minimizing server throughput for low-delay live streaming in content delivery networks
DESCRIPTION
Large-scale live streaming systems can experience bottle- necks within the infrastructure of the underlying Content Delivery Network. In particular, the “equipment bottleneck” occurs when the fan-out of a machine does not enable the concurrent transmission of a stream to multiple other equipments. In this paper, we aim to deliver a live stream to a set of destination nodes with minimum throughput at the source and limited increase of the streaming delay. We leverage on rateless codes and cooperation among destination nodes. With rateless codes, a node is able to decode a video block of k information symbols after receiving slightly more than k encoded symbols. To deliver the encoded symbols, we use multiple trees where inner nodes forward all received symbols. Our goal is to build a diffusion forest that minimizes the transmission rate at the source while guaranteeing on-time delivery and reliability at the nodes. When the network is assumed to be lossless and the constraint on delivery delay is relaxed, we give an algorithm that computes a diffusion forest resulting in the minimum source transmission rate. We also propose an effective heuristic algorithm for the general case where packet loss occurs and the delivery delay is bounded. Simulation results for realistic settings show that with our solution the source requires only slightly more than the video bit rate to reliably feed all nodes.TRANSCRIPT
Minimizing ServerThroughput for Low-DelayLive Streaming in ContentDelivery Networks
F. Zhou, S. Ahmad, E. Buyukkaya,R. Hamzaoui and G. Simon
Live Stream Delivery
Content Provider
encodersingestserver
CDN
originserver
edgeservers Clients
Recent large-scale live video streaming failedSuperbowlKorean Telecom and smart TVs
2 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Live Stream Delivery
Content Provider
encodersingestserver
CDN
originserver
edgeservers Clients
Recent large-scale live video streaming failedSuperbowlKorean Telecom and smart TVs
2 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Motivations
Where is the bottleneck in today’s CDN ?
ISP 1 ISP 2 ISP 3
origin servers
reflectors
edge servers
last-mile ?last-mile ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mileedge servers in ISP → peering link
infrastructure ?
adaptive streaming → last-mileedge servers in ISP → peering link
3 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Motivations
Where is the bottleneck in today’s CDN ?
ISP 1 ISP 2 ISP 3
origin servers
reflectors
edge servers last-mile ?
last-mile ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mileedge servers in ISP → peering link
infrastructure ?
adaptive streaming → last-mileedge servers in ISP → peering link
3 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Motivations
Where is the bottleneck in today’s CDN ?
ISP 1 ISP 2 ISP 3
origin servers
reflectors
edge servers
last-mile ?
last-mile ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mileedge servers in ISP → peering link
infrastructure ?
adaptive streaming → last-mileedge servers in ISP → peering link
3 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Motivations
Where is the bottleneck in today’s CDN ?
ISP 1 ISP 2 ISP 3
origin servers
reflectors
edge servers
last-mile ?last-mile ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mileedge servers in ISP → peering link
infrastructure ?
adaptive streaming → last-mileedge servers in ISP → peering link
3 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Motivations
Where is the bottleneck in today’s CDN ?
ISP 1 ISP 2 ISP 3
origin servers
reflectors
edge servers
last-mile ?last-mile ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mileedge servers in ISP → peering link
infrastructure ?
adaptive streaming → last-mileedge servers in ISP → peering link
3 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Motivations
Where is the bottleneck in today’s CDN ?
ISP 1 ISP 2 ISP 3
origin servers
reflectors
edge servers
last-mile ?last-mile ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mile
peering link ?
adaptive streaming → last-mileedge servers in ISP → peering link
infrastructure ?
adaptive streaming → last-mileedge servers in ISP → peering link
3 / 13 Gwendal Simon Minimizing Server Throughput in CDN
The upload capacity equipment bottleneck
2CPU
NIC
6
4 3 5
2 ⇓ ∞ ⇑
∞ ∞ ∞
4 / 13 Gwendal Simon Minimizing Server Throughput in CDN
The upload capacity equipment bottleneck
2CPU
NIC
6
4 3 5
2 ⇓ ∞ ⇑
∞ ∞ ∞
4 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Our proposalno cooperation
cooperation between nodes
Objectives :minimizing source throughputmaintaining a low transmission delay
5 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Our proposalno cooperation cooperation between nodes
Objectives :minimizing source throughputmaintaining a low transmission delay
5 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Our proposalno cooperation cooperation between nodes
Objectives :minimizing source throughputmaintaining a low transmission delay
5 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Rateless codes
encoder decoder
one chunkk symbols
infinite nb. ofencoded symbols k + ε symbols k decoded symbolsone chunk
k symbolsinfinite nb. of
encoded symbols k + ε symbols k decoded symbolsone chunkk symbols
infinite nb. ofencoded symbols k + ε symbols k decoded symbolsone chunk
k symbolsinfinite nb. of
encoded symbols k + ε symbols k decoded symbolsone chunkk symbols
infinite nb. ofencoded symbols k + ε symbols k decoded symbols
Main advantages :adaptive : no fixed code ratelow complexitysuitable for multi-source systems (Wu’2008)
6 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Rateless codes
encoder decoder
one chunkk symbols
infinite nb. ofencoded symbols k + ε symbols k decoded symbolsone chunk
k symbolsinfinite nb. of
encoded symbols k + ε symbols k decoded symbolsone chunkk symbols
infinite nb. ofencoded symbols k + ε symbols k decoded symbolsone chunk
k symbolsinfinite nb. of
encoded symbols k + ε symbols k decoded symbolsone chunkk symbols
infinite nb. ofencoded symbols k + ε symbols k decoded symbols
Main advantages :adaptive : no fixed code ratelow complexitysuitable for multi-source systems (Wu’2008)
6 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Multi-tree delivery (1/2)
Main objective for the delivery of one video chunk :minimize the number of trees (packets)
Main constraint in the trees :each node should be in K treesupload capacity constraint c on nodes
s
1
2
3
4
s
1
2 3
s
2
1 3
4
s
3
4 1
2
s
4
K = 3, c = {2, 2, 3, 1}
7 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Multi-tree delivery (1/2)
Main objective for the delivery of one video chunk :minimize the number of trees (packets)
Main constraint in the trees :each node should be in K treesupload capacity constraint c on nodes
s
1
2
3
4
s
1
2 3
s
2
1 3
4
s
3
4 1
2
s
4
K = 3, c = {2, 2, 3, 1}
7 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Multi-tree delivery (2/2)
Additional constraints
Do not introduce much delaysum of delays over all paths in any tree below D
Do not introduce much packet lossoverall probability of all paths in any tree below P
8 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Our contributions
Two algorithms :1. without last constraints :
an optimal O(Kn2) algorithm2. with limited tree height :
an efficient O(Kn3) heuristic
Algorithm performances depend on the context :over-provisioned system
“source rate = video rate” is achievableunder-provisioned system
source has to compensate missing resources
9 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Our contributions
Two algorithms :1. without last constraints :
an optimal O(Kn2) algorithm2. with limited tree height :
an efficient O(Kn3) heuristic
Algorithm performances depend on the context :over-provisioned system
“source rate = video rate” is achievableunder-provisioned system
source has to compensate missing resources
9 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Simulations
Video and rateless code settings :H.264 with bitrates from 320 kbps to 3.2 MbpsOne chunk is one GOP : 0.5 secondsOne UDP packet is 1,000 bytesRaptor code with redundancy 5%
Network and node settings :from 5 to 25 nodesupload capacity follows log-normal distribution
mean capacity is {512, 1,024, 2,048} kbpshomogeneous packet loss probability and RTT
10 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Scalability
5 10 15 20 250
1,000
2,000
3,000
4,000
5,000
number of nodes
transmiss
ionrate
(inkbps) 512 kbps
11 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Scalability
5 10 15 20 250
1,000
2,000
3,000
4,000
5,000
number of nodes
transmiss
ionrate
(inkbps) 512 kbps
1024 kbps
11 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Scalability
5 10 15 20 250
1,000
2,000
3,000
4,000
5,000
number of nodes
transmiss
ionrate
(inkbps) 512 kbps
1024 kbps2048 kbps
11 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Decoding lag
5 10 15 20 250
200
400
600
800
1,000
number of nodes
playback
lag(in
ms)
512 kbps1024 kbps2048 kbps
12 / 13 Gwendal Simon Minimizing Server Throughput in CDN
Future works
Real implementation. We currently have :a fully-developed program that just workssome contacts with a small CDN company
Academic work :resource management for multiple flowsmore dynamic algorithms
13 / 13 Gwendal Simon Minimizing Server Throughput in CDN