Evaluating Quality of Web Services: A Risk-driven Approach

fizzlargeΑσφάλεια

3 Νοε 2013 (πριν από 3 χρόνια και 9 μήνες)

80 εμφανίσεις

26/04/2007

BIS'07 Poznan, Poland

1

Evaluating Quality of Web Services:


A Risk
-
driven Approach

Natallia Kokash

Vincenzo D’Andrea

26/04/2007

BIS'07 Poznan, Poland

2

Introduction


Service
-
centric systems


Quality of Service (QoS) Issues


QoS
-
driven service selection


Risk
-
driven service selection


Risk analysis


SOA risks


Failure risk


Experimental results


Conclusions and Future Work


Risk management for SOA


References

26/04/2007

BIS'07 Poznan, Poland

3

Service
-
centric systems

s
2

s
1

+

+

s
3

s
5

+

s
4

+

Client

Provide
r

Partners

s
0

Sequential operator

s
1
; s
2

Parallel operator

s
1
| s
2

Choice operator

s
1
+ s
2

Web service

s
i

Start state

t
0

End state

t

s
i

+

|

26/04/2007

BIS'07 Poznan, Poland

4

Quality of Service Issues


QoS for web services:


Domain
-
independent


Throughput, capacity, latency, response time (duration), availability,
reliability, reputation, execution cost (price)


Domain
-
dependent


Currency converters: accuracy


Hotel booking: prices, number of the rooms, availability rate



How to:

1.
specify QoS?

2.
measure QoS?

3.
specify user requirements and/or preferences about QoS?

4.
match user requirements with existing services in terms of QoS?

5.
rank services according to user preferences?

6.
predict QoS factors under certain environmental conditions?

7.
choose web services to guarantee certain QoS level of their
composition?

26/04/2007

BIS'07 Poznan, Poland

5

QoS
-
driven service selection


Problems in quality
-
driven service selection:


Lack of QoS statistics


Volatility of QoS factors


Multidimensionality


Subjectivity


Context
-
dependence



Approaches


Multi
-
attribute optimization
[Ardagna and Pernici 2005, Zeng et al. 2004,
Yu et al. 2005 ]


Constraints satisfaction
[Martin
-
Diaz et al. 2005]


Genetic algorithms
[Canfora et al. 2006]


Fuzzy
[Lin et al. 2005]



Problems with existing approaches


Simplified models (e.g., one service for one task)


Dependences among QoS factors are ignored


Context is not taken into account

26/04/2007

BIS'07 Poznan, Poland

6

Risk analysis


Example:


Movie:

title=
Rainmaker
, format=
DVD
, languages=
Italian,
English


Convert DVD to AVI:

language=
English


SimpleDivX converter
: time=
2 hours
, language =
Italian


Impact on time:

2 hours are lost


Reason:

Unexpected service behaviour

(discrepancy with
specification)



Requires assessment of inherently uncertain events
and circumstances


Two dimensions:


how likely the uncertainty is to occur (probability)


what the effect would be if it happened (impact)

26/04/2007

BIS'07 Poznan, Poland

7

SOA Risks



Threats


Loss of service, data, users


Unexpected service behavior, changes


Performance problems


Contract violation


Assessment


Likelihoods and implications of threats


Analysis of user expectations


Service testing


User feedback, reputation systems


Mitigation


Service selection, redundancy, redesign


Runtime monitoring


Contracts and policies


26/04/2007

BIS'07 Poznan, Poland

8

Risk management for SOA

Information gathering
(
service discovery
,
QoS data
)
Decision making
(
are risks acceptable
?
)
Analysis of system artifacts
(
service testing
,
conformance evaluation
)
System configuring
(
service selection
,
adaptation
,
composition
,
...)
Risk identification
(
Business risks
,
technical risks
)
Risk analysis
Risk prioritization
Risk mitigation
(
policies
,
SLAs
,
...)
No
Redesign
loop
Fix configuration
Define controls
Yes
Evolutional
loop
History
26/04/2007

BIS'07 Poznan, Poland

9

Risk
-
driven service selection

Loss function



defines the cost of service failure

(money, time, resources)

Choose the composition that maximizes the
expected profit:

26/04/2007

BIS'07 Poznan, Poland

10

Failure risk


probability that some fault occurs


resulting impact of this fault on the
composite service

where is the

probability of the service failure
.


Loss function includes:


Expenses to invoke failed service (its cost and response
time)


Service failure can cause rollback of the transaction,
therefore expenses to execute precedent services are
also included


The provider may have to pay penalty to a user whose
request was not accomplished.

26/04/2007

BIS'07 Poznan, Poland

11

Failure risk of service
compositions

b
-
g

b
-
e

+

+

g
-
t

e
-
t

+

g
-
e

+

b
-
g

b
-
e

+

+

g
-
t

e
-
t

+

g
-
e

+

26/04/2007

BIS'07 Poznan, Poland

12

Failure risk: examples

Success rate = 0.5; execution cost = 1; penalty = 2

1.75

1.375

1.5625

1.625

1.25

s
1

s
2

s
1

s
2

s
3

+

+

s
4

s
2

s
1

s
3

+

+

s
1

s
3

+

+

s
2

s
4

+

+

s
1

s
2

s
3

+

+

26/04/2007

BIS'07 Poznan, Poland

13

Risk
-
driven selection algorithm


Select an execution path with minimum risk value


Notation:


c


composition


q(s
i
)


quality parameter (response time, execution cost)


p(s
i
)


probability of success


q
max



resource limit


Objective function:

where

26/04/2007

BIS'07 Poznan, Poland

14

Experimental results (1)


Goal:

Compare QoS of compositions
chosen by our algorithm with QoS of
compositions chosen by other methods


Zeng et al. [2004]


QoS factors:

price, duration, reputation, success
rate, availability



Objective function:

linear combination of scaled
QoS factors


Scaling:
QoS factors range from 0 to 1


Weights reflect user preferences

26/04/2007

BIS'07 Poznan, Poland

15

Experimental results (2)


100 simulated service compositions


10 services in each composition

26/04/2007

BIS'07 Poznan, Poland

16

Conclusions and Future work


A novel risk
-
based method for assessing QoS
of web services is proposed



Real world case studies


Comparative analysis of existing service
selection algorithms


Risk management framework for automatic
web service compositions



Questions?

26/04/2007

BIS'07 Poznan, Poland

17

References

1.
[Ardagna and Pernici 2005] Ardagna, D., Pernici, B.: ”Global and Local QoS
Constraints Guarantee in Web Service Selection,”
IEEE International Conference on
Web Services
, 2005, pp. 805

806.

2.
[Canfora et al. 2006] Canfora, G., di Penta, M., Esposito, R., Villani, M.
-
L.: “QoS
-
Aware
Replanning of Composite Web Services”,
Proceedings of the International Conference
on Web Services
, 2005.

3.
[Claro et al. 2005] Claro, D., Albers, P., Hao, J
-
K.: “Selecting Web Services for Optimal
Composition”,
Proceedings of the ICWS 2005 Second International Workshop on
Semantic and Dynamic Web Processes
, 2005, pp. 32
-
45.

4.
[Gao et al. 2006] Gao, A., Yang, D., Tang, Sh., Zhang, M.: “QoS
-
driven Web Service
Composition with Inter Service Conflicts”,
APWeb: 8th Asia
-
Pacific Web Conference
,
2006, pp. 121


132.

5.
[Lin et al. 2005] Lin, M., Xie, J., Guo, H., Wang, H.: “Solving QoS
-
driven Web Service
Dynamic Composition as Fuzzy Constraint Satisfaction,
IEEE International Conference
on e
-
Technology, e
-
Commerce and e
-
Service
, 2005, pp. 9
-
14.

6.
[Martin
-
Diaz et al. 2005] Martin
-
Diaz, O., Ruize
-
Cortes, A., Duran, A., Muller, C.: ”An
Approach to Temporal
-
Aware Procurement of Web Services”,
International
Conference on Service
-
Oriented Computing
, 2005, pp. 170

184.

7.
[Zeng et al. 2004] Zeng, L., Benatallah, B., et al.: ”QoS
-
aware Middleware for Web
Services Composition”,
IEEE Transactions on Software Engineering
, Vol. 30, No. 5, 2004,
pp. 311

327.

8.
[Yu et al. 2005] Yu, T., Lin, K.J.: ”Service Selection Algorithms for Composing Complex
Services with Multiple QoS Constraints”,
International Conference on Service
-
Oriented
Computing
, 2005, pp. 130

143.