Reliability Study of wireless corba using Petri net and end to end instantenous reliability measure

RELIABILITY STUDY OF
WIRELESS NETWORK
Supervised by:
Prof. Dr. Hassan Farahat
Prof. Assist. Ibrahim Tarrad
Presented by:
Eng. Ahmed Samir Koriem
Master Degree Seminar- Al-Azhar University-
Communication and electronics
eng.ahmedkoriem@gmailcom

Aim of Thesis ?
•Real Methodology for study the reliability of the
wireless network using Petri Net modeling tool
•Study End to end instantaneous Reliability {EIR(t)}
measure
•Study CORBA Network as a wireless network
Master Degree Seminar- Ahmed Samir Koriem- eng.ahmedkoriem@gmailcom

Why study the reliability for wireless network
not for wired network??
Wireless networking has rapidly increased in popularity over recent years.
• Most of the available research works have embraced wireless technology
and businesses as a great impact on their operational efficiency.
• Recently, the benefits of using wireless technology in such networking are
more than those of wired networking (mobility).
• Wireless networks inherit the unique handoff characteristics, which lead
to different communication patterns among the communicating network
resources.

Why CORBA network?
• Currently, CORBA network represents a Practical wireless mobile
computing network.
• The available practical benchmarking results of CORBA network helps
us to understand the working mechanisms of the resources of such
practical network.
• To the best of our knowledge, there is no work concentrated with the
performance modeling of such network. Recently, the reliability of
CORBA network using Markov Chain modeling technique has studied.
However, this work has many problems.

Visited Domain
Home Domain
Terminal
Domain
Access
Bridge
Access
Bridge
Access
Bridge
Access
Bridge
Static
Host
Static
Host
Terminal
Bridge
GIOP
Tunnel
ab1
ab2
mh1
GTP Messages
Wireless CORBA Architecture
Home
Location
Agent

Basic resources of wireless CORBA network
MH MH
Wired Network
(Fixed/Static network)
MH MH
MSS
MH
MH MH MH
Wireless Link
FIFO Channel
Wireless Cell Wireless Cell
MH: mobile host
MSS: mobile support station
SH: static host
HLA: Home Location Agent
SH
SH
SH
SH SH
HLA
MSS MSS
MSS

1- System states in the MS
scheme:
(a) normal communication
(b) handoff procedure.
2- System states in the SM scheme:
(a) location-querying
(b and c) location-forwarding.
(2-a)
(2-b)
(2-c)
(1-a)
(1-b)
3- System states in the MM scheme
(a) MH in handoff
(b) MH in handoff
(c) both MH & MH in handoff
(3-a) (3-b) (3-c)
Four Communication Schemes of wireless CORBA
Network
4- System states in the SS scheme

System states in
the MS scheme at
Normal
Communication
System states in the MS
scheme at Handoff
procedure
Markov model for
the MS scheme
  denotes the handoff rate= 1/ handoff time
  denotes the handoff completion rate= 1/ Service time
Markov Chain Model for MS Scheme
MSS
SH
MH
MSS1
SH
MH
HLA
MSS2
(a) (b)
S1S2



Markov Chain Model Calculations
t
et )(
1 )( 




 





t
et )(
2 )( 




 





The probabilities of the system in states S1 and S2, at time t are calculated
analytically, as follows:

End to End Instantaneous Reliability
 s ss tRttEIR )()()( 
•As MH moves and performs handoff operation, the communication structure
(i.e communication patterns that describe the communication processes
among the active network resources) will vary with time. Analytically, the end-
to-end Expected Instantaneous Reliability (EIR) measure can be calculated as
follows



)(
1
)()(
snc
i
is tRtR
s(t) : is the probability that the system is in state S at time t, where a system state s can be
defined as the communication structure, therefore S changes with time t.
Rs(t) : is the reliability of the system in state S at time t. Rs(t) can be expressed as follows:
n(s) : is the number of engaged resources (n(s) =2,3,…, ) in system state S,
c : is the type of a ith resource, which may take a value mh, mss, sh, or hla for MH, MSS,
SH, or HLA respectively.
Ri(t) : is the reliability of ith resource.
EIR(t) : is a function composed not only of failure parameters but also of service
parameters introduced by state probability s(t).

][)(][)()( )()(
2
)()(
1
ttt
s
t
ss
mssmhmssmh
eeteetRt 
 

Time (t)(unit time) 100 200 300 400 500
Reliability measure =
EIR(t) Analytically
0.818730661330 0.670319934917 0.548811349136 0.449328772724 0.367879270041
The constant failure parameters for the four resources of wireless CORBA :
(i.e., MH, MSS, SH, and HLA) are mh, mss, sh and hla, respectively. Exponential distribution is utilized as
the service and failure distributions for model simplicity. Thus, the End-to-End EIR at a generic time t,
is given by
mh =0.001
mss=0.001
sh =0.0001
hla =0.001

Conclusion of the analytical model
• New Reliability measures for wireless communication EIRPN(T) that
measure the changing in the wireless communication system structure
with time.
• Handoff rate and Handoff completion rate are new parameters which
assumed in the markov chain model, handoff completion rate means the
service time that the system spent in program execution.

The problems in the analytical technique (Markov
Chain)
• Firstly, the analytical technique divides the reliability analysis of
CORBA network into four parts, as explained earlier. Such
reliability aspects cannot give us the complete reliability picture
for the practical wireless network in an accuracy way.
• Secondly, the communication processes as well as the handoff
processes of the network resources have been represented by
random variables  and . Such parameters cannot describe the
practical communication mechanisms or the handoff operations
that can occur among the different resources of the practical
wireless network.

How to get the solution for that
problems???

Petri Nets
 Petri Net (PN) is a Mathematical formalism and a graphical
tool for the formal description of the logical interactions
among parts or of the flow of activities in complex systems.
 Used as a visual communication aid to model
the system behavior.
 consists of three types of components: places
(circles), transitions (rectangles/bar) and arcs
(arrows):
• Places represent possible states of the system
• Transitions are events or actions which cause the
change of state
• Every arc simply connects a place with a transition
or a transition with a place.
A place
A transition
A token
Input Arc
Output Arc

 The original PN did not convey any notion of time.
 For performance and dependability analysis it is necessary
to introduce the duration of the events associated to PN
transitions.
Generalized Stochastic Petri Nets
(GSPN)
 Generalized Stochastic Petri Nets (GSPN) is able to model
real systems with an appropriate granularity of time.

GSPN Reliability model to describe the concept of
EIR End to END
•p1 and t1 models the handoff rate 
•p2 and t2 models the handoff completion rate

•p3, t6 and p10 models the failure process (with
rate MH) that can occur to MH before starting
its handoff process
rate MSS) that can occur to MSS before
starting
its handoff process
rate MH) that can occur to MH through
its handoff process
rate MSS) that can occur to MSS through
its handoff process.
Handoff
communication
process completed
p1

p2
t2
 
t1
p10
p3 p4
p9
t5
p8
p6
t4
p7
p5
t3
t6
MH has Failed
through Handoff
processes
MSS has Failed
through Handoff
processes
MH
MSS
MH has Failed before
Handoff processes
MSS has Failed
before Handoff
processes
Failure rate
of MH
Failure rate
of MSS
MH
Failure rate
of MH MSS
Failure rate
of MSS
 Handoff in progress

MH
MH
MSS
MSS
MH Movement
MH Movement

Reachability graph of GSPN Reliability model
M0: p1 p3 p4
M1: p2 p5 p6 M3:p1 p4 p10
M4: p1 p5 p6 M5: p2 p6 p7 M6: p2 p5 p8 M10A: p1 p9 p10
M7: p1 p6 p7 M8: p1 p5 p8
M11A: p1 p7 p8
M9: p2 p7 p8
M2: p1 p3 p9
t1
t5
t6
t4
t3
t2
t6 t5
t4
t3
t4
t2
t3
t2t4

The EIR(t) reliability measure can be calculated from
the Petri net reliability model
Probability that the MH is failed = Prob(p7  1 token) and Prob(p10  1 token).
Probability that the MSS is failed = Prob(p8  1 token) and Prob(p9  1 token).
Where, the place pi identifies the condition of failed component MH, or MSS.
Initial EIRPN(t) =1  { Prob(p7  1 token) + Prob(p8  1 token) +Prob(p9  1 token) +
Prob(p10  1 token) }
EIRPN(t) = Initial EIRPN(t) +  absorbing states that shown in the Reachability
= Initial EIRPN(t) + Marking M10A + Marking M11A
•Probability that the condition:
M10A: p1=1, p9=1, and p10=1
M11A: p1=1, p7=1, and p8=1 is true at time t.

Comparison between EIRPN(t) of Petri net
models and EIR(t) of Markov chain models
Time (t)
(unit time)
Reliability measure =
EIR(t) Analytically
Reliability measure = EIRPN(t)
Petri Net methodology
Percentage of accuracy
between Petri Net &
Analytical methodologies
100 0.818730661330 0.818722316146 99.9999 %
200 0.670319934917 0.670317501561 99.9999 %
300 0.548811349136 0.548810995919 100 %
400 0.449328772724 0.449328901792 100 %
500 0.367879270041 0.367879577849 100 %

Main contributions in our research
• I: Build GSPN performance model for describing the dynamic behavior (e.g. services
processes, and communication processes) of CORBA wireless network under the
desired real workload (e.g. executing parallel programming). Such model will give
the complete picture for the network communication service instead of using a
random parameter for representing such service.
• II: Build GSPN performance model for describing the occurrence of handoff
processes among the network cells. Such model will give the complete picture for the
network handoff process instead of using a random parameter for representing such
service.
• III: Build GSPN reliability model for describing the dynamic behavior of the various
failure network resources.
• V: Use the modeling power of the graphical representation of Petri net technique to
concise the above mentioned three models (I, II, III) into a new generic reliability model
for wireless network.
• IV: Calculate the reliability metric EIR(t) from the generic reliability model and
compare its results with those obtained from the analytical technique

Dynamic Behavior of CORBA Network
at the execution of a program consists of M parallel tasks
(Main Concept ofWorkload)
As a Coordinator
MH11 Distributes
a part of M-
tasks
All the desidered
MHs
Executing M-tasks
Concurrently
Coordinator MH11
Collects the first
step of M-task
results from other
MHs, and exchanges
information between
them
MH11 
 MH14
 MH13
 MH12
 MH11
 MH11
Next step of executing the remaing M-tasks

Steps of execution a program consists of M
parallel tasks on the CORBA network
MSS4 MSS3
MSS2
MH11
MSS1
Wireless Link
FIFO Channel
1
SH
SH
SH
HLA
MH21
MH12MH22
MH13MH23
MH14MH24
2
3
4
5
5
5
5
6
5
6
5
7
5
8
5
9
5
9
5
10
11
7
5
7
5
8
5
Step-1: Assume the wireless MH11 receives the desired program
from the user. Then, the coordinator MH11 divides the entire
program into subtasks and distributes these tasks as follows:
MH11 services task-1, MH12 services task-2, MH13 services
task-3, and MH14 services task-4,
Step-2: MH11 sends task-2, task-3, and task-4 to MSS1 to distribute
them to their MHs destinations.
Step-3: MSS1 asks HLA about the locations of the destination MSSs.
Step-4: MSS1 receives the requested information from HLA.
Step-5: Based on parallel fashion, MSS1 sends task-2 to MSS2, task-3
to MSS3, and task-4 to MSS4.
Step-6: Each MSS sends the desired task to its destination MH as
follows: MSS2 sends task-2 to MH12, MSS3 sends task-3 to
MH13, and MSS4 sends task-4 to MH14.
Step-7: All the wireless MHs execute the M-tasks concurrently.
Step-8: Except the coordinator MH11, each MH sends the results of
execution its task to its associated MSS. Thus, MH12, MH13,
and MH14 send their results to MSS2, MSS3, and MSS4
respectively.
Step-9: MSS2, MSS3, and MSS4 send the results to MSS1.
Step-10: MSS1 sends all the first results of M-tasks to the coordinator
MH11.
Step-11: The coordinator MH11 exchanges the information
between the M-tasks.
Step-12: Repeat Steps 6 to11 until the M-tasks programming are
finished.

 A GSPN model for describing
the execution processes of M
parallel tasks in the COBRA
network
 we calculate the
performance measure
"Mean Time to Complete
the Program (MTCP)"
MTCP = TATS / πinitial
Where
MTCP : is the time from the submission of the
user’s program to the CORBA network until the
user receiving his service.
πinitial : is the steady state probability of being in
the initial state (marking) M0. Initial marking M0
represents the state that a place p1 contains a
token (s).
TATS : is the mean time spent in the initial state
M0.
p1p2p3
●
●
●●
● ●
p4
p5p6
p7
p10
p8
p9
p11
p12
p16
p17
p18
p13
p14
p15
p19
p20
p21
p22
p23
p24
p25
t1t2
t3t4
Step1Step2Step3Step4
t5
Step5
t6
t7
t8
t9
t10
t11
t12
t13
t14
t15 t16
Step6
Step7
Step6 Step7
Step6
Step7Step6
Step7
Step10 Step11,
Step12
Step8, Step9
MSS1MSS3
MSS2MSS4
MH11
MH12
MH13
MH14
Task1
Task2Task4
Task3
t2
t2
t2
t2
MH11 Exe.
Task1
MH13 Exe.
Task3
MH12 Exe.
Task2
MH14 Exe.
Task4
Entire
Program
Tasks at
MH11
Tasks at
MSS1
HLA
MSS1
distribute
Tasks
4

Steps of execution M-tasks through the
occurrence of MH handoff
Wired Network
MH1
MSS1
Wireless Cell-1
MH1
MSS2
Wireless Cell-2
HLA
2- MH1 sends a "new
connection with MSS2
message" to MSS1
3- MSS1 sends a " MH1 new
connection message" to MSS2
4- MSS2 requests
HLR to "update
MH1's location".
6- HLA sends a "MH1
registration cancellation
message" to the old MSS1
8- MSS1 asks
MH1 to start
handoff process
to MSS2
1- MH1 neither
sends nor
receives any
message within
the current cell
(it breaks its
connection with
MSS1)
10- MH1 sends "location
registration message" to MSS2 at
Cell-2 and asks MSS2 to modifies its
location list
9- MH1 moves from cell-1 to cell-2
5- HLA updates its record
indicating the current serving
MSS2 of the new MH1 and sends
" MH1 registration ACK
message" to the new MSS2
7- The old MSS1 sends a
"MH1 cancellation ACK
message" to HLA
Handoff Processes
Step-71: MH12 neither sends nor receives any message within
the current cell. (MH12 breaks its connection with
MSS2).
Step-72: MH12 sends a "New Connection with MSS1" message
to MSS2.
Step-73: MSS2 sends a "MH12 New Connection" message to
MSS1.
Step-74: MSS1 sends a "Update MH12's Location" message to
HLR.
Step-75: HLR updates its record indicating the future serving
MSS1 of the new MH12. Then, HLR sends "MH12
Registration ACK" message to the new MSS1.
Step-76: HLR sends a "MH12 Registration Cancellation"
message to the old MSS2.
Step-77: The old MSS2 sends a "MH12 Cancellation ACK"
message to HLR.
Step-78: MSS2 asks MH12 to start handoff process to MSS1.
Step-79: MH12 performs handoff process from cell-2 to cell-1.
Step-710: MH12 sends "location registration" message to MSS1 at
cell-1. Then, MH12 asks MSS1 to modify its location list.
Step-711: MH12 completes the execution of its current task.

A GSPN model for describing the handoff process that can be occurred in the COBRA network
 describes the handoff process
of MH12 from cell-2 of MSS2
to cell-1 of MSS1
THandoff = TIstate / πinitial
where
THandoff : is the time of traveling the
token from a place p1 throughout the
handoff model until the token returns
to this place.
πinitial : is the steady state probability of
being in the initial state (marking) M0.
Initial marking represents the state that
a place p1 contains a token (s).
TIState : is the mean time spent in the
initial state M0.
p1
t1
●●●●
p2p3
p4
p5p6p7p8p10 p9
p14
p13p12
p11
t2t3t4t5
t6t7
t11
t10t9t8
Step1Step2Step3
Step4
Step5
Step6Step7
Step8 Step9
Step10
Step11
MSS2 MH12MSS1
HLA
Cell2  Cell1 MH12

Refinement the GSPN model of Handoff process
Handoff
Place
Handoff
Transition
THandoff
PHandoff
From the most of available reduction methods, we have chosen one structure reduction technique
that is suitable to reduce the structure of the GSPN handoff model. We reduce the structure of the
GSPN handoff model. Accordingly, this GSPN model can be refined to one place PHandoff and one
transition THandoff as depicted. The time associated with THandoff models the "mean time spent in the
GSPN handoff model".

GSPN model that describes how to incorporate the concise handoff process model of handoff process
(represented by p26 and the time associated with t20) into the GSPN model of execution processes of
M parallel tasks.This model describes the handoff process of MH11 to MSS2, MSS3, or MSS4.
MSS2
MH11
MSS4 MSS3
p16 p13 p10 p19
t10
t17t18
t19
p26
t20
p23
t14
p24
Step7
MH11 may complete
Execution of Task1 (t10)
or Perform Handoff (t17,
t18, or t19)
MH12 Complete
Execution of Task2
Handoff Time
between Two Cells
t17: MH11 Perform
Handoff to MSS2
OR
t18: MH11 Perform
Handoff to MSS3
OR
t19: MH11 Perform
Handoff to MSS4
Handoff Process
between Two Cells
Step8,
Step9
•This model describes how to
incorporate the concise handoff process
model (represented by p26 and the
time associated with t20) into the Petri
net model of program execution

GSPN model that describes how to incorporate two concise handoff process models such
asThe model describes the handoff process of MH11 to MSS2, MSS3, or MSS4., andThe
model describes the handoff process of MH14 to MSS1, MSS2, or MSS3. into the GSPN
model of the execution processes of M parallel tasks .
MSS2
MH11
MSS4 MSS3
p16 p13 p10 p19
t10
t17t18t19
Step7
MSS1
MH14
MSS3 MSS2
p13 p10 p7 p22
t13
t21t22t23
Step7
p23
t14
p24
Step8,
Step9
p26
p27
t20
t24
t25
3
44
4
3
2
4
Handoff Time
between Two Cells
Handoff Time
between Two Cells
MH14 may
Perform
Handoff
(t21, t22, or t23)
MH11 may
complete
Execution of
Task1
MH11 may Perform
Handoff (t17, t18, or t19)
MH14 may
complete
Execution
of Task4
Two Handoff
Process occur at
the same time
Handoff Process -I Handoff Process -J

MTCP without the occurrence of handoff process and the
handoff time of one handoff process under the effect of
distance among the active network resources
•MTCP time & Handoff process time increase as the distance D
(D=200, 400,…, 1000 meters) among the resources increases.
0
500
1000
1500
2000
2500
3000
3500
4000
2004006008001000
PerformanceTime(ms)
DISTANCE between resources(meter)
MTCP
HO time

MTCP with the occurrence of one handoff process and MTCP with the
occurrence of two handoff processes under the effect of distance
among the active network resources
•The time of MTCP time with two handoff processes larger than The time of MTCP time
with one handoff process.
0
500
1000
1500
2000
2500
3000
3500
4000
2004006008001000
PerformanceTime(ms)
MTCP with one HO
occurence
MTCP with Two HO
occurence

(i). Only handoff time THandoff,
(ii). Only performance time MTCP,
(iii). Performance time MTCP under the
occurrence of one handoff process,
(v). Performance time MTCP under the
occurrence of two handoff processes
Study the relation between the performance time (MTCP or
THandoff) and the distance among the network resources in the
following cases:
0
500
1000
1500
2000
2500
3000
3500
4000
2004006008001000
PerformanceTime(ms)
Distance between resources (meter)
MTCP without HO
occurence
HO process
MTCP with one HO
occurence
MTCP with two HO
occurence

How to use Petri net tool to create reliability model to
study the reliability of wireless CORBA according to the
main concept of End to End Instantaneous reliability EIR(t)
?????

A generic CORBA performance model for giving concise description for the
dynamic behavior of Petri net models of Program Execution and under one
handoff process.
p1
t1 p2
p3
p4
t2
t3
t5
●
t4
Start Program
End Program
Handoff
Handoff INHandoff OUT
Program
Available
Program
Processing
May Perform
Handoff
t1 = t2= TMTCP (Figure 4.1)/3 + D/2
t3 = t4= THandoff (Figure 4.2)/2 + D/2
t5 = TMTCP (Figure 4.1)/3
For example: If D =1000m
t1 = t2= 2449/3 + 1000/2
t3 = t4= 1240/2 + 1000/2
t5 = 2449/3

MTCP without the occurrence of handoff process and the
handoff time of one handoff process under the effect of
distance among the active network resources
0
500
1000
1500
2000
2500
3000
3500
4000
2004006008001000
PerformanceTime(ms)
MTCP
HO time

The accuracy of the results obtained from the
generic reliability model
1000800600400200Distance D
2449204916491249846TMTCP of Figures 5.2 and 5.9
12401040840640440THandoff of Figures 5.4 and 5.9
33182777223616951154TMTCP with one handoff of Figures
5.5 and 5.10
33152770218515941002TMTCP with one handoff Obtained
from the generic performance model of
Figure 6.1
99.91%99.75%97.72%94.04%86.83%Accuracy results of the generic
model of Figure 6.1 compared to the
results of model of Figure 5.5

A GSPN model for illustrating the failure processes of one (e.g. MH, MSS, or HLA), two,
three, or four resources through the execution processes of the desired program under the
occurrence of one handoff process between two cells.
Handoff
Processing
p1
t1
p2
p3
p4
p7
p6
p5
p8
t2
t3
t5
t6
t7
t8
●
●
●
●
t9
t4
p11
p10
p9
HLA
 MH
 MSS
Start Program
End Program
Handoff
Program
Available
Program
Processing
Which
Resource
fails?
MH Fails
MSS Fails
HLA Fails
Available
Resources
Failure
Rate
May Perform
Handoff

A GSPN model for describing the failure processes of the various network
resources when one or more of MHs perform handoff processes between two
cells
p1
t1
p2
p3
p4
t2
t3
t5
t4
Start Program
End Program
Handoff
Program
Available
Program
Processing
Handoff
Processing
May Perform
Handoff
p7
p6
p5
t6
t7
t8
t9
p11
p10
p9
HLA
 MH
 MSS
Which
Resource
fails?
MH Fails
MSS Fails
HLA Fails
Available
Resources
Failure
Rate
●
●
●
●
p8

From Table 6.2
t1 = t2= t3 = t4= TMTCP (Figure 5.2)/2 + D/4
t9 = t10= TMTCP (Figure 5.2)/2 + D/3
t5 = t6= t7 = t8= THandoff (Figure 5.4)/3 + D/4
For Example: If D=1000m
t1 = t2= t3 = t4= 2449 /2 + 1000/4
t9 = t10= 2449/2 + 1000/3
t5 = t6= t7 = t8= 1240/3 + 1000/4
p1
t1
p2
p7
p3
t2t7
t9
t8
8
Start
Program
Program
Execution
Program Available
Through
Program
Processing
Handoff
Processing
May Perform
Handoff
t3
p4
End
Program
t10
Handoff IN - I
Handoff OUT - I
Handoff
Processing
p6
t5
t6
Handoff
t4
p5
Program
Execution
Handoff IN - J
Handoff OUT - J
Other Handoff
Process Occurred
Through
Program
Processing
●
May Perform
Handoff
A generic CORBA performance model for
giving concise description for the
dynamic behavior of Petri net models of
Program Execution and underTwo
handoff process.

The accuracy of the results obtained from the
generic reliability model
1000800600400200Distance D
2449204916491249846TMTCP of Figures 5.2 and 5.9
12401040840640440THandoff of Figures 5.4 and 5.9
36453051245718641270TMTCP with Two handoff of Figures
5.8 and 5.10.
36423011238417541122TMTCP with Two handoffs Obtained from
the generic performance model of Figure
6.6.
99.91%98.69%97.03%94.18%88.45%Accuracy results of the generic
model of Figure 6.6 compared to the
results of original model of Figures 5.8
and 5.10.

GSPN model for
studying the
reliability of COBRA
network
when it executes
the required
program through
the occurrence of
two
handoff processes.
Through the
program execution
may one, two,
three,
or four resources fail
p1
t1
p2
p7
p3
t2t7
t9
t8
Start
Program
Program
Execution
Program
Available
Through
Program
Processing
Handoff
Processing
May
Perform
Handoff
p8
p9
p10
t10
HLAMH MSS
Which Resource
fails?
Available
Resources
p14
p13
p12
MH Fail(s) MSS Fail(s) HLA Fails
Failure Rate
HLAMH MSS
Which Resource
fails?
Available
Resources
p21
p20
Failure
Rate
t3
p4
End
Program
t14
May
Perform
HandoffHandoff IN - I
Handoff OUT - I
Handoff
Processing
p6
t5
t6
Handoff
t4
t11 t12 t13
p5
p11
p15 p17
p16
p18
p19
t15
t16
t17 t18
Program
Execution
Handoff IN - J
Handoff OUT - J
Other Handoff
Process Occurred
Through
Program
Processing
●
●● ●
●● ●

GSPN model for
studying the
reliability of COBRA
network
when it executes the
required program
through the
occurrence of two
handoff processes.
Through the handoff
processes may be
one, two,
three, or four
resources fail.
p1
t1
p2p7
p3
t2t7
t9
t8
Start
Program
Program
Execution
Program
Available
Through
Program
Processing
Handoff
Processing May Perform
Handoff
t3
p4
End
Program
t14
May
Perfor
m
HandoffHandoff IN-I
Handoff OUT-I
Handoff
Processin
g
p6
t5
t6
Handoff
t4
p5
Program
Execution
Handoff IN-J
Handoff OUT-J
Other Handoff Process
Occurred through
Program Execution
Through
Program
Processing
p8
p9
p10
t10
HLAMH MSS
Available
Resources
p12
p13
p14
Failure Rate
HLAMH MSS
Which
Resource
fail
Available
Resources
p19
p20
MH Fail(s)
Failure Rate
t13
t12t11
p11
p15 p17
p16
p18
p21
t15
t18t17t16
●● ●
●● ●
MSS Fail(s) HLA Fails
Which
Resource
fails?

EIR(t) reliability results in case of CORBA network executes its required
program through the occurrence of one handoff process.Through the
execution processes of the program may one, two, three, or four
resources failed.
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
1.02
200 400 600 800 1000
Distance (Meter)
ExpectedReliabilityEIR(t)
one resource failed
two resouces failed
three resources failed
four resources failed
•When the transmission range among the wireless MHs increases, the reliability results of the
network decrease
•When no. of failed resources increase, the reliability results of the network decrease

EIR(t) reliability results in case of CORBA network executes its
required program through the occurrence of one handoff process.
Through the handoff process may be one, two, three, or four
resources failed.
0.7
0.75
0.8
0.85
0.9
0.95
1
200 400 600 800 1000
Distance(Meter)
one resource failed
two resources failed
three resources failed
four resources failed
network decrease

EIR(t) reliability results in case of CORBA network execute its required
program through the occurrence of two handoff processes.Through the
program execution may be one, two, three, or four resources fail.
0.8
0.82
0.84
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
200 400 600 800 1000
Distance(Meter)
One resource failed
Three resources failed
Four resources failed
network decrease

EIR(t) reliability results in case of CORBA network execute its required
program through the occurrence of two handoff processes.Through the
handoff processes may be one, two, three, or four resources fail.
0.7
0.75
0.8
0.85
0.9
0.95
1
200 400 600 800 1000
Distance(Meter)
One resource failed
Three resources failed
Four resources failed
network decrease

Conclusion
• We have proposed novel and flexible GSPN framework to assess the
reliability of wireless networks based on the End-to-End Expected
Instantaneous Reliability (EIR) measure (Unlike traditional reliability
technique)
• We have built a GSPN performance model for describing the dynamic
behavior (e.g. services processes, and communication processes) of
wireless network resources under real workload (e.g. executing parallel
programming)
• We have built GSPN performance model for describing the occurrence of
handoff processes among the network cells. We have used the modeling
power of the Petri net technique to concise this handoff model.
• We have constructed a new generic GSPN reliability model consists of three
concise models: the execution program concise GSPN model, the handoff
concise GSPN model, and the reliability concise GSPN model that describes
the dynamic behavior of failure processes of network resources (e.g. MH,
MSS, HLA)

Conclusion
• The modeling power of the developed generic GSPN reliability model allows us to
study the reliability of the wireless network when the resources failures through
the execution program or/ and through the handoff processes.
• The main contribution of the developed GSPN reliability model can be illustrated
through the capability of studying some novel reliability metrics that incorporate
the effect of the mobility characteristic introduced by handoff processes, the
distance between the communicating wireless resources, the failure processes of
various network resources, and the main time to complete the desired program
that runs on the network resources
• Due to the modeling power of the developed GSPN model, we can easily describe
numerous number of handoff processes of wireless resources.
• From our investigation in the reliability area, we have found that it is very difficult
to describe such complicated aspects by analytical or simulation methods.

Reliability Study of wireless corba using Petri net and end to end instantenous reliability measure

More Related Content

Viewers also liked (13)

Similar to Reliability Study of wireless corba using Petri net and end to end instantenous reliability measure (20)

Recently uploaded (20)

Reliability Study of wireless corba using Petri net and end to end instantenous reliability measure

Editor's Notes