Expected Quorum Overlap Sizes of Optimal Quorum Systems with the Rotation Closure Property for Asynchronous Power-Saving Algorithms in Mobile Ad Hoc Networks Presented by Jehn-Ruey Jiang Department of Computer Science and Information Engineering National Central University
41
Embed
Presented by Jehn-Ruey Jiang Department of Computer Science and Information Engineering
Expected Quorum Overlap Sizes of Optimal Quorum Systems with the Rotation Closure Property for Asynchronous Power-Saving Algorithms in Mobile Ad Hoc Networks. Presented by Jehn-Ruey Jiang Department of Computer Science and Information Engineering National Central University. Outline. - PowerPoint PPT Presentation
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
Expected Quorum Overlap Sizes of Optimal Quorum Systems with the Rotation Closure Property for Asynchronous Power-Saving Algorithms in Mobile Ad Hoc Networks
Presented by
Jehn-Ruey JiangDepartment of Computer Science and Information Engineering
National Central University
2/38
Outline
Mobile Ad hoc Networks Quorum-Based Asynchronous Power
Saving Algorithm Expected Quorum Overlap Size The f-Torus Quorum System Analysis and Simulation Results of
The blue ones do not know the existence of the red ones, not to
mention the time when they are awake.
The red ones do not know the existence of the blue ones, not to
mention the time when they are awake.
13/38
Asynchronous PS Algorithms (1/2)
Try to solve the network partitioning problem to achieve Neighbor discovery Wakeup prediction
Without synchronizing hosts’ clocks
14/38
Asynchronous PS Algorithms (2/2)
Three existent asynchronous PS algorithms
Dominating-Awake-Interval
Periodical-Fully-Awake-Interval
Quorum-Based (QAPS)
15/38
Quorum System
What is a quorum system?A collection of mutually intersecting subsets of an universal set U, where each subset is called a quorum.E.G. {{1, 2},{2, 3},{1,3}} is a quorum system under U={1,2,3}, where {1, 2}, {2, 3} and {1,3} are quorums.
Not all quorum systems are applicable to QAPS algorithms
Only those quorum systems with the rotation closure property are applicable. [Jiang et al. 2005]
16/38
Optimal Quorum System (1/2)
Quorum Size Lower Bound for quorum systems satisfying the rotation closure property:k, where k(k-1)+1=n, the cardinality of the universal set, and k-1 is a prime power(k n ) [Jiang et al. 2005]
17/38
Optimal Quorum System (2/2)
Optimal quorum system FPP quorum system
Near optimal quorum systems Grid quorum system Torus quorum system Cyclic (difference set) quorum system E-Torus quorum system
18/38
Numbering Beacon Intervals
0 1 2 3
4 5 6 7
8 9 10 11
12 13 14 15
And they are organized
as a n n array
n consecutive beacon intervals are numbered as 0 to n-1
Mobile Ad hoc Networks Quorum-Based Asynchronous Power
Saving Algorithm Expected Quorum Overlap Size The f-Torus Quorum System Analysis and Simulation Results of
EQOS Conclusion
30/38
Performance Metrics
SQOS: smallest quorum overlap sizefor worst-case neighbor sensibility
MQOS: maximum quorum overlap separationfor longest delay of discovering a neighbor
EQOS: expected quorum overlap sizefor average-case neighbor sensibility
New Contribution
f-torus quorum system
31/38
New Contributio
n
32/38
33/38
New Contributio
n
34/38
35/38
36/38
Conclusion (1/2)
We have proposed to evaluate the average-case neighbor sensibility of a QAPS algorithm by EQOS
We have proposed a new quorum system, called the fraction torus (f-torus) quorum system, for the construction of flexible mobility-adaptive PS algorithms.
We have analyzed and simulate EQOS for the FPP, grid, cyclic, torus, e-torus and f-torus quorum systems
37/38
Conclusion (2/2)
f-torus quorum systems may be applied to other applications: location management, information dissemination/retrieval data aggregation
in mobile ad hoc networks (MANETs)and/or wireless sensor networks (WSNs)
38/38
Thanks
39/38
Rotation Closure Property (1/3)
Definition. Given a non-negative integer i and a quorum H in a quorum system Q under U = {0,…, n1}, we define rotate(H, i) = {j+ijH} (mod n).
E.G. Let H={0,3} be a subset of U={0,…,3}. We have rotate(H, 0)={0, 3}, rotate(H, 1)={1,0}, rotate(H, 2)={2, 1}, rotate(H, 3)={3, 2}
40/38
Rotation Closure Property (2/3)
Definition. A quorum system Q under U = {0,…, n1} is said to have the rotation closure property ifG,H Q, i {0,…, n1}: G rotate(H, i) .