You are on page 1of 4

(IJCSIS) International Journal of Computer Science and Information Security,

Vol. 9, No. 7, July 2011



Recovery function of Components of Additive Model
of Biometric System Reliability in UML
Zoran osi(Author)
director
Statheros d.o.o.
Katel Stari, Croatia
zoran.cosic@statheros.hr
Jasmin osi (Author)
IT Section of Police Administration
Ministry of Interior of Una-sana canton

Biha, Bosnia and Hercegovina
jacosic@gmail.com
Miroslav Baa (Author)
professor
Faculty of Organisational and Informational science
Varadin, Croatia
miroslav.baca@foi.hr


Abstract- Approaches The development of biometric systems is
undoubtedly on the rise in the number and the application areas.
Modelling of system reliability and system data analysis after
failure and the time of re-establishing the operating regime is of
crucial importance for users of the system and also for producers
of certain components. This paper gives an overview of the
mathematical model of biometric system function recovery and its
application through the UML model.

Keywords- Additive model, Biometric system, reliability, recovery
function, UML, component,
I. INTRODUCTION

Many models of reliability of biometric systems are applicable
only to specific parts or components of that same system. For
more complex considerations must be taken into account
models based on Markov processes that consider the reliability
of the system as a whole, which includes components of the
system. In this paper the approach to restoring the functions of
a biometric system that had failure at some of its components is
elaborated. The basic model is an additive model which
assumes a serial dependence between the components [1] (Xie
& Wohlin).
UML is also becoming standard in the process of system
design so the manufacture of component systems greatly
benefits from the UML view. The authors introduce the
concept of UML modelling in the process of restoring function
analysis of biometric systems. The paper defines the conceptual
class diagram in UML, which provides a framework for
analyzing the function recovery of biometric systems.
II. ADDITIVE RELIABILITY MODEL

Reliability [2] as the probability [2] (number between 0 and 1
or 0% and 100%) can be represented as a ratio between the
number of successful tasks and the total number of tasks in the
time specified for the operation of the system:

1
( )
( )
( )
n t
R t
n t
=
(1)

where :
( ) R t
- assessment of reliability,
1
( ) n t
- number of successful assignments in time t,
( ) n t
- total number of tasks performed in time t,
t - time specified for the operation of the
system.

The value
( ) R t
represents the estimated reliability due to the
fact that the number of tasks n(t) is final. Therefore, the actual
reliability R(t) is obtained when the number of tasks n(t) tends
to infinity.

( )
( )
( ) lim R t
n t
R t

=
(2)
R ( t ) = 1 F ( t ) = P ( T > t ) (3)

Where the R ( t ) 1indicates the reliability function. Thus F ( t )
can be called non reliability function. Approximate form of the
function F ( t ), is shown in Figure 1. It is a continuous and
monotonically increasing function:

F ( 0 ) = 0
F ( t ) 1, when t

Density failure function is marked with F( t ), and from
probability theory we know that:


1
R ( t ) is function of reliability
1 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 7, July 2011

( )
( )
dF t
f t
dt
=
(4)
where F(t) is probability distribution function.
Failure intensity [3], [4], ( t ) represents the density of
conditional probability of failure at time t provided that until
that moment there was no failure.

) (
) (
) (
t R
t f
t =
(5)
Or according to the model of Xie and Wohlin:
0 ,
) (
) ( > = t
dt
t d
t

(6)
where (t) is mean value of the expected system failure.

It is also assumed that the intensity of the failure of the entire
system is the sum of the intensity of failures of its components:
(7)

So it follows that the expectations of failure of the system are
(6):

(8)

III. BIOMETRIC SYSTEM RECOVERY FUNCTION
Term recovery consider biometric system as a system that is
maintained after a long period of use or recovered after failure of
particular components. Biometric system components, after the
failure, are maintained or exchanged and then continue to be part of
the system. When considering the reliability problems of generic
biometric system along with a random event that includes the
appearance of failure within the system, it is necessary to consider
other random event and that is recovery the system after failure.
To this event corresponds a new random variable

that indicates
the time of recovery. As a characteristic of random variable


indicators similar to those being considered for the analysis of time
without failure are used.

A. Distribution recovery function, refresh frequency

is a random variable [3], [4] which marks the time of recovery


of the components in failure, then the probability of recovery is as a
function of time:

P ( < ) ( )
1
t F t = (9)

F
1
(t) probability distribution function of random variable

.
The probability of non-recovery G(t) is defined as:

P t G = ) ( ( > t) = ) ( 1
1
t F (10)


Refresh frequency
) (
1
t F
is the probability density function of
random variable

:

dt
t dG
dt
t dF
t f
) ( ) (
) (
1
1
= = (11)
From here it follows that:

) (
1
t F dt t f
t
) (
0
1

= (12)

dt t f t G
t
) ( 1 ) (
0
1

= (13)

B. Intensity of recovery function

) (t
is the conditional probability density function2 of completion
of recovery of components (repair) within time t, provided that
recovery is not completed until the moment t.
Intensity recovery function is conditional probability density
function of the end of the recovery in time t, provided that recovery
is not complete until that moment t, we have:

1 1
( ) ( ) ( )
( )
( ) ( ) ( )
F t f t G t
t
G t G t G t

'
'
= = = (14)

0 0
( )
( )
( )
t t
dG t
t dt
G t
=

(15)
0
0
ln ( ) ( )
t
t
G t t dt =

(16)
( )
0
( )
t
t dt
G t e

= (17)
0
( )
1
F (t)= 1- e
t
t dt

(18)



2 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 7, July 2011


C. Time recovery function
a. Mean recovery time

Mean time of recovery
M
(

) is the mathematical expectation of


random variable

whose probability density function is


) (
1
t f
,
i.e.
M ( ) dt t tf ) (
0
1

= (19)
M ( ) dt t G

=
0
) ( (20)

b. Recovery time variance

Recovery time variance
2
0

is characterized by deviation of
duration of recovery

from his mean recovery time.




2
0
| | | |
2
0 0
2
) ( ) ( 2 ) (

= = = dt t G dt t tG E E V
(21)

c. Availability of system after recovery time

Probability [5], that the system after time t will be available for
functioning is the expression (10).

Where intensity recovery function can be defined as:

(22)

MTTR mean time to repair

The process of transition from the state of failure to the state of
availability can be represented as in Figure 1:


Figure 1


Condition 1 represents a functional system and condition 2
represents a system that has been repaired after a failure.
The transition of system from condition 1 to 2 is represented
with failure intensity function , the transition from condition 2
to condition 1 is defined with recovery intensity function .
IV. RECOVERY FUNCTION OF BIOMETRIC SYSTEM IN UML
A. Generalized biometric system

Generalized biometric system model[7], [8], [9] is a schematic view
of Wyman biometric system model that depicts serial dependence
of a system components and can be summarized, in this exploitation
period of time, as shown on Figure 2.

Figure 2

The system shown in Figure 2 works in time t0 without failure.

After the failure the system is recovered in time t1, after recovery
occurs time period of re-operation t2.
Parameter which defines the conditions created by failure is
intensity of failure of particular component .

The intensity of the component failure can be expressed as:

1
1
EL
n n

O
= =
O

(23)
Where is:

n- number of correct parts of the confidence interval
(1 ) 0, 75 =

O
- lower limit of confidence for the mean time between failures.

Recovery time of the system is the function of the recovery
intensity as described by the expression (22).

B. The conceptual class-diagram model of system
recovery

During the study [8] of the problem of reliability of generic
biometric system, object-relational approach of description of the
problem provides easier and clearer description of the sequence
analysis of events within the system during the verification of
failure.
Figure 3 shows the diagram of classes of the recovery of biometric
system model:
3 http://sites.google.com/site/ijcsis/
ISSN 1947-5500
(IJCSIS) International Journal of Computer Science and Information Security,
Vol. 9, No. 7, July 2011


Figure 3

Class Biometric system is a set of components of that system
and is in relation to class Failure which contains data on the
Component in failure, time of occurrence of failure and failure
intensity.
Class Recovery is in relation to class Biometrical system
because it contains information about the component, the time
of recovery of component and calculated recovery intensity of
component. Class Recovery is in relation to the class
Availability, which is a function of data on failure intensity and
the recovery intensity, with the class Mean time which contains
data of recovery start time, duration and results of recovery,
with the class Recovery intensity. Furthermore it is possible, at
the level of class diagrams to present and other factors of
reliability and facilitate access to their prediction based on
historical data (logs) of the system functioning.
V. CONCLUSION AND FURTHER RESEARCH
Information about the system failure must be considered in the
context of the whole biometric system and its performance in
time.
In accordance with the above information on the exploitation
of biometric systems must be part of a comprehensive analysis
of the functioning and also information on recovery of the
system and its functionality at any given time. The time to put
the system into operation condition is often placed in clearly
defined time frames that are stipulated in contracts or SLA
addenda to the contract. The parameters monitoring processes
associated with the reliability of the system are often
complicated and laborious so UML approach to description of
problem simplifies the same. UML also imposes as general or
universal standard for descriptions of appearance.
Further work of the authors will be directed toward
specialization of model taking into consideration the other
models of reliability dependence and different system failure
probability distributions.

REFERENCES

[1] An additive reliability model for the modular software failure data
M.Xie, C.Wohlin - 2007
[2] Teorija pouzdanosti tehnikih sistema, Vojnoizdavaki novinski centar,
Beograd 2005,
[3] Pouzdanost brodskih sustava Ante Buka, Ivica egulja Pomorstvo -
2008
[4] Pouzdanost tehnikog sustava brodskog kompresora Zoran osi
magistarski rad - 2007
[5] Eksploatacija i razvitak telekomunikacijskog sustava, Mr.sc Juraj Buzoli
, Split 2006
[6] Zasnivanje otvorene ontologije odabranih segmenata biometrijske
znanosti - Markus Schatten Magistarski rad FOI 2007
[7] Early reliability assessment of UML based software models Vittorio
Cortellessa, Harshinder Singh, Bojan Cukic WOSP02 , July 24-26,
2002 Rome Italy
[8] Modelling biometric systems in UML Miroslav Baa, Markus
Schatten, Bernardo Golenja, JIOS 2007 FOI Varadin
[9] Reliability, Availability and Maintainability in Biometric Applications
2003-2007 Optimum Biometric Labs A WHITE PAPER Version r1.0,
Date of release: January 2, 2008, SWEDEN

AUTHORS PROFILE
Zoran osi, CEO at Statheros ltd, and business consultant in business process
standardization field. He received BEng degree at Faculty of nautical
science , Split (HR) in 1990, MSc degree at Faculty of nautical science ,
Split (HR) in 2007 , actually he is a PhD candidate at Faculty of
informational and Organisational science Varadin Croatia. He is
a member of various professional societies and program
committee members. He is author or co-
author more than 20 scientific and professional papers. His main
fields of interest are: Informational security, biometrics and privacy,
business process reingeenering,
Jasmin osi has received his BE (Economics) degree from University of
Biha, B&H in 1997. He completed his study in Information Technology
field (dipl.ing.Information Technlogy) in Mostar, University of Demal
Bijedi, B&H. Currently he is PhD candidate in Faculty of Organization
and Informatics in Varadin, University of Zagreb, Croatia. He is
working in Ministry of the Interior of Una-sana canton, B&H. He is a
ICT Expert Witness, and is a member of Association of Informatics of
B&H, Member of IEEE and ACM. His areas of interests are Digital
Forensic, Computer Crime, Information Security and DBM Systems. He
has presented and published over 20 conference proceedings and journal
articles in his research area
Miroslav Baa is currently an Associate professor, University of Zagreb,
Faculty of Organization and Informatics. He is
a member of various professional societies and program
committee members, and he is reviewer of several international
journals and conferences. He is also the head of the Biometrics centre in
Varadin, Croatia. He is author or co-
author more than 70 scientific and professional papers and two books.
His main research fields are computer forensics, biometrics and privacy
professor at Faculty of informational and Organisational science
Varadin Croatia



4 http://sites.google.com/site/ijcsis/
ISSN 1947-5500