1 Replication Strategies in Unstructured Peer-to-Peer Networks Edith Cohen, Scott Shenker ACM SIGCOMM Computer Communication Review, Proceedings of the 2002 conference on Applications, technologies, architectures, and protocols for computer communications, vol. 32 issue 4 Presentation by Tony Sung, MC Lab, IE CUHK 16th December 2004
22
Embed
Replication Strategies in Unstructured Peer-to-Peer Networks
Replication Strategies in Unstructured Peer-to-Peer Networks. Edith Cohen, Scott Shenker ACM SIGCOMM Computer Communication Review, Proceedings of the 2002 conference on Applications, technologies, architectures, and protocols for computer communications, vol. 32 issue 4 - PowerPoint PPT Presentation
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
1
Replication Strategies in Unstructured Peer-to-Peer Networks
Edith Cohen, Scott Shenker
ACM SIGCOMM Computer Communication Review, Proceedings of the 2002 conference on Applications, technologies, architectures, and protocols for computer communications, vol. 32 issue 4
Presentation by Tony Sung, MC Lab, IE CUHK16th December 2004
2
Introduction
What is an Unstructured P2P Network?
Centralized
Decentralized
Structured
Unstructured
3
Introduction
Locating Objects in an Unstructured P2P Network
Probing
How to Reduce Probe Count?
No Probing is better than Random Probing
By Replication
4
Introduction
Current Replication Strategies
… Implicit
Objective of the Paper:
“Designs an explicit replication strategy.”“What is the optimal way to replicate data?”
5
Introduction
Two Starting Points
Uniform Replication
Proportional Replication
6
Paper’s Outline Introduction Model and Problem Statement
Defining an Allocation and the Expected Search Size Bounded Search Size and Insoluble Queries Heterogeneous Capacities and Bandwidth
Allocation Strategies Uniform and Proportional Characterizing Allocations Between Uniform and Proportional
The Square-root Allocation How much we can gain?
Square-root* and Proportional* Allocations Square-root* Allocation Proportional* Allocation
Distributed Replication Path Replication Replication with Sibling-number Memory Replication with Probe Memory Obtaining the Optimal Allocation Simulations
Conclusion
7
Today’s Outline Introduction Model and Problem Statement
Defining an Allocation and the Expected Search Size Bounded Search Size and Insoluble Queries Heterogeneous Capacities and Bandwidth
Allocation Strategies Uniform and Proportional Characterizing Allocations Between Uniform and Proportional
The Square-root Allocation How much we can gain?
Square-root* and Proportional* Allocations Square-root* Allocation Proportional* Allocation
Distributed Replication Path Replication Replication with Sibling-number Memory Replication with Probe Memory Obtaining the Optimal Allocation Simulations
• Thus minimizes system resources spent on insoluble queries
• Minimizes maximum utilization rate.
• More relevant when the replication is of copies rather than of pointers
11
Allocation Strategies, Uniform & Proportional
Expected Search Size Aq(p)
Uniform
Aq(p) = 1/ρΣ(qi/pi)
= 1/ρΣqim
= m/ρ
Proportional
Aq(p) = 1/ρΣ(qi/pi)
= 1/ρΣ1
= m/ρ
12
Allocation Strategies, Characterizing Allocations
Consider space allocations for two items pi, pj and qi, qj
Range of allocation defined by x, 0 < x < 1,
pi/(pi +pj) = x
pj/(pi +pj) = (1-x)
x = qi/(qi +qj) [Proportional] or 0.5 [Uniform]
ESS proportional to qi/x + qj/(1-x) and is convex.
ESSmin occurs at which is independent of p.
13
Allocation Strategies, Characterizing Allocations
Consider space allocations for two items pi, pj and qi, qj
14
Allocation Strategies, Between Uniform & Prop.
15
Allocation Strategies, Between Uniform & Prop.
16
Allocation Strategies, Short Conclusion
ESS of Uniform and Proportional Allocation is equal, and is equal to m/ρ
For one special case (m=2), ESS is a convex function and is minimum for a square-root allocation
For any allocation p that lies between Uniform and Proportional, its ESS is at most m/ρ.
If p is different from Uniform or Proportional then its ESS is strictly less than m/ρ.
17
The Square-root Allocation
18
How much can we gain?
For uniform and proportional allocation,
ESS = m/ρ
For Square-root allocation,
ESS = (Σqi1/2)2/ρ
which depends on the query distribution
Define gain factor as ESSuniform/ESSSR
It is shown that ESSuniform/ESSSR ≤ m(u + l - mlu)
When l = 1/m or u = 1/m, the only legal allocation is pi = 1/m, and gain factor = 1If l << 1/m, and gain factor is roughly mu.
19
How much can we gain?
20
How much can we gain?
21
Materials Left
Natural extension of Square-root and Proportional Allocation that are defined when l is fixed for a maximum search size. Similar Results
Distributed Replication Protocols for achieving Square-root Allocation Path replication, converges but unstable Replication with sibling-number memory, better Replication with probe memory, better Confirmed with Simulation
22
Conclusion
Modeled different replication strategies Uniform Proportional In-between, especially Square-root
Uniform and Proportional forms two extremes of all legal allocations