Updates in Highly Unreliable, Replicated Peer-to-Peer Systems

1
Updates in Highly Unreliable, Replicated
Peer-to-Peer Systems

• Anwitaman Datta, Manfred Hauswirth, Karl Aberer
• (EPFL)
• Presented by Zhiyuan Troy Zhan

2
Outline

• Motivation
• System Model Algorithm
• Analytical Model Analysis
• Related Work
• Conclusion

3
Motivation

• Peer-to-Peer System is not just about file
  sharing.
• Data items can be added, deleted and updated
  frequently.
• Peer commerce
• Shared calendars/address books
• Trust management
• Medical information sharing
• Replication is used to improve fault-tolerance
  and response time.

4
Motivation contd

• How to disseminate updates to other peers is the
  target problem
• Consistency guarantee
• Scalability, underlying system assumption,
  resource consumption
• Challenges
• Huge number of peers
• Peers can go online/offline at any time
• Often lack of global knowledge

5
Motivation contd

• Contributions
• Address the update dissemination problem with low
  online probabilities of peers (lt30) and no
  global knowledge.
• Present a fully decentralized, efficient and
  robust communication scheme based on rumor
  spreading.
• A generic analytical model of combined push/pull
  technique.

6
Motivation - Problem Statement

• Assumptions
• Low percentage of online peers, impossible to
  achieve any kind of quorum.
• Transactional consistency is not required,
  eventual consistency is desirable in most
  applications.
• Update conflicts is very rare, the paper does not
  handle it.
• Probabilistic guarantee of successful search are
  sufficient.
• Total of replicas is substantially lower than
  total of peers (i.e. 1000 vs 1000,000).
• Consecutive updates can be distributed sparsely
  over time.
• Communication overhead is the major performance
  measurement.

7
Outline

• Motivation
• System Model Algorithm
• Analytical Model Analysis
• Related Work
• Conclusion

8
System Model

• A peer-to-peer overlay network.
• Each peer has its own local knowledge, i.e.
  routing table, replica list, etc.
• Peers can go offline at any time.
• A communication channel can be established
  between any two online peers, otherwise, assume
  each other offline.

9
Algorithm Push Phase

• Executed when disseminating updates.
• At replica p, upon receiving message
  Push(U,V,Rf,t)
• IF Push(U,V,Rf,t) not processed THEN
• Select a random subset Rp of replicas with
  RpRfr
• With probability PF(t), send Push(U, V,
  RfRpp, t1) to Rp-Rf
• Set Push(U,V,Rf,t) as processed

10
Algorithm Push Phase,contd

• U actual update data item
• V version vector. Contains global version
  identifiers (GUID, can be computed locally),
  altered data items are treated as distinct
  coexist versions.
• R, Rf, Rp replicas.
• PF(t) a function of t.

11
Algorithm Pull Phase

• Executed when a peer recovers from failure, or
  reconnects, or receives no updates for a while,
  or receives pull message but not sure whether
  itself is in sync.
• Contact online replicas
• Inquire for missed updates based on version
  vectors

12
Outline

• Motivation
• System Model Algorithm
• Analytical Model Analysis
• Related Work
• Conclusion

13
General Assumptions

• Assume an update U is initiated for R online
  replicas.
• In general, the online population in push round
  t
• Ron(t)Ron(t-1)xR-Ron(t-1)y
• x1-p, yq
• p probability of an online peer going offline in
  one push round q probability of an offline peer
  coming online in one push round.
• p,q are typically small and may vary in different
  rounds.
• ASSUME p is constant and omit peers coming
  online
• Ron(t)Ron(t-1)x
• ASSUME fr is constant.

14
Analysis of Pull Phase Round 0

• Total number of messages msg(0)Rfr
• New replicas which receive the update
  newreplicas(0)Ron(0)fr
• Online replicas that do not receive the update
  Ron(0)(1-fr)
• Message length (size, denote U as the update
  message size) ML(0)URfrB (B size of data
  required to describe one replica meta data),
  only consider U and Rf

15
Analysis of Pull Phase Round 1

• of messages in round 1 msg(1)Ron(0)frxPF(1)
  Rfr(1-fr)
• of replicas that newly pushed with updates
  after round 1 newreplicas(1)Ron(0)x(1-fr)
  1-(1-fr)Ron(0)frxPF(1)
• Length of message ML(1)URB(frfr(1-fr))
  URB(1-(1-fr)2)

16
Analysis of Pull Phase Round tgt2

• Define fd_aware(t) and faware(t)
• fd_aware(t) Increment in fraction of online
  replicas which are aware of the update after
  round t
• faware(t) Total fraction of online replicas
  which are aware of the update at the beginning of
  round t
• faware(t) faware(t-1) fd_aware(t-1)

17
Analysis of Pull Phase Round tgt2, contd

• newreplicas(t)Ron(t-1)(1-faware(t-1))x 1 -
  (1-fr)Ron(t-1)fd_aware(t-1)xPF(t) in paper
• newreplicas(t)Ron(t-1)(1-faware(t))x 1 -
  (1-fr)Ron(t-1)fd_aware(t-1)xPF(t) I think
• Given fd_aware(t)(1-faware(t))1-(1-fr)Ron(t-1)
  fd_aware(t-1)xPF(t) , we have
• faware(t) faware(t-1) fd_aware(t-1)1-(1-faware(
  t-1))(1-fr)Ron(t-2)fd_aware(t-2)xPF(t-1) .
• faware(t) rapidly grows to 1

18
Analysis of Pull Phase Round tgt2, contd

• If the partial list is ignored
• msg(t)Ron(t-1) fd_aware(t-1)xRF(t) Rfr
• If the partial list is considered
• msg(t)Ron(t-1) fd_aware(t-1)xRF(t)
  Rfr(1-fr)t - (1)
• ML(t)URB(1-(1-fr)t1) - (2)
• Both (1) and (2) are proved in the paper by
  induction on t.

19
Analysis of Pull Phase

• Case1 a replica p comes online after a push
  phase is over
• Trivial, assume other online replicas have got
  the update already.
• Case2 p comes online during the push phase,
  suppose faware fraction of the replicas Ron are
  aware of the updates, the probability of p
  getting the update in m attempts is
• 1-1-(Ron faware /R)m -(3)
• Query similar to Pull, but may need majority
  logic, or version scheme, or hybrid of two, to
  identify the latest updates

20
Analytical Results
Varying initial online population Ron(0), 1
21
Analytical Results contd
Varying initial online population Ron(0), gt5
22
Analytical Results contd
23
Analytical Results contd
24
Analytical Results contd, Parameter tuning
25
Analytical Results contd, scalability
26
Discussions

• Comparison with Gnutella
• Parameter self-tuning (Optimization)

27
Outline

• Motivation
• System Model Algorithm
• Analytical Model Analysis
• Related Work
• Conclusion

28
Related Work

• Replication and updates in DB
• iAnywhere Solutions Server-based approach
• Bayou assumes significantly less replicas, less
  updates, disconnections are short
• Some other approaches assume availability of
  resource and replicas in general

29
Related Work contd

• Group communication and lazy epidemic schemes
• Similar work Bimodal multicast, epidemic updates
• None has done special case study of bimodal
  behavior and the utility of epidemic algorithms
  in a highly unreliable environment

30
Outline

• Motivation
• System Model Algorithm
• Analytical Model Analysis
• Related Work
• Conclusion

31
Conclusion

• This paper provides an analytical model to
  demonstrate the significant reduction of message
  overhead using combined push and pull
  techniques.
• Totally decentralized solution, no global
  knowledge is needed.
• The paper is available at citeseer. Will appear
  in ICDCS 2003.