pmac2wn07-slide-v0
Download
Report
Transcript pmac2wn07-slide-v0
Modeling and Analysis of Regional
Registration Based Mobile Service
Management
Authors: Ing-Ray Chen and Ding-Chau Wang
Presented by Chaitanya ,Geetanjali and Bavani
Overview
Introduction
Related Work
Protocol Description
Model
Numerical Data And Analysis
Simulation
Introduction
Multicasting
Single sender, multiple receivers
Efficient Multicast Protocol
Minimum data duplication
Minimum distance travelled
Mobile Multicasting
Multicasting in mobile networks
Challenges for Mobile Multicast Protocol
Dynamic group membership
Dynamic member topology
Basic Schemes for Mobile Multicasting
Remote Subscription (RS):
Mobile host (MH) always needs to subscribe to its multicast
group when it enters or changes a foreign network
Handoff frequency of mobile host is proportional to the
update frequency of the multicast tree
Advantage: Optimal shortest path
Disadvantage: Extra Overhead for reconstructing multicast
tree
Bi-directional Tunneling (BT):
A MH receives multicast data by way of its home network
using unicast mobile IP tunnels from its Home Agent(HA)
Advantages:
Handles both source and recipient mobility
No need to update the multicast tree when MH’s
location is changed
Disadvantages:
Packet delivery path is not optimal
Limited scalability
Related work
mMOM
Hybrid approach of BT and RS
Every MH must re-register with FA after a period of
residence time(Life time)
A MH applies either BT or RS based on its mobility
If the MH is highly mobile, BT will be used
If the MH is immobile, RS will be used
Advantage: Simple and Practical
Disadvantage: Does not allow care of address to be used in
mobile IP
Multicasting with Multicast Agents
Multicast agents(MA)
MA provides multicast services to the mobile group
members in the multiple foreign agents.
MA maintains a list of multicast groups and FAs
that have visiting mobile members for that group.
Routing with multicast agents
Multicasting with Multicast Agents (Contd.)
MA joins the multicast group on behalf of mobile
group members in it service area.
MA tunnels multicast packets for these groups to FA
FA delivers packets to mobile hosts
Advantages:
Stable structure
Avoids frequent modifications to multicast tree
Disadvantages:
Lacks flexibility
Single point failure
Range Based Mobile Multicast (RBMoM)
Introduces a Mobile Multicast HomeAgent(MMA)
MMA multicasts packets to FA to which the MH is currently
attached
Each MHA must always be one of the multicast group member
Figure: Setup of a new MHA
Range Based Mobile Multicast (RBMoM) (Contd.)
The MHA information is recorded at MHs HAs agent table
MMA handoff s occur if a MH is out of current MMA’s service
range
Advantages:
Dynamic MMA
Disadvantages:
Communication overhead and performance penalty to the
network
Protocol Description
Proposed protocol
User-oriented Regional Registration based Mobile
Multicast (URRMoM)
Combines advantages of RS and BT
Each MH’s can autonomously determine its optimal
service area (MMA) based on its dynamic mobility and
service characteristics
Advantages:
No need to maintain Agent table
Minimizes network traffic
Simple , scalable and efficient.
Protocol Description (Contd.)
MMA is responsible for tunneling multicast packets to FA as
long as the FA is within the regional MMA’s service area.
Each MMA will be a member of multicast tree
Each MH should have one MMA
MMA of the MH will change as it roams in the network
Each MH keeps a counter to record the number of subnets it
has crossed within the service area of its MMA
Regional service area of MMA = Number of subnets covered
by the MMA
When the FA is MMA
The MMA of the MH will be updated to the current FA
The counter in the MH will be reset to 0 after the MMA reset
When the FA is not a MMA
The counter in the MH will increment by 1
When the counter in the MH reaches to the regional size
(R) the multicast
The new FA will subscribe to tree and become a new
MMA for the MH
Types of Moves in RRMoM
1.
Intra-Regional:
Occurs? -> Whenever a MH performs a location
handoff “within” a multicast service area of a regional
MMA.
Change in MMA? -> only if the new FA it enters into is
itself a MMA for other MHs.
In this case, the MH’s MMA is updated to the current
FA.
Type of Moves - II
2.
Inter-Regional:
Occurs? -> Whenever a MH moves across a service area
(the counter reaches R), thus incurring a multicast service
handoff.
Change in MMA? -> The MH’s MMA always changes.
If the new FA is itself a MMA, then the MH’s MMA is
simply updated to the current FA.
Otherwise, the current FA becomes the MH’s new MMA. A
multicast tree subscription event is triggered to add the
new MMA to the multicast tree.
Hypothesis
There exists an optimal service area size that will
minimize the network traffic generated due to mobile
multicast services.
It depends on:
the mobility of MHs
population of the MHs
the size and topology of the network.
MODEL
Is fixed
Each
node is a
subnet
with FA
A MH can move
in four
directions
randomly with
equal
probability
Relationship between λ and μ
Let μ = MH’s residence time in FA (exponentially
distributed)
Let λ = arrival rate of a single MH to any FA in this n x n
homogenous network
Then, λ = μ / (n2-1)
M/M/∞/M
Let M = number of MHs belonging to the multicast
group
The arrival-departure process of M members to a FA
(a subnet) is modeled as M/M/∞/M model
Solving the linear equations for Pi and using
P0 = (1-1/n2)M (Substituting , λ = μ / (n2-1))
Average number of members in the multicast group residing
under one FA:
a MMA on average
covers R subnets
The average number of multicast members
a MMA covers: R.
Average number of MMAs in the system is :
Probability that a FA in which a MH just enters is a MMA,
(PMMA),
SPN – Behavior of a MH in network
Performance Metrics
Total cost is given by:
CMaintenance : Cost incurred per unit time due to control packets for tree
management = MMA Subscription cost + MMA Un-subscription cost
Let rsub = Rate at which a member subscribes a new MMA to the
multicast tree after it has crossed R subnets
Let β = Average number of hops separating a MMA and multicast source.
Let τ = Average per-hop communication cost.
Total Subscription Rate = rsub x M
Total Unsubscription Rate =
Total Cost
CService = Cost per unit time for delivering multicast packets from the
multicast source to MHs in the multicast group.
CService = Cost per packet delivery x rate at which packets are generated
= number of hops for multicast packet delivery from the
multicast source to MMAs
= number of hops through which packets are tunneled
from various MMAs to M MHs.
NUMERIC DATA AND ANALYSIS
Figure shows the total traffic
generated as a function of the
service area size R expressed in
terms of the number of subnets
Optimal service area size under
which the network traffic
generated is minimized
As the mesh network becomes
Cost vs. Regional Area Size (R)
with varying n.
larger, the optimal service area
size becomes larger and larger
Cost vs. R with Varying Number of MHs
Figure shows the network
traffic generated vs. R as M
varies in an 8 by 8 mesh
network.
As M increases the optimal R
decreases.
Effect of the Distance between Source and MMA
Figure shows that when β
increases, the optimal range R
increases for the case when M is
fixed at 100
Here,
β =average number of hops to
reach the source for multicast tree
subscription/un-subscription
Comparison of URRMoM vs RS and RBMoM
Figure compares the network
traffic generated due to
maintenance vs. the network size
n for URRMoM vs. RS and
RBMoM at optimizing R values
under the same set of parameter
values.
URRMoM always produces the
least amount of network traffic
compared with RS and RBMoM
SIMULATION
SMPL has been used to conduct a simulation study to
validate the analytical results reported in Numerical Data
and Analysis section.
To ensure statistical significance of simulation results, a
batch mean analysis (BMA) technique has been adopted
SIMULATION RESULTS
Simulation Results - Cost vs. R with
varying n.
Simulation Results - Cost vs. R with
varying Number of MHs
Simulation Results – Comparison of URRMoM vs RS and RBMoM
CONCLUSIONS
Proposed and analyzed user-oriented regional registration based mobile
multicast (URRMoM) approach
Combines distinct performance advantages of remote subscription and bi-
directional tunneling
Mathematical model to analytically determine the optimal service area size
under which the overall network traffic generated due to multicast tree
maintenance and multicast packet delivery can be minimized
Effect of key parameters on the optimal regional area size
Reasons for the sensitivity analysis has been provided
FUTURE WORK
In the future, empirical validation of URRMoM in
an experimental testbed is planned to be
performed
THANK YOU