Quantifying Impact of Mobility on Data
Availability in Mobile Ad Hoc Networks
Takahiro Hara
IEEE TRANSACTIONS ON MOBILE COMPUTING,
VOL. 9, NO. 2, FEBRUARY 2010.
1
1
Mobility model
•
1. Random Walk (RW)
– At every unit of experimental time, each MH randomly determines a movement direction
and speed from 0 to V [m/s].
•
2. Random WayPoint (RWP)
– MHs selects a random destinations with the speed. After reaching the destination, it
again pauses, and then, repeats this behavior.
•
3. Manhattan Mobility (MM)
– MM emulates the node movement on streets where nodes only travel on the pathways in
the map.
•
4. Reference Point Group Mobility (RPGM)
– Each group has a logical “center” called a reference point. MHs moves to the reference
point (nearby) based on the RWP model.
•
5. Random Waypoint with Locality (RWP-L)
–
–
The concept of Home area.
MHs choose a random destination insider the home area with high prob. H and one outside the region
with prob. 1-H. H: the home area ratio.
2
2
metrics
• For Data availability (data storage capacity)
– 1. Average size of partitions (Network)
– 2. Distribution of partition sizes (Network)
– 3. Size of partitions belonged to (Node)
– 4. Change in size of partitions belonged to (Node)
– 5. Distribution of connected nodes (Node)
• For Data distribution (data replication)
– 6. Total number of connected nodes (Direct)
– 7. Total number of data-reachable nodes (Indirect)
3
1. Average size of partitions
l: 幾個interval
m: 節點數
n1:表示第1個t時間內的partition個數
Time
t
t
t
t
T = l4* t
t
l=6
t
2. Distribution of partition sizes
5
5
3. Size of partitions belonged to
Mk: Mobile node k
6
Q:在ti時間，某1個節點，有不同二個partition的連結。
6
4. Change in size of partitions belonged to
7
7
5. Distribution of connected nodes
8
8
6. Total number of connected nodes
7. Total number of data-reachable nodes
9
9
7. Total number of data-reachable nodes
Ri denotes a set of mobile nodes that mobile nodes in Ni connected to from the
beginning of ti until the end of the observation time l’ * t
Cs,f,Mj denotes a set of mobile nodes that Mj have connected to during the duration
from ts to tf .
Ni denotes a set of mobile nodes that Mk first connected
to at the beginning of ti but
10
have never connected to before that.
10
1. Average size of partitions
Reference Point Group Mobility
Map:2500m x 2500m
Node:300
Transmission range: 100m
11
11
2. Distribution of partition sizes
1. Group mobility
2. Random Way Point
3. Manhattan Mobility
4. Random Waypoint with Locality
5. Random Walk
12
12
3. Size of partitions belonged to
1. Group mobility
2. Random Way Point
3. Manhattan Mobility
4. Random Waypoint with Locality
5. Random Walk
13
13
4. Change in size of partitions belonged to
14
14
5. Distribution of connected nodes
15
15
16
6. Total number of connected nodes
17
7. Total number of data-reachable nodes
18
T as 10,000,000 seconds
t =1, l = 10,000,000
The first 1000s is removed.
19
19
20
20
Random
Walk
Random
Way
Point
Property
Many small partitions
Mobility is very low in the long term
Not good in terms of storage capacity and data distribution
Data
distribution
Protocol should not rely on data sharing with a large number of
nodes but should data with a small number of connected nodes.
Data
diffusion
Protocol should consider effective data disseminations to reduce
the traffic
Property
Few large partitions
Mobility is middle
Good in storage capacity when the node belongs to a large
partition
Data
distribution
Which nodes it shares data by considering several factors such as
the number of neighboring nodes and stability of wireless links
Data
diffusion
Designer does not need to be very nervous for the performance in
terms of distribution rapidness.
It should address the reduction of unnecessary data redistribution
to reduce excessive data traffic
21
21
Manhattan
Mobility
Reference
Point
Group
Mobility
Property
Many small partitions but much fewer isolated nodes than
other models.
Data
distribution
It is effective to share data among nodes in the same partition.
Data
diffusion
Reduce data redistribution because maximum capacity of
partitions is small.
Property
Large partition (Best in storage capacity)
Very high connectivity among nodes in the same group
Data
distribution
Share data among nodes in the same group.
Data
diffusion
Address the reduction of unnecessary data redistribution
22
22
Random
Waypoint
with
Locality
Property
The node whose home area is the center of the entire area
gave better performance than the node whose area is a
corner of the entire area.
Data
distribution
Total number of connected nodes is greatly affected by the
increase in node density because the increased number of
nodes can bridge isolated partitions.
Data
diffusion
Similar to Random Walk model, it should aggressively
disseminate data to connected mobile nodes.
23
23
Related work
[8] J. Hahner et.al. “Quantifying Network Partitioning in Mobile Ad Hoc Networks,”
Proc. Int’l Conf. Mobile Data Management, pp. 174-181, 2007.
•
In [8]
– Network wide metrics:
• Number of partitions
• Average size of partitions
• Average partition change rate
– Node-centric metrics:
• Node partition change rate
• Node separation time
– The first two metrics represent the capacity of data storage (memory space) of each partition
– The larger the partition is, the more data can be stored in it.
– The other three metrics just represent how frequently members of each partition change or how
long before each pair of two nodes disconnects.
24
24
Motivation
– More specifically, these metrics cannot distinguish whether only one
node disconnects from the partition or the partition is split into two
partitions with the same size.
– Also, they do not represent how many nodes each node connects with
at a certain interval. Thus, they do not truly represent the dynamism of
partitions.
– In this paper, we propose new metrics to represent the dynamism of
partitions in MANETs.
25