PeertoPeer Applications Part2: Peer selection and dynamics of a p2p network

1
Peer-to-Peer Applications-Part-2 Peer
selection anddynamics of a p2p network

• CS 7270
• Networked Applications Services
• Lecture-10

2
Reading

• Understanding KaZaA by J. Liang et al.
• Based on 2004 paper
• Measurement-based study
• Focuses on KaZaA (FastTrack architecture), but
  most of the observations are important for
  general p2p unstructured systems

3
Understanding KaZaA

• Jian Liang
• Rakesh Kumar
• Keith Ross
• Polytechnic University
• Brooklyn, N.Y.

4
Internet Traffic
CF CacheLogic
5
KaZaA/FastTrack Operation

• Top file sharing system
• 3 million active nodes
• four clients KaZaA, KaZaA-lite, Grokster and
  iMesh
• Good availability and scalability
• Proprietary protocol signaling traffic encrypted
• in contrast with Gnutella and e-mule

6
Purpose of Measurement Study

• Try to understand highly successful file-sharing
  system
• Overlay topology and dynamics
• Peer selection
• Index management

7
Existing Tools and Projects

• FastTrack encryption algorithm
• available from a Web site http//gift-fasttrack.b
  erlios.de/
• KaZaA Media Desktop (KMD) software architecture
• http//kazaasearch.narod.ru/

8
Big Picture of Overlay

• Two layer hierarchy
• Ordinary Node (ON)
• Super Node (SN)
• SNs are generally more powerful machines (CPU,
  network bw) and they are NOT behind NATs

9
FastTrack architecture

• Each ON has a parent SN node
• For each shared file, ON uploads to parent SN
• Filename, ContentHash, file descriptors
  (metadata)
• Parent SN provides ON with SN refresh list
• Up to 200 alive SNs, then stored at ON cache
• For each SN, the list includes IP address, port
  number, SN workload (defined as ?), freshness,
  and timestamp
• SNs also exchange SN refresh lists
• Each SN maintains local index for all children
  ONs
• Each SN maintains TCP connections with other SNs
• Overlay net
• If an SN cannot answer a query, it forwards query
  to other SN peers
• TTL-limited flooding
• Actual file transfer is directly between peers
  (not through overlay) using HTTP
• All signaling traffic is encrypted

10
Measurement Apparatus

• KaZaA Sniffing Platform
• KaZaA Probing Tool

11
KaZaA Sniffing Platform

• Poly (Ethernet)
• Home (cable modem)

12
KaZaA Probing Tool

• Campus home based probing
• Probe arbitrary SNs
• Retrieve their SN refresh lists
• Obtain workload of probed SN

13
Signaling Protocol
ON-SN session initial (repeat for 5 SNs)
SN-SN session initial
14
TCP Connections Evolution at instrumented SN node
Poly campus 4 6 hour measurement
Cable modem 7-11 hour measurement
15
Some basic calculations

• Estimate total number of SNs, assuming about 3M
  users (typical in 2004)
• About 25000-40000 SNs
• Estimate probability of SN-SN link
• About 0.1

16
Signaling Sessions Lifetime

• Measured over a period of 12 hours
• Avg duration 34mins (ON-SN) and 11mins (SN-SN)
• 30-40 of connections (both types) last for less
  than 30 seconds!
• What causes short-lived ON-SN connections?
• What causes short-lived SN-SN connections?

17
Parent selection

• Recall that ON receives a list of 200 SNs from
  its parent SN
• Then, it can select a new parent
• How would you select the parent SN?

18
SN workload vs of connections
7 - 11 hours TCP connections evolution
7 - 11 hours workload values evolution
19
Peer Selection the workload of the SN clearly
matters
20
Locality in Peer Selection (graphs show
percentage of SNs in the SN list having common
prefix with child ON and parent SN)
21
Peer Selection it appears that RTT also
matters 40 of ON-SN connections have
RTT 22
Index Management 1) No index exchange between
SNs2) SN purges metadata of ON as soon as that
child disconnects from parent3) Highly skewed
contribution of metadata by different peers
23
Firewall evasion NAT circumvention

• Found only 3.6 of SNs use the default 1214 port
  number
• Earlier KaZaA clients used this default port, but
  it was easy for net-admins to block them
• 18,887 SNs (96.3) use non-default and
  dynamically assigned port numbers
• How is this done?
• NAT traversal if peer B is behind NAT, then peer
  A contacts Bs parent, and the latter asks B to
  initiate connection to A

24
Summary of Results

• 20,000 40,000 active supernodes
• Each SN connects to approx. 0.1 of other SNs
• Highly dynamic connections over 35 SN-SN
  durations are less than 30 sec.

25
Summary of results

• Peer selection uses IP prefix match, workload,
  RTT and freshness
• No index exchange between SNs, but query
  forwarding
• Skewed content distribution 20 peers provide
  70 metadata for sharing

26
Design Principles forUnstructured P2P Overlays

• Distributed design
• No infrastructure
• Also helps with legal attacks...
• Exploit heterogeneity
• Hierarchy
• Self organization
• Load balancing based on workload metric
• Explicit locality awareness
• Shuffle connections in core overlay

27
Design Principles forUnstructured P2P Overlays

• Properly designed gossip mechanisms
• peers have a fresh list of SNs
• Firewall circumvention
• dynamic port numbers
• improves availability
• NAT circumvention