Optimal Set of Video Representations in Adaptive Streaming

Optimal Set of VideoRepresentations inAdaptive StreamingL. Toni1, R. Aparicio2, G. Simon2,A. Blanc2 and P. Frossard1

1: EPFL2: Telecom Bretagne

Context : Video Delivery Chain

Content provider

ingestserver

CDN

originserver

edgeservers

end-users

Adaptive Streaming

1 video stream = k representations

many works on adaptive video deliverymany works on client adaptationwhat about video encodingfor adaptive streaming ?

2 / 17 Gwendal Simon Optimal Video Representations in DASH


Content provider

ingestserver

CDN

originserver

edgeservers

end-users

Adaptive Streaming


many works on adaptive video deliverymany works on client adaptationwhat about video encodingfor adaptive streaming ?



Content provider

ingestserver

CDN

originserver

edgeservers

end-users

Adaptive Streaming


many works on adaptive video delivery

many works on client adaptationwhat about video encodingfor adaptive streaming ?



Content provider

ingestserver

CDN

originserver

edgeservers

end-users

Adaptive Streaming


many works on adaptive video delivery

many works on client adaptation

what about video encodingfor adaptive streaming ?



Content provider

ingestserver

CDN

originserver

edgeservers

end-users

Adaptive Streaming


many works on adaptive video deliverymany works on client adaptation

what about video encodingfor adaptive streaming ?


Encoding a set of representations

For a content provider, what does it mean ?

For each video in the catalog, deciding :How many representationsWhat resolutionsWhat bit-rates

Today :Recommendations from vendors (e.g. Apple)Self-tuned by content providers (e.g. Netflix)












Our objective

Finding the best set of representations

for each video in a catalog

maximizing QoE of clients

with limited infrastructure cost


Our objective






Our objective






Our contributions

1. We formulate an optimization problem

2. We study how optimal are recommended sets

3. We identify guidelines for encoding parameters


Our contributions





Our contributions





Model


Definitions

(v , r , s) : representation of video v encoded atresolution s and at bit-rate reach user is associated with one video resolutionfvs(r) ∈ [0, 1] : user satisfaction

normalized QoE per resolution

0 2,000 4,000 6,000 8,0000.6

0.7

0.8

0.9

1

rate (in kbps)

(1-V

QM

)no

rmal

ized

224p360p720p1080p


Definitions



0 2,000 4,000 6,000 8,0000.6

0.7

0.8

0.9

1

rate (in kbps)

(1-V

QM

)no

rmal

ized

224p360p720p1080p


Definitions



0 2,000 4,000 6,000 8,0000.6

0.7

0.8

0.9

1

rate (in kbps)

(1-V

QM

)no

rmal

ized

224p360p720p1080p


Constraints

The global CDN capacity C

The total number of representations K

The fraction of users that must be served P


ILP

max{ααα,βββ}

∑u∈U

∑v∈V

∑r∈R

∑s∈S

fvrs · αuvrs (1a)

s.t. αuvrs ≤ βvrs , u ∈ U, v ∈ V, r ∈ R, s ∈ S (1b)

βvrs ≤∑u∈U

αuvrs , v ∈ V, r ∈ R, s ∈ S (1c)

(bminvs − br ) · βvrs ≤ 0, v ∈ V, r ∈ R, s ∈ S (1d)

(br − bmaxvs ) · βvrs ≤ 0, v ∈ V, r ∈ R, s ∈ S (1e)∑

r∈R

αuvrs ≤

{1, if v = vu

& s = su0, otherwise

u ∈ U, v ∈ V, s ∈ S (1f)

∑v∈V

∑r∈R

∑s∈S

br · αuvrs ≤ cu , u ∈ U (1g)

∑u∈U

∑v∈V

∑r∈R

∑s∈S

br · αuvrs ≤ C, (1h)

... (1i)


Recommended setperformance


Process

1. We define some configurations :A model for QoE based on four test sequencesA population of usersSettings for the content provider

2. We use CPLEX to compute the optimalrepresentations and obtain the best QoE possible

3. We compare with the best QoE achievable with therecommended sets


Process





Process





How far from the optimal ?

For a catalog of four videos in a regular configuration

20 30 40 50 600.8

0.85

0.9

0.95

1

Apple(40 rep)

number of representations K

avg.

user

satisfaction




20 30 40 50 600.8

0.85

0.9

0.95

1

Apple(40 rep)


avg.

user

satisfaction




20 30 40 50 600.8

0.85

0.9

0.95

1

Apple(40 rep)

19 rep.


avg.

user

satisfaction


Guidelines


Process

1. We define multiple configurations with variation ofThe popularity of videosThe characteristics of clientsThe constraints of the service provider

2. We identify trends and derive useful guidelines forthe settings of encoding parameters


Process

1. We define multiple configurations with variation ofThe popularity of videosThe characteristics of clientsThe constraints of the service provider

2. We identify trends and derive useful guidelines forthe settings of encoding parameters


How many representations per video ?

224p 360p 720p 1080p0

2

4

6

Resolution

numbero

frepresentations

cartoon documentary sport movie

nb. of representations depends on the content1

slightly more representationsfor high resolutions

2



224p 360p 720p 1080p0

2

4

6

Resolution

numbero

frepresentations




2



224p 360p 720p 1080p0

2

4

6

Resolution

numbero

frepresentations




2


Conclusion


Conclusion

For more details, read the paper !

Takeaway :Optimal encoding parameters for adaptive streamingRecommendations far from being optimalSome useful guidelines

Future works :How to take into account dynamic configurations ?How to optimize encoding in a data-center ?


Conclusion





Conclusion





Technology

Optimal Set of Video Representations in Adaptive Streaming