A Survey and Comparison of PeertoPeer Overlay Network Schemes

1
A Survey and Comparison of Peer-to-Peer Overlay
Network Schemes
May. 07. 2005 Joonmyung Kang eliot_at_postech.ac.kr
DPNM Lab., Dept. of CSE, POSTECH
2
Presentation Outline

• Introduction
• P2P Review
• P2P Overlay Networks
• Structured vs. Unstructured P2P Overlay Networks
• Structured P2P Overlay Networks
• CAN
• Chord
• Pastry
• Unstructured P2P Overlay Networks
• Gnutella
• Freenet
• KaZaA
• P2P for Internet Telephony
• Conclusions
• References

3
Review of P2P

• What is P2P?
• A class of systems and applications that employ
  distributed resources to perform a critical
  function in a decentralized manner
• Goals
• Resource aggregation - CPU, disk,
• Cost sharing/reduction
• Improved scalability/reliability
• Interoperability - heterogeneous peers
• Increased autonomy at the network edge
• Anonymity/privacy
• Dynamic (join, leave), self organizing
• Ad hoc communication and collaboration

4
P2P Overlay Networks

• In a P2P overlay network, network links are
  application-layer connections implemented on top
  of another underlying network such as an IP
  network
• Thus, nodes in the overlay network do not need to
  worry about traditional IP routing

5
Abstract P2P Overlay Network Architecture
6
Structured vs. Unstructured P2P1

• Structured P2P
• Based on Distribute Hash Table
• The network topology is tightly controlled
• Node can join and leave the system, but the
  topology need to be reconstructed.
• A hash table associates data with keys
• Key is hashed to find bucket in hash table
• Each bucket is expected to hold items/buckets
  items
• In a Distributed Hash Table (DHT), nodes are the
  hash buckets
• Key is hashed to find responsible peer node
• Data and load are balanced across nodes

7
Structured vs. Unstructured P2P2

• Unstructured P2P
• Random overlay networks
• No central index
• Start with nodes that know about peer servers
  and flood along the network
• Features
• Scalability Flooding limited by TTL
• Unreliable/ inefficient searching
• Applications
• Gnutella
• Freenet

Peer 1
Peer 5
Peer 3
Peer 4
Peer 2
Peer 7
Peer 6
8
Goals re-visited

• If present gt find it
• Flooding is not scalable
• Blind search is inefficient

P2P systems
Structured
Unstructured

• Efficient searching
• Proximity
• Locality

• Data availability
• Decentralization
• Scalability
• Load balancing
• Fault tolerance

• Maintenance
• Join/leave
• Repair

Query time, number of messages, network usage,
per node state
9
Structured-DHT-based P2P overlay networks
10
Distributed Hash Tables

• Types of search
• Central index (Napster)
• Distributed index with flooding (Gnutella)
• Distributed index with hashing (Chord)
• Basic operations
• find(key), insert(key, value), delete(key)

11
CANContent Addressable Network

• Each key maps to one point in the d-dimensional
  space
• Each node responsible for all the keys in its
  zone.
• Divide the space into zones.

C
D
E
A
B
12
CAN 2
13
Chord1

• Chord provides fast distributed computation of a
  hash function, mapping keys to nodes responsible
  for them
• Assigns keys to nodes with consistent hashing (
  based on SHA-1 )
• A chord node needs only a small amount of routing
  information about other nodes
• A node resolves the hash function by
  communicating with other nodes
• With high probability, the number of nodes that
  must be contacted to find a successor is an
  N-node network is O(log N)
• Only O(log N) fingers need be stored
• When an Nth node joins or leaves the network,
  only an O(1/N) fraction of the keys are moved

14
Chord2
This figure shows chord identifier circle, which
consists three nodes 0, 1 and 3. Then identifiers
should be ordered in identifier circle modulo 8
(82m, m3). In this example, key 1 is located at
node 0, key 2 at node 1 and key 6 at node 3.
15
Pastry

• Prefix-based
• Route to node with shared prefix (with the key)
  of ID at least one digit more than this node.
• Neighbor set, leaf set and routing table.

d471f1
d467c4
d46a1c
d462ba
d4213f
Route(d46a1c)
d13da3
65a1fc
16
Other schemes

• Tapestry
• Distributed TRIE
• Viceroy
• Kademlia
• SkipGraph
• Symphony

17
Comparison
18
Unstructured P2P Overlay Networks

• Napster
• Gnutella
• Freenet
• KaZaA

19
Napster

• Centralized index
• File names gt
• active holder machines
• Sophisticated search
• Easy to implement
• Ensure correct search
• Centralized index
• Lawsuits
• Denial of service
• Can use server farms

P1
P5
S
P2
P4
P2
Where is quit playing games ?
FTP
P3
20
Gnutella

• Flooding
• Overlay network
• Decentralized
• Robust
• Not scalable.
• Use TTL. Query can fail
• Can not ensure correctness

P
P
P
P
P
P
P
P
P
21
FreeNet

• File is cached on reverse search path
• Anonymity
• Replication, cache
• Similar keys on same node
• Empirical log(N) lookup
• TTL limits search
• Only probabilistic guarantee
• Transaction state
• No remove( )
• Use cache replacement

P
2
1
P
P
3
12
7
11
4
6
P
10
P
5
P
9
8
P
22
KaZaA (FastTrack)

• Super-nodes
• Election
• capacity
• bandwidth, storage, CPU
• and availability
• connection time
• public address
• Use heterogeneity of peers
• Inherently non-scalable
• If flooding is used

P
P
P
P
P
P
P
P
P
P
P
P
23
P2P for IP telephony
REGISTER alice_at_postech.edu gt141.223.82.24
INVITE alice_at_postech.edu
Postech.edu
141.223.82.24
Alices host 141.223.82.24
Bobs host
24
Conclusions

• Summary
• P2P Review ( Concept and Features )
• P2P Overlay Networks
• Structured vs Unstructured P2P overlay networks
• CAN, Chord, Pastry
• Freenet, gnutella, KaZaA
• P2P IP Telephony

25
References

• A Survey and Comparison of Peer-to-Peer Overlay
  Network Schemes
• Eng Keong Lua, Jon Crowcroft, Marcelo Pias, Ravi
  Sharma and Steven Lim
• IEEE Communications survey and tutorial, march
  2004
• A Scalable Content-Addressable Network
• Sylvia Ratnasamy, Paul Francis, Mark Handley,
  Richard Karp
• SIGCOMM01, August 27-31, 2001
• Chord A scalable peer-to-peer lookup protocol
  for internet applications
• I. Stoica, R. Morris, D. Karger, M.F. Kaashoek,
  and H. Balakrishnan
• IEEE/ACM Transactions on Networking, vol. 11, no.
  1, pp. 17-32, 2003
• Pastry Scalable, distributed object location
  and routing for large-scale peer-to-peer systems
• A. Rowstron and P. Druschel
• Proceedings of the Middleware, 2001

26
Question and Discussion
Question??