and quality fluctuations
into many sub videos and then link them
– Playback stalls when playing a new video – Non-adaptive playback
– Combine branched video and HAS [Krishnamoorthi et al., ACM CCR 2013]
31
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
32
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
33
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
34
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
35
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
36
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
37
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
38
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
39
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
Chunk1
40
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
Chunk1 Chunk2
41
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
Chunk1 Chunk2 Chunk3
42
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
Chunk1 Chunk2 Chunk4 Chunk3
43
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
Chunk1 Chunk2 Chunk4 Chunk3 Chunk5
44
– Video is split into chunks – Easy firewall traversal and caching – Easy support for interactive VoD
– Multiple encodings of each fragment (defined in manifest file) – Clients adapt quality encoding based on buffer/network conditions
Chunk1 Chunk2 Chunk4 Chunk3 Chunk5
45
– The video can include branch points, with multiple branch choices – User selects which segment to play back next
– Arbitrary sequence of chunks from one or more videos
– Goal: Seamless playback even if user decision at last possible moment
46
– The video can include branch points, with multiple branch choices – User selects which segment to play back next
– Arbitrary sequence of chunks from one or more videos
– Goal: Seamless playback even if user decision at last possible moment
47
– The video can include branch points, with multiple branch choices – User selects which segment to play back next
– Arbitrary sequence of chunks from one or more videos
– Goal: Seamless playback even if user decision at last possible moment
48
– The video can include branch points, with multiple branch choices – User selects which segment to play back next
– Arbitrary sequence of chunks from one or more videos
– Goal: Seamless playback even if user decision at last possible moment
49
– The video can include branch points, with multiple branch choices – User selects which segment to play back next
– Arbitrary sequence of chunks from one or more videos
– Goal: Seamless playback even if user decision at last possible moment
50
– The video can include branch points, with multiple branch choices – User selects which segment to play back next
– Arbitrary sequence of chunks from one or more videos
– Goal: Seamless playback even if user decision at last possible moment
51
– The video can include branch points, with multiple branch choices – User selects which segment to play back next
– Arbitrary sequence of chunks from one or more videos
– Goal: Seamless playback even if user decision at last possible moment
52
– The video can include branch points, with multiple branch choices – User selects which segment to play back next
– Arbitrary sequence of chunks from one or more videos
– Goal: Seamless playback even if user decision at last possible moment
53
– The video can include branch points, with multiple branch choices – User selects which segment to play back next
– Arbitrary sequence of chunks from one or more videos
– Goal: Seamless playback even if user decision at last possible moment
54
– The video can include branch points, with multiple branch choices – User selects which segment to play back next
– Arbitrary sequence of chunks from one or more videos
– Goal: Seamless playback even if user decision at last possible moment
55
– The video can include branch points, with multiple branch choices – User selects which segment to play back next
– Arbitrary sequence of chunks from one or more videos
– Goal: Seamless playback even if user decision at last possible moment
the prefetching problem as an optimization problem
– Maximizes expected playback quality while avoiding stalls
determine:
1. When different chunks should be downloaded 2. What quality level should be selected for each of these chunks 3. How to manage playback buffers and (multiple) TCP connections such as to ensure smooth playback experience without excessive workahead (buffering)
insights into the importance of careful adaptive policies
the prefetching problem as an optimization problem
– Maximizes expected playback quality while avoiding stalls
determine:
1. When different chunks should be downloaded 2. What quality level should be selected for each of these chunks 3. How to manage playback buffers and (multiple) TCP connections such as to ensure smooth playback experience without excessive workahead (buffering)
insights into the importance of careful adaptive policies
the prefetching problem as an optimization problem
– Maximizes expected playback quality while avoiding stalls
determine:
1. When different chunks should be downloaded 2. What quality level should be selected for each of these chunks 3. How to manage playback buffers and (multiple) TCP connections such as to ensure smooth playback experience without excessive workahead (buffering)
insights into the importance of careful adaptive policies
the prefetching problem as an optimization problem
– Maximizes expected playback quality while avoiding stalls
determine:
1. When different chunks should be downloaded 2. What quality level should be selected for each of these chunks 3. How to manage playback buffers and (multiple) TCP connections such as to ensure smooth playback experience without excessive workahead (buffering)
insights into the importance of careful adaptive policies
*Software: http://www.ida.liu.se/~nikca/mm14.html
the prefetching problem as an optimization problem
– Maximizes expected playback quality while avoiding stalls
determine:
1. When different chunks should be downloaded 2. What quality level should be selected for each of these chunks 3. How to manage playback buffers and (multiple) TCP connections such as to ensure smooth playback experience without excessive workahead (buffering)
into the importance of careful adaptive policies
*Software: http://www.ida.liu.se/~nikca/mm14.html
the prefetching problem as an optimization problem
– Maximizes expected playback quality while avoiding stalls
determine:
1. When different chunks should be downloaded 2. What quality level should be selected for each of these chunks 3. How to manage playback buffers and (multiple) TCP connections such as to ensure smooth playback experience without excessive workahead (buffering)
into the importance of careful adaptive policies
*Software: http://www.ida.liu.se/~nikca/mm14.html
the prefetching problem as an optimization problem
– Maximizes expected playback quality while avoiding stalls
determine:
1. When different chunks should be downloaded 2. What quality level should be selected for each of these chunks 3. How to manage playback buffers and (multiple) TCP connections such as to ensure smooth playback experience without excessive workahead (buffering)
into the importance of careful adaptive policies
*Software: http://www.ida.liu.se/~nikca/mm14.html
playback quality
Current segment
Beginning of next segment
Current segment
Current segment
Current segment
Current segment
first chunk next Current segment
first chunk next Current segment
first chunk next Current segment
first chunk next Current segment
extra workahead first chunk next Current segment
extra workahead first chunk next Current segment
extra workahead first chunk next Current segment
extra workahead first chunk next Current segment
first chunk next Current segment
first chunk next Current segment Selected path
Selected path current segment
current segment
current segment
current segment
22
current segment first chunk next
22
current segment first chunk next
– for seamless playback without stalls
Playback schedule
– for seamless playback without stalls
Playback schedule Download schedule
– for seamless playback without stalls – Current segment: e.g., 2 and 3
– for seamless playback without stalls – Current segment: e.g., 2 and 3
Download completion time
– for seamless playback without stalls – Current segment: e.g., 2 and 3
Download completion time
– for seamless playback without stalls – Current segment: e.g., 2 and 3
Download completion times
Playback deadlines Download completion time
– for seamless playback without stalls – Current segment: e.g., 2 and 3
Time of playback deadline
– for seamless playback without stalls – Current segment: e.g., 2 and 3
Time of playback deadline Playback deadlines
Startup delay
– for seamless playback without stalls – Current segment: e.g., 2 and 3
Playback deadlines
Playtime of earlier chunks
– for seamless playback without stalls – Current segment: e.g., 2 and 3
Startup delay Playback deadlines
– for seamless playback without stalls – First chunks next segment: e.g., 4, 7, and 10
– for seamless playback without stalls – First chunks next segment: e.g., 4, 7, and 10
– for seamless playback without stalls – First chunks next segment: e.g., 4, 7, and 10
Download completion times Download completion times
– for seamless playback without stalls – First chunks next segment: e.g., 4, 7, and 10
Download completion times Time at which branch point is reached Playback deadline (shared) for chunks 4, 7, and 10
Download completion times Download completion times
– Schedule new downloads and new TCP connections at the end of a chunk download – Assume that an additional TCP connection will not increase the total download rate – New connections are initiated only if it is not expected to lead to playback deadline violations
– At the end of a chunk download, schedule new downloads and new TCP connections – Assume that an additional TCP connection will not increase the total download rate – New connections are initiated only if it is not expected to lead to playback deadline violations
– At the end of a chunk download, schedule new downloads and new TCP connections – Assume that an additional TCP connection will not increase the total download rate – New connections are initiated only if it is not expected to lead to playback deadline violations
– At the end of a chunk download, schedule new downloads and new TCP connections – Assume that an additional TCP connection will not increase the total download rate – New connections are initiated only if it is not expected to lead to playback deadline violations
– At the end of a chunk download, schedule new downloads and new TCP connections – Assume that an additional TCP connection will not increase the total download rate – New connections are initiated only if it is not expected to lead to playback deadline violations
– Decide number of chunks to download next – Decide quality level of chunks – Maximize expected weighted playback
– Differ in number of candidate schedules and how aggressive they are (in choosing qualities)
– Decide number of chunks to download next – Decide quality level of chunks – Maximize expected weighted playback
– Differ in number of candidate schedules and how aggressive they are (in choosing qualities)
– Decide number of chunks to download next – Decide quality level of chunks – Maximize expected weighted playback
– Differ in number of candidate schedules and how aggressive they are (in choosing qualities)
– Decide number of chunks to download next – Decide quality level of chunks – Maximize expected weighted playback
– Differ in number of candidate schedules and how aggressive they are (in choosing qualities)
– Decide number of chunks to download next – Decide quality level of chunks – Maximize expected weighted playback
– Differ in number of candidate schedules and how aggressive they are (in choosing qualities)
– Decide number of chunks to download next – Decide quality level of chunks – Maximize expected weighted playback
– Policies differ in number of candidate schedules and how aggressive they are (in choosing qualities)
Policy Connections Schedules considered Objective All schedules 1≤ci≤Cmax
QM, where M=ne+|ξb|-m
increasing quality 1≤ci≤Cmax Optimized maintainable quality 1≤ci≤Cmax Q
M+Q-1 Q-1
i=1
ne
∑ qili + ∑ qili
i=ne+1 ne+|ξb|
– Constraint: Qualities of consecutive chunks are non-increasing
– Constraint: Chosen quality must be sustainable for the remaining chunks
Policy Connections Schedules considered Objective All schedules 1≤ci≤Cmax
QM, where M=ne+|ξb|-m
increasing quality 1≤ci≤Cmax Optimized maintainable quality 1≤ci≤Cmax Q
M+Q-1 Q-1
i=1
ne
∑ qili + ∑ qili
i=ne+1 ne+|ξb|
– Constraint: Qualities of consecutive chunks are non-increasing
– Constraint: Chosen quality must be sustainable for the remaining chunks
Policy Connections Schedules considered Objective All schedules 1≤ci≤Cmax
QM, where M=ne+|ξb|-m
increasing quality 1≤ci≤Cmax Optimized maintainable quality 1≤ci≤Cmax Q
M+Q-1 Q-1
i=1
ne
∑ qili + ∑ qili
i=ne+1 ne+|ξb|
– Constraint: Qualities of consecutive chunks are non-increasing
– Constraint: Chosen quality must be sustainable for the remaining chunks
Policy Connections Schedules considered Objective Single connection 1 Q Greedy bandwidth 1≤ci≤Cmax
i=1
ne
∑ qili + ∑ qili
i=ne+1 ne+|ξb|
∑ qili
i=j j+m
which do not use multiple connections
Policy Connections Schedules considered Objective Single connection 1 Q Greedy bandwidth 1≤ci≤Cmax
i=1
ne
∑ qili + ∑ qili
i=ne+1 ne+|ξb|
∑ qili
i=j j+m
which do not use multiple connections
Policy Connections Schedules considered Objective Single connection 1 Q Greedy bandwidth 1≤ci≤Cmax
i=1
ne
∑ qili + ∑ qili
i=ne+1 ne+|ξb|
∑ qili
i=j j+m
which do not use multiple connections
Worst case scenario
Worst case scenario
– Chunks per segment: 5 – Branches per branch point: 4 – Branch points: 3
Worst case scenario
– Chunks per segment: 5 – Branches per branch point: 4 – Branch points: 3
– Chunks per segment: 5 – Branches per branch point: 4 – Branch points: 3
Segment length
– Chunks per segment: 5 – Branches per branch point: 4 – Branch points: 3
Branch options
– Chunks per segment: 5 – Branches per branch point: 4 – Branch points: 3
Branch points
– Chunks per segment: 5 – Branches per branch point: 4 – Branch points: 3
Branch points
– Stalls at every branch point – Note: High playback rate is misleading on its own
– Stalls at every branch point – Note: High playback rate is misleading on its own
– Stalls at every branch point – Note: High playback rate is misleading on its own
– Much lower stall probability – Tradeoff is somewhat lower playback quality
– Much lower stall probability – Tradeoff is somewhat lower playback rate
– Much lower stall probability – Tradeoff is somewhat lower playback rate
– Much lower stall probability – Tradeoff is somewhat lower playback rate
– Higher playback rate – More stalls
– Higher playback rate – More stalls
– Higher playback rate – More stalls
– Good (slightly higher) playback rate – Much more stalls
– Good (slightly higher) playback rate – Much more stalls
– Good (slightly higher) playback rate – Much more stalls
– Lower playback rate – More stalls
– Lower playback rate – More stalls
– Lower playback rate – More stalls
– Playback rate decrease with RTT – Stall probability increase with RTT
– Playback rate decrease with RTT – Stall probability increase with RTT
– Playback rate decrease with RTT – Stall probability increase with RTT
Segment length
Segment length
Segment length
100%
Segment length
Branch options
Branch options
Branch options
Branch options
TCP connection 1 TCP connection N
competing traffic
– E.g., Single connection policy has much more stalls when competing flows
TCP connection 1 TCP connection N
competing traffic
– E.g., Single connection policy has much more stalls when competing flows
TCP connection 1 TCP connection N
competing traffic
– E.g., Single connection policy has much more stalls when competing flows
TCP connection 1 TCP connection N
competing traffic
– E.g., Single connection policy has much more stalls when competing flows
TCP connection 1 TCP connection N
competing traffic
– E.g., Single connection policy has much more stalls when competing flows
TCP connection 1 TCP connection N
– Start playback
– Suspend download
– Resume download
– Start playback
– Suspend download
– Resume download
buffer > Tmin
– Start playback
– Suspend download
– Resume download
buffer > Tmin
– Start playback
– Suspend download
– Resume download
buffer > Tmax buffer > Tmin
– Start playback
– Suspend download
– Resume download
buffer > Tmax buffer > Tmin
– Start playback
– Suspend download
– Resume download
buffer > Tmax buffer > Tmin
– Start playback
– Suspend download
– Resume download
buffer > Tmax buffer > Tmin buffer > Tmin
– Start playback
– Suspend download
– Resume download
buffer > Tmax buffer > Tmin buffer > Tmin
– Start playback
– Suspend download
– Resume download
buffer > Tmax buffer > Tmin buffer > Tmin
– Start playback
– Suspend download
– Resume download
buffer > Tmax buffer > Tmin buffer > Tmin
Example
Example Tsingle = 8 = 4 # branches = 2
Example Tsingle = 8 = 4 # branches = 2 Tmin = 16 Tmax = 20
Example
achieve seamless streaming without playback interruptions
while ensuring sufficient workahead to avoid stalls
levels and number of parallel connections so as to provide best possible video quality, given current conditions
achieve seamless streaming without playback interruptions
while ensuring sufficient workahead to avoid stalls
levels and number of parallel connections so as to provide best possible video quality, given current conditions
achieve seamless streaming without playback interruptions
while ensuring sufficient workahead to avoid stalls
levels and number of parallel connections so as to provide best possible video quality, given current conditions
achieve seamless streaming without playback interruptions
while ensuring sufficient workahead to avoid stalls
levels and number of parallel connections so as to provide best possible video quality, given current conditions
achieve seamless streaming without playback interruptions
while ensuring sufficient workahead to avoid stalls
levels and number of parallel connections so as to provide best possible video quality, given current conditions
Software: http://www.ida.liu.se/~nikca/mm14.html
www.liu.se
Vengatanathan Krishnamoorthi, Niklas Carlsson, Derek Eager, Anirban Mahanti, Nahid Shahmehri
Software: http://www.ida.liu.se/~nikca/mm14.html