SCAN: A Dynamic, Scalable, and Efficient Content Distribution Network

1
SCAN A Dynamic, Scalable, and Efficient Content
Distribution Network
Yan Chen, Randy H. Katz, John D.
Kubiatowicz yanchen, randy, kubitron_at_CS.Berkeley
.EDU EECS Department UC Berkeley
2
Outlines

• Motivation
• Goal and Challenges
• Previous Work
• SCAN Architecture and Components
• Evaluation Methodology
• Results
• Conclusions

3
Motivation Scenario World Cup 2002
4
Goal and Challenges
Provide content distribution to clients with good
latency and staleness, while retaining efficient
and balanced resource consumption of the
underlying infrastructure

• Dynamic choice of number and location of replicas
  
• Clients QoS constraints latency, staleness
• Servers capacity constraints
• Efficient resource consumption
• Small delay, bandwidth consumption for replica
  update
• Small replica management cost
• Scalability millions of objects, clients and
  servers
• No global network topology knowledge

5
Previous Work

• Replica Placement
• Research communities optimal static replica
  placement
• Assume clients distributions, access patterns
  IP topology
• No consideration for clients QoS or servers
  capacity constraints
• CDN operators un-cooperative, ad hoc placement
• Centralized CDN name server cannot record replica
  locations place many more than necessary (ICNP
  02)
• Update Multicast
• No inter-domain IP multicast
• Most application-level multicast (ALM) unscalable
• Split root as common solution, suffers
  consistency overhead

6
SCAN Scalable Content Access Network
Dynamic replica placement d-tree construction
data source
data plane
Web content server
network plane
DOLR with locality
7
Components of SCAN

• Decentralized Object Location Routing (DOLR)
• Properties needed
• Scalable location with guaranteed success
• Search with locality
• Improve the scalability of d-tree each member
  only maintains states for its parent and direct
  children
• Simultaneous Dynamic Replica Placement and d-tree
  Construction
• Replica search Singular, Localized or Exhaustive
• Replica placement on DOLR path Lazy or Eager

8
Replica Search

• Singular Search

data plane
s
c
network plane
DOLR mesh
9
Replica Search

• Localized search

data plane
client child
s
parent
proxy
sibling
c
server child
DOLR mesh
network plane
10
Replica Placement Eager
data plane
s
proxy
c
network plane
DOLR mesh
DOLR path
11
Replica Placement Lazy
data plane
client child
s
proxy
c
network plane
DOLR mesh
12
Evaluation of Alternatives

• Two dynamic overlay approaches
• Overlay_naïve Singular search Eager placement
• Overlay_smart Localized search Lazy placement
• Compared with static placement IP multicast
• Overlay_static With global overlay topology
• IP_static With global IP topology (ideal)
• Metrics
• Number of replicas deployed, load distribution
• Multicast performance Relative Delay Penalty
  (RDP) and bandwidth consumption
• Tree construction traffic (packets and bandwidth)

13
Methodology

• Network Topology
• 5000-node network with GT-ITM transit-stub model
• SCAN nodes placed randomly or on transit nodes
• NS-like Packet-level Network Simulations
• Workloads
• Synthetic flash crowd all clients access a hot
  object in random order
• Real Web server traces NASA and MSNBC

Web Site Period Duration Requests Clients objects
MSNBC 8/2/1999 1011am 1.6M 140K 4186
NASA 7/1/1995 All day 64K 5177 3258
14
Methodology Sensitivity Analysis

• Various Client/Server Ratio
• Various Server Density
• Various Latency Capacity Constraints
• Various Network Topologies
• Average over 5 topologies with different setup
• All Have Similar Trend of Results
• Overlay_smart has close-to-optimal (IP_static)
  number of replicas, load distribution, multicast
  performance with reasonable amount of tree
  construction traffic

15
Number of Replicas Deployed and Load Distribution

• Overlay_smart uses only 30-60 of replicas than
  overlay_naïve and very close to IP_static
• Overlay_smart has two times better load
  distribution than od_naïve, overlay_static and
  very close to IP_static

16
Multicast Performance

• 85 of overlay_smart Relative Delay Penalty (RDP)
  less than 4
• Bandwidth consumed by overlay_smart is very close
  to IP_static, and is only 1/3 of bandwidth by
  overlay_naive

17
Tree Construction Traffic

• Including join requests, ping messages,
  replica placement and parent/child registration

• Overlay_smart consumes 3 - 4 times of traffic
  than overlay_naïve, and the traffic of
  overlay_naïve is quite close to IP_static
• Far less frequent event than access update
  dissemination

18
Conclusions

• P2P networks can be used to construct CDNs
• SCAN Scalable Content Access Network with good
  QoS, efficiency and load balancing
• Simultaneous dynamic replica placement d-tree
  construction
• Leverage DOLR to improve scalability and locality
• In particular, overlay_smart recommended
• Localized search Lazy placement
• Close to optimal number of replicas, good load
  distribution, low multicast delay and bandwidth
  penalty at the price of reasonable construction
  traffic

19
Results on Web Server Traces

• Limited simulations, most URLs have very few
  requests
• Overlay_smart uses only one third to half
  replicas than overlay_naïve for hot objects

20
SCAN Scalable Content Access Network
Dynamic replica placement d-tree construction
data source
data plane
Web content server
network plane
DOLR with locality
21
Replica Search

• Singular Search

data plane
s
c
network plane
DOLR mesh
22
Replica Search
data plane
client child
s
parent
proxy
sibling
c
server child
network plane
23
Dynamic Replica Placement naïve

• Singular Search

• Eager Placement

data plane
parent candidate
s
proxy
c
network plane
Tapestry mesh
Tapestry overlay path
24
Dynamic Replica Placement smart

• Localized search
• Lazy placement

• Greedy load distribution

data plane
parent candidates
client child
s
parent
proxy
sibling
c
server child
network plane