adaptive delivery of live video stream: infrastructure cost vs. qoe

76
Adaptive Delivery of Live Video Streams Infrastructure Cost vs. QoE Gwendal Simon

Upload: gwendal-simon

Post on 17-Jul-2015

416 views

Category:

Technology


0 download

TRANSCRIPT

Adaptive Delivery ofLive Video StreamsInfrastructure Cost vs. QoEGwendal Simon

Context andMotivations

2 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Context

Target applications : live streaming platforms

Target network : CDN-based delivery architecture

Content Provider

encodersingestserver

CDN

originserver

edgeservers Clients

3 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Context

Target applications : live streaming platforms

Target network : CDN-based delivery architecture

Content Provider

encodersingestserver

CDN

originserver

edgeservers Clients

3 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Problem

one ingested stream = a dozen of representations

= tens of Mbps to deliverto each edge server

thousands of streams = a huge stress on theCDN infrastructure

4 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Problem

one ingested stream = a dozen of representations= tens of Mbps to deliver

to each edge server

thousands of streams = a huge stress on theCDN infrastructure

4 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Problem

one ingested stream = a dozen of representations= tens of Mbps to deliver

to each edge server

thousands of streams = a huge stress on theCDN infrastructure

4 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Motivations

Our goal : find a better trade-off betweenThe Quality of Experience (QoE) at the user sideThe CDN infrastructure cost

Our approach : optimization, bounds, heuristics

5 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Motivations

Our goal : find a better trade-off betweenThe Quality of Experience (QoE) at the user sideThe CDN infrastructure cost

Our approach : optimization, bounds, heuristics

5 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Contributions

Content Provider

encodersingestserver

CDN

originserver

edgeservers Clients

Contribution 2Optimizing the delivery in under-provisioned network

CDN

originserver

edgeservers

Contribution 1Optimizing transcoding in the ingest server

ingestserver

originserver

6 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Contributions

Content Provider

encodersingestserver

CDN

originserver

edgeservers Clients

Contribution 2Optimizing the delivery in under-provisioned network

CDN

originserver

edgeservers

Contribution 1Optimizing transcoding in the ingest server

ingestserver

originserver

6 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Contributions

Content Provider

encodersingestserver

CDN

originserver

edgeservers Clients

Contribution 2Optimizing the delivery in under-provisioned network

CDN

originserver

edgeservers

Contribution 1Optimizing transcoding in the ingest server

ingestserver

originserver

6 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Case study : Twitch

7 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

What is a broadcaster

Anybody who screencasts, encodes, and uploads

online online

nb. of viewers

timet1 t ′

1 t2 t ′2

session 1 session 2

8 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Data retrieval

Three months in 2014 : from Jan. 6th to Mar. 6th

time

allchannelsof one

snapshot

9 :00 9 :05 9 :10 9 :15 9 :20every five minutes → one snapshot

Dataset available : http ://dash.ipv6.enstb.fr/dataset/twitch/

9 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

How many online broadcasters

0 10 20 30 40 50 60 70 80 900

2K

4K

6K

8K

10K

Days

Nb.

ofon

linechannels

min max

10 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

How much bandwidth is needed

0 10 20 30 40 50 60 70 80 900

1

2

Days

Bandw

idth

(Tbp

s)min max

11 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Characteristics of the ingested videos

240p 360p 480p 720p 1080p0

0.2

0.4

0.6

Representation

Sessions

ratio

0.1 1 100

0.25

0.5

0.75

1

Video bit-rate (Mbps)CD

Fof

thesessions

720p

12 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Characteristics of the ingested videos

240p 360p 480p 720p 1080p0

0.2

0.4

0.6

Representation

Sessions

ratio

0.1 1 100

0.25

0.5

0.75

1

Video bit-rate (Mbps)CD

Fof

thesessions

720p 480p 1080p

12 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

OptimizingTranscoding

13 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Implementing adaptive streaming

ingestserver

deliverynetworkbroadcaster

How many representations ?What bit-rates ?What resolutions ?

broadcaster

broadcaster

Popular user-generated platformsneed transcoding-as-a-service offers

broadcaster

broadcasterbroadcaster

Transcode according to :stream type

stream popularitystream resolution and bit-rate

14 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Implementing adaptive streaming

ingestserver

deliverynetworkbroadcaster

How many representations ?What bit-rates ?What resolutions ?

broadcaster

broadcaster

Popular user-generated platformsneed transcoding-as-a-service offers

broadcaster

broadcasterbroadcaster

Transcode according to :stream type

stream popularitystream resolution and bit-rate

14 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Implementing adaptive streaming

ingestserver

deliverynetwork

broadcaster

How many representations ?What bit-rates ?What resolutions ?

broadcaster

broadcaster

Popular user-generated platformsneed transcoding-as-a-service offers

broadcaster

broadcasterbroadcaster

Transcode according to :stream type

stream popularitystream resolution and bit-rate

14 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Implementing adaptive streaming

ingestserver

deliverynetwork

broadcaster

How many representations ?What bit-rates ?What resolutions ?

broadcaster

broadcaster

Popular user-generated platformsneed transcoding-as-a-service offers

broadcaster

broadcasterbroadcaster

Transcode according to :stream type

stream popularitystream resolution and bit-rate

14 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A dataset for transcoding ingested videos

originalvideoyuv

broadcaster-prepared video

1080p2.75 Mbps

encoding

cloud-transcoded repr.

360p1.6 Mbps

transcoding

measure CPU cycles

reference video360p

3 Mbps

estimatin

g

QoE

500 1000 1500 2000 2500303234363840

Rate (in kbps)

PSNR(in

dB)

224p 360p 720p 1080p

500 1000 1500 2000 25000.60.81

1.21.41.6

Rate (in kbps)

CPU

(inGHz)

224p 360p 720p 1080p

15 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A dataset for transcoding ingested videos

originalvideoyuv

broadcaster-prepared video

1080p2.75 Mbps

encoding cloud-transcoded repr.

360p1.6 Mbps

transcoding

measure CPU cycles

reference video360p

3 Mbps

estimatin

g

QoE

500 1000 1500 2000 2500303234363840

Rate (in kbps)

PSNR(in

dB)

224p 360p 720p 1080p

500 1000 1500 2000 25000.60.81

1.21.41.6

Rate (in kbps)

CPU

(inGHz)

224p 360p 720p 1080p

15 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A dataset for transcoding ingested videos

originalvideoyuv

broadcaster-prepared video

1080p2.75 Mbps

encoding cloud-transcoded repr.

360p1.6 Mbps

transcoding

measure CPU cycles

reference video360p

3 Mbps

estimatin

g

QoE

500 1000 1500 2000 2500303234363840

Rate (in kbps)

PSNR(in

dB)

224p 360p 720p 1080p

500 1000 1500 2000 25000.60.81

1.21.41.6

Rate (in kbps)

CPU

(inGHz)

224p 360p 720p 1080p

15 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Problem Formulation

a type of videoa resolutiona bit-rate

Ingested streams

a max download capacitya max display sizea stream to watch

End-users

limited CPU resourceslimited delivery capacity

Constraints

decide for each stream :nb of representations

their resolutionstheir bit-rates

16 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Problem Formulation

a type of videoa resolutiona bit-rate

Ingested streams

a max download capacitya max display sizea stream to watch

End-users

limited CPU resourceslimited delivery capacity

Constraints

decide for each stream :nb of representations

their resolutionstheir bit-rates

16 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Problem Formulation

a type of videoa resolutiona bit-rate

Ingested streams

a max download capacitya max display sizea stream to watch

End-users

limited CPU resourceslimited delivery capacity

Constraints

decide for each stream :nb of representations

their resolutionstheir bit-rates

16 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Problem Formulation

a type of videoa resolutiona bit-rate

Ingested streams

a max download capacitya max display sizea stream to watch

End-users

limited CPU resourceslimited delivery capacity

Constraints

decide for each stream :nb of representations

their resolutionstheir bit-rates

16 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Integer Linear Program (ILP)

max{ααα,βββ}

∑i∈I

∑o∈O

∑u∈U

fiou · αiou (1a)

s.t. i ∈ I, o ∈ O, u ∈ U (1b)∑o∈O

αiou ≤ diu , i ∈ I, u ∈ U (1c)

∑i∈I

∑o∈O

(ro − cu

)· αiou ≤ 0, u ∈ U (1d)

∑i∈I

∑o∈O

∑u∈U

αiou ≥ R · N, (1e)

βiom ≤

{1, if (vi = vo & si = so & bi > bo) ‖

(vi = vo & si > so & bi ≥ bo)0, otherwise

i ∈ I, o ∈ O,m ∈ M (1f)∑i∈I

∑o∈O

pio · βiom ≤ Pm, m ∈ M (1g)

... (1h)

17 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

How far from the optimal ?

1. Define simulation settings using three datasets :A collection of broadcasters’ streamsA population of end-viewersA transcoding dataset of QoE and CPU measures

2. Compute optimal representations with ILP

3. Compare ILP optimum with current solutions :Full-cover strategies : the smallest rate per resolutionCloud-transcoding providers sets (e.g. Zencoder)

18 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

How far from the optimal ?

1. Define simulation settings using three datasets :A collection of broadcasters’ streamsA population of end-viewersA transcoding dataset of QoE and CPU measures

2. Compute optimal representations with ILP

3. Compare ILP optimum with current solutions :Full-cover strategies : the smallest rate per resolutionCloud-transcoding providers sets (e.g. Zencoder)

18 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

How far from the optimal ?

1. Define simulation settings using three datasets :A collection of broadcasters’ streamsA population of end-viewersA transcoding dataset of QoE and CPU measures

2. Compute optimal representations with ILP

3. Compare ILP optimum with current solutions :Full-cover strategies : the smallest rate per resolutionCloud-transcoding providers sets (e.g. Zencoder)

18 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

How far from the optimal ?

The 50 most popular channels are transcoded

20 40 60 80 100

29

30

31

32

33

number of machines

Avg.

PSNR

(indB

)

19 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

How far from the optimal ?

The 50 most popular channels are transcoded

20 40 60 80 100

29

30

31

32

33

Full-Cover

number of machines

Avg.

PSNR

(indB

)

19 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

How far from the optimal ?

The 50 most popular channels are transcoded

20 40 60 80 100

29

30

31

32

33

Full-Cover Zencoder

number of machines

Avg.

PSNR

(indB

)

19 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Heuristic in a nutshellInput : streams + total CPU

Process each stream

20 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Heuristic in a nutshellInput : streams + total CPU

Process each stream

Fix a stream CPU based on optimum :1) stream popularity2) stream CPU < 10GHz3) video type4) resolution

20 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Heuristic in a nutshellInput : streams + total CPU

Process each stream

Fix a stream CPU based on optimum :1) stream popularity2) stream CPU < 10GHz3) video type4) resolution

Find stream representations w.r.t CPU

20 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Heuristic in a nutshellInput : streams + total CPU

Process each stream

Fix a stream CPU based on optimum :1) stream popularity2) stream CPU < 10GHz3) video type4) resolution

Find stream representations w.r.t CPU

Any pending stream ? Next stream

Output : Representations

yesno

20 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Algorithms performance

0 2 000 4 000 6 000 8 000 10 000 12 000 14 000 16 000 18 000 20 00031

31.5

32

32.5

33

Total CPU (GHz)

Avg.

PSNR(dB)

Full-Cover

21 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Algorithms performance

0 2 000 4 000 6 000 8 000 10 000 12 000 14 000 16 000 18 000 20 00031

31.5

32

32.5

33

Total CPU (GHz)

Avg.

PSNR(dB)

Full-Cover Zencoder

21 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Algorithms performance

0 2 000 4 000 6 000 8 000 10 000 12 000 14 000 16 000 18 000 20 00031

31.5

32

32.5

33

Total CPU (GHz)

Avg.

PSNR(dB)

Full-Cover Zencoder Heuristic

21 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

More details in

L. Toni, R. Aparicio-Pardo, K. Pires, A. Blanc, G. Simon and P. Frossard.Optimal Selection of Adaptive Streaming Representations

ACM Transactions on Multimedia Computing, Communications and Applications, 2015.

R. Aparicio, K. Pires, A. Blanc and G. Simon.Transcoding Live Video Streams at a Massive Scale in the Cloud

in Proc. of ACM MMSys, 2015.

22 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Optimizing Delivery

23 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Model

ISP 1 ISP 2 ISP 3

origin servers

reflectors

edge servers

Assumption : The upload capacity of theequipments is the main resource to save

24 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Model

ISP 1 ISP 2 ISP 3

origin servers

reflectors

edge servers

Assumption : The upload capacity of theequipments is the main resource to save

24 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Our idea in a nutshell

Do not send all representations to edge servers

CDNcoordinator

1

edge serversreport to the

CDN coordinatorabout the

requests theygot during thelast period

12

The CDN coor-dinator decidesutility scoresfor all repre-sentations andall edge servers

23 sources deliverrepresentations sothat the overallutility score is

maximal, subjectto network

under-provisioning

3

25 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Our idea in a nutshell

Do not send all representations to edge servers

CDNcoordinator

1

edge serversreport to the

CDN coordinatorabout the

requests theygot during thelast period

12

The CDN coor-dinator decidesutility scoresfor all repre-sentations andall edge servers

23 sources deliverrepresentations sothat the overallutility score is

maximal, subjectto network

under-provisioning

3

25 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Our idea in a nutshell

Do not send all representations to edge servers

CDNcoordinator

1

edge serversreport to the

CDN coordinatorabout the

requests theygot during thelast period

1

2The CDN coor-dinator decidesutility scoresfor all repre-sentations andall edge servers

23 sources deliverrepresentations sothat the overallutility score is

maximal, subjectto network

under-provisioning

3

25 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Our idea in a nutshell

Do not send all representations to edge servers

CDNcoordinator

1

edge serversreport to the

CDN coordinatorabout the

requests theygot during thelast period

1

2The CDN coor-dinator decidesutility scoresfor all repre-sentations andall edge servers

2

3 sources deliverrepresentations sothat the overallutility score is

maximal, subjectto network

under-provisioning

3

25 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Our idea in a nutshell

Do not send all representations to edge servers

CDNcoordinator

1

edge serversreport to the

CDN coordinatorabout the

requests theygot during thelast period

12

The CDN coor-dinator decidesutility scoresfor all repre-sentations andall edge servers

2

3 sources deliverrepresentations sothat the overallutility score is

maximal, subjectto network

under-provisioning

3

25 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A simple example of delivery

46

45

mobile ftth other mobile

reflectors with uploadcapacity expressed in Mbps

46

45

mobile ftth other mobile

One streamTwo representations

low-quality 1 Mbpshigh-quality 3 Mbps

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

23

11

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

∑e ulow + uhigh = 23

Can we do better ?

26 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A simple example of delivery

46

45

mobile ftth other mobile

reflectors with uploadcapacity expressed in Mbps

46

45

mobile ftth other mobile

One streamTwo representations

low-quality 1 Mbpshigh-quality 3 Mbps

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

23

11

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

∑e ulow + uhigh = 23

Can we do better ?

26 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A simple example of delivery

46

45

mobile ftth other mobile

reflectors with uploadcapacity expressed in Mbps

46

45

mobile ftth other mobile

One streamTwo representations

low-quality 1 Mbpshigh-quality 3 Mbps

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

23

11

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

∑e ulow + uhigh = 23

Can we do better ?

26 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A simple example of delivery

46

45

mobile ftth other mobile

reflectors with uploadcapacity expressed in Mbps

46

45

mobile ftth other mobile

One streamTwo representations

low-quality 1 Mbpshigh-quality 3 Mbps

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

23

11

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

∑e ulow + uhigh = 23

Can we do better ?

26 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A simple example of delivery

46

45

mobile ftth other mobile

reflectors with uploadcapacity expressed in Mbps

46

45

mobile ftth other mobile

One streamTwo representations

low-quality 1 Mbpshigh-quality 3 Mbps

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

23

11

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

∑e ulow + uhigh = 23

Can we do better ?

26 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A simple example of delivery

46

45

mobile ftth other mobile

reflectors with uploadcapacity expressed in Mbps

46

45

mobile ftth other mobile

One streamTwo representations

low-quality 1 Mbpshigh-quality 3 Mbps

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

23

11

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

∑e ulow + uhigh = 23

Can we do better ?

26 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A simple example of delivery

46

45

mobile ftth other mobile

reflectors with uploadcapacity expressed in Mbps

46

45

mobile ftth other mobile

One streamTwo representations

low-quality 1 Mbpshigh-quality 3 Mbps

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

23

11

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

∑e ulow + uhigh = 23

Can we do better ?

26 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A simple example of delivery

46

45

mobile ftth other mobile

reflectors with uploadcapacity expressed in Mbps

46

45

mobile ftth other mobile

One streamTwo representations

low-quality 1 Mbpshigh-quality 3 Mbps

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

46

45

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

otherulow = 4uhigh = 4

mobileulow = 6uhigh = 1

26

14

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

23

11

mobileulow = 6uhigh = 1

ftthulow = 3uhigh = 9

mobileulow = 6uhigh = 1

otherulow = 4uhigh = 4

∑e ulow + uhigh = 23

Can we do better ?

26 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

What we have done

J. Liu, G. Simon, G. Texier, and C. Rosenberg.User-centric discretized delivery of rate-adaptive livestreams in underprovisioned CDN networksIEEE Journal in Selected Areas in Communications, 2014.

A linear program to compute the optimal deliveryOnly on small-scale network and few movies

Some fast optimal algorithms for special casesA fast heuristic with near-optimal performances

27 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

What we have done

J. Liu, G. Simon, G. Texier, and C. Rosenberg.User-centric discretized delivery of rate-adaptive livestreams in underprovisioned CDN networksIEEE Journal in Selected Areas in Communications, 2014.

A linear program to compute the optimal deliveryOnly on small-scale network and few movies

Some fast optimal algorithms for special casesA fast heuristic with near-optimal performances

27 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

What we have done

J. Liu, G. Simon, G. Texier, and C. Rosenberg.User-centric discretized delivery of rate-adaptive livestreams in underprovisioned CDN networksIEEE Journal in Selected Areas in Communications, 2014.

A linear program to compute the optimal deliveryOnly on small-scale network and few movies

Some fast optimal algorithms for special cases

A fast heuristic with near-optimal performances

27 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

What we have done

J. Liu, G. Simon, G. Texier, and C. Rosenberg.User-centric discretized delivery of rate-adaptive livestreams in underprovisioned CDN networksIEEE Journal in Selected Areas in Communications, 2014.

A linear program to compute the optimal deliveryOnly on small-scale network and few movies

Some fast optimal algorithms for special casesA fast heuristic with near-optimal performances

27 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A glimpse of simulation results

provisioning only 15 of a full delivery

0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

User satisfaction

CDF

ofus

ers

28 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A glimpse of simulation results

provisioning only 15 of a full delivery

0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

User satisfaction

CDF

ofus

ers

A naive approach

28 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A glimpse of simulation results

provisioning only 15 of a full delivery

two thirds of users donot get the best repr.

a fifth of usersexperience qualityhalf the optimal

0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

User satisfaction

CDF

ofus

ers

A naive approach

28 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A glimpse of simulation results

provisioning only 15 of a full delivery

0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

User satisfaction

CDF

ofus

ers

A naive approachOur optimal solution

28 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

A glimpse of simulation results

provisioning only 15 of a full delivery

room for improvement

0 0.2 0.4 0.6 0.8 10

0.2

0.4

0.6

0.8

1

User satisfaction

CDF

ofus

ers

A naive approachOur optimal solution

28 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Conclusion

29 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Takeaway

Adaptive streaming can be harmful for infrastructures

Current implementations are far from optimal

Smarter solutions exist

Significant improvements can be obtained

30 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Takeaway

Adaptive streaming can be harmful for infrastructures

Current implementations are far from optimal

Smarter solutions exist

Significant improvements can be obtained

30 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Takeaway

Adaptive streaming can be harmful for infrastructures

Current implementations are far from optimal

Smarter solutions exist

Significant improvements can be obtained

30 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams

Takeaway

Adaptive streaming can be harmful for infrastructures

Current implementations are far from optimal

Smarter solutions exist

Significant improvements can be obtained

30 / 30 Gwendal Simon Adaptive Delivery of Live Video Streams