Software Systems Engineering - nessi

crookpatedspongySoftware and s/w Development

Dec 2, 2013 (3 years and 9 months ago)

66 views




Part 1:

Overview of S
-
Cube


Part 2:

Service engineering research in S
-
Cube and links with
industrial use cases


Part 3:

Collaboration opportunities and wrap
-
up



Outline



Details on the S
-
Cube
Integrated Research
Framework


Research Areas


Service Engineering


Cross
-
layer Adaptation


End
-
to
-
end Quality Provision


Case
studies: a methodological
approach





Details on
the

IRF





S
-
Cube Integrated Research Framework


S
-
Cube focuses on long
-
term research


Main research focus:


Software service and systems


Adaptivity and evolution

-
of services

-
of agile service networks



S
-
Cube developed an Integrated Research Framework


4 views on research issues



S
-
Cube evolved a methodology for case studies
documentation




Integrated Research Framework Views


Conceptual Research Framework


Reference Life Cycle


Logical Run Time Architecture


Logical Design Environment





Integrated Research Framework

Conceptual Research Framework

Quality Definition, Negotiation & Assurance
(SQDNA)
Business
Process
Management
(BPM)
Service
Composition
& Coordination
(SCC)
Service
Infrastructure
(SI)
Design
Specifications
Local Design
Capabilities
Design
Specifications
Local Design
Capabilities
Design
Specifications
Local Design
Capabilities
Local A&M
Capabilities
Local A&M
Capabilities
Local A&M
Capabilities
A&M
Specifications
A&M
Specifications
A&M
Specifications
Adaptation and Monitoring Specifications
Integrated Adaptation and Monitoring Capabilities
Engineering
and
Design
(SED)
Adaptation
and
Monitoring
(SAM)
Local QDNA
Capabilities
QDNA
Specifications
Integrated QDNA
Capabilities
QDNA
Specifications
QDNA Capab.
QDNA Spec.
Monit. Capab.
Monit. Spec.
©

S
-
Cube


6



©

S
-
Cube


7

Integrated Research Framework

Reference Life
-
Cycle

Identify
adaptation
need
Identify
adaptation
strategy
Enact
adaptation
Early
requirements
engineering
Requirements
engineering
&
design
Construction
Deployment &
provisioning
Operation &
management


Integrated Research Framework

Logical Run
-
Time Architecture

Communication backbone
Software
service
Service
Container
Human
provided
service
Human
Service
Interface
Monitoring
Engine
Adaptation
Engine
Monitoring
logic
Resource
Resource
Broker
Discovery
and
Registry
Infrastr
.
Service
composition
Resource
Monitoring
logic
Negotiation
Engine
Negotiation
logic
Run
-
time QA
Engine
Test cases,
run
-
time
models, …
`
Application
-
specific
components
Core services
Management
interface
©

S
-
Cube


8



Integrated Research Framework

Logical Design Environment

Business
Process
Management
Service
Composition &
Coordination
Service
Infrastructure
SC&C
Modellers

BPEL
Modelling
Transformation
and Generation
Service
Modellers
BPM
Modellers

ASN

Choreography
(BPMN)
BPM ↔ SC&C
transformation

ASN+BPMN ↔ Service
Choreography

Service
Choreography
↔ Orchestration
SLA
Modellers
Quality
properties
translformation

PPM ↔
QoS
Deployment
KPI
Modeller
PPM
Modeller
QoS Modeller
Service
Compositions
Deployment
Service
Deployment
ASN and BPMN
Deployment
Verification
Monitoring transformations
and code generation
Adaptation transformation and code
generation
A&M
Modeller
A&M
Modeller
A&M
Configurators
SC&C Monitoring
& Adaptation
Deployment
BPM
Verification
SC
Verification
Quality
properties
translformation

KPI ↔
PPM
BPM Monitoring
Deployment
©

S
-
Cube


9





Research Areas in Service
Engineering



www.s
-
cube
-
network.eu

Service
Engineering


for

Service
-
Based

Applications

(SBA)



Adaptation & Monitoring

(SAM)

Business

Process

Mgt. (BPM)

Composition &

Coordination

(SCC)

Infra
-

structure

(SI)

Quality Definition, Negotiation & Assurance

(SQDNA)

©

S
-
Cube


3/#

Adaptation & Monitoring

(SAM)

Engineering & Design

(SED)



Adaptation design


Focus on activation of adaptation strategies


Instance
-
level adaptation


Context


Model
-
based



Main Ingredients of an
Adaptable

SBA [6]

Bucchiarone, C. Cappiello, E. di Nitto, R. Kazhamiakin, V. Mazza, and M. Pistore,

“Design for adaptation of Service
-
Based applications: Main issues and requirements,” in WEOSA 2009



Life
-
cycle for SBAs

WESOA, 2009



Adaptation Strategies


To mantain aligned the application behaviour with the context
and system requirements


Service substitution


Re
-
execution


(Re
-
)negotiation


(Re
-
)composition


Compensation


Log/Update adaptation Information


Fail



WESOA, 2009



Adaptation Triggers


The adaptation may be motivated by a variety of triggers


Component Services


Service functionality


Service quality


SBAs context


Business context


Computational context


User context

WESOA, 2009



Adaptation Strategies & Triggers


To re
-
align the application within the system and/or context
requirements


Each trigger can be associated with a set of adaptation
strategies

WESOA, 2009



Design Guidelines for triggers and adaptation
strategies


Design adaptable SBAs implies relate adaptation triggers and
adaptation strategies together


Modeling adaptation triggers


Realizing adaptation strategies


Associating adaptation strategies to triggers


Design approaches


Built
-
in adaptation

-
Adaptation needs and adaptation configuration known a priori


Abstraction
-
based adaptation

-
Adaptation need fixed, but adaptation configuration not known a
priori


Dynamic adaptation

-
Adaptation needs not known at design time or cannot be
enumerated


WESOA, 2009



Designing reliable

service compositions




Focus on service compositions


To design more reliable service
-
based processes


inserting monitors, adaptation strategies, changing process structure,



Which is the best choice?


Context
-
awareness (user
-
dependent)


Based on quality evaluation


Cappiello, C.; Pernici, B.:
QUADS: Quality
-
Aware Design of dependable Service
-
based
processes
. 2010



Cappiello, 2010



Preventive and Corrective strategies

Cappiello, 2010




Quality evaluation



According to quality evaluation techniques for service compositions



Representing users



Importance of each user


Importance of quality dimensions for each user

Cappiello, 2010



Comparing strategies

Domain specific!



Evaluating alternative strategies for a service

Example: Redundancy selected



additional quality constraints for redundant services
are derived at design time



basis for service selection at run time

www.s
-
cube
-
network.eu

Cross
-
Layer

Adaptation



Adaptation & Monitoring

(SAM)

Business

Process

Mgt. (BPM)

Composition &

Coordination

(SCC)

Infra
-

structure

(SI)

Engineering & Design

(SED)

Quality Definition, Negotiation & Assurance

(SQDNA)

©

S
-
Cube


3/#

Adaptation & Monitoring

(SAM)

R.Kazhamiakin / WP
-
JRA
-
1.2


Month 24 Review Meeting

WP Vision

Key Challenges

©

S
-
Cube


4/#


Comprehensive and integrated adaptation and monitoring
principles, techniques, and methodologies


Across SBA layers, across SBA boundaries, across SBA life
-
cycle


Context
-

and HCI
-
aware A&M


Improve SBA adaptation based on the contextual knowledge


Mixed initiative SBA adaptation


From self
-
adaptation to human
-
in
-
the
-
loop adaptation

WP Vision

Integrated A&M Framework

R.Kazhamiakin / WP
-
JRA
-
1.2


Month 24 Review Meeting

Monitoring
mechanisms

Adaptation
mechanisms

Monitored
events

Adaptation
requirements

Adaptation
strategies

detect

trigger

achieve

realize

Conceptual Model
:


A&M taxonomy

Instantiations
:


Cross
-
layer A&M


Proactive adaptation


HCI and context
-
awareness


Self
-
adaptation

Conceptual architecture
:


Results from
SotA


Research results of S
-
Cube members

©

S
-
Cube


5/#

R.Kazhamiakin / WP
-
JRA
-
1.2


Month 24 Review Meeting

Monitoring
mechanisms

Adaptation
mechanisms

Monitored
events

Adaptation
requirements

Adaptation
strategies

Cross
-
layer integrated
monitoring mechanisms

Cross
-
layer integrated and
coordinated adaptation mechanisms

Means to identify
adaptation needs
across layers

Means to identify
adaptation strategies
across layers

Event
propagation and
alignment

Adaptation
coordination

Adaptation
effectiveness

Adaptation
compatibility
and integrity

Key Results Achieved

Cross
-
layer Adaptation and Monitoring

REQUIREMENTS

©

S
-
Cube


9/#

R.Kazhamiakin / WP
-
JRA
-
1.2


Month 24 Review Meeting

©

S
-
Cube


31
/#

Cross
-
layer Adaptation and Monitoring

Integrated Monitoring Mechanisms


Results


Unify monitoring of business properties and low
-
level service properties
centered around service compositions at run
-
time
[1]

-
Instance and class properties in the same model (ASTRO)

-
Rich notation for basic events and probes (Dynamo)


Integrate run
-
time and design
-
time events monitoring
[2]

-
Monitoring of run
-
time properties and model changes (EMF events, from design
environment) in the same framework (
WildCat

model for monitored data)


Multifactor Monitoring
[3]

-
Unified language for querying various factors: service behavior, service quality,
service context, service structure

-
Formal model for the representation of queries: event calculus

-
Automated translation rules for different factors

1.
Baresi
, Guinea,
Pistore
,
Trainotti
.
Dynamo+ASTRO
: an Integrated Approach for BPEL Monitoring. In
ICWS 2009

2.
Morin,
Ledoux
, Ben
Hassine
,
Chauvel
,
Barais
,
Jezequel
.
Unifying Runtime Adaptation and Design
Evolution

In CIT 2009

3.
Zisman
,
Spanoudakis
,
Mahbub
. A Monitoring Approach for Run
-
time Service Discovery. Under
submission.

R.Kazhamiakin / WP
-
JRA
-
1.2


Month 24 Review Meeting

©

S
-
Cube


32
/#

Cross
-
layer Adaptation and Monitoring

Identification of Adaptation Needs


Results


Domain assumptions
[1]

-
Explicit encoding of domain assumptions

-
Associate them to requirements

-
Monitor assumption violations

-
verify requirements at run time to trigger adaptation



Replacement policies
[2]

-
Explicit encoding of rules for service substitution

-
Take into account the SBA execution point

-
Consider variety of aspects:
QoS
, structural, behavior, context, changing
requirements

-
Take into account availability of new services


1.
Gehlert
,
Bucchiarone
,
Kazhamiakin
, Metzger,
Pistore
, Pohl: Exploiting Assumption
-
Based
Verification for the Adaptation of Service
-
Based Applications. SOA@SAC Conference, 2010

2.
Mahbub
,
Zisman
. Replacement Policies for Service
-
Based Systems. MONA+ workshop,
ICSOC/
ServiceWave

conference, 2009


Assumptions

Domain

Requirements

SBA


Non
-
functional req’s


Functional req’s


Composition (BPEL)


Service protocols


QoS models


External services


User context

R.Kazhamiakin / WP
-
JRA
-
1.2


Month 24 Review Meeting

©

S
-
Cube


33
/#

Cross
-
layer Adaptation and Monitoring

Identification of Adaptation Strategies


Results


Adaptation based on quality factor analysis

-
Identify influential factors for KPI violations across layers: data mining techniques

-
Identify adaptation requirements: analysis of the dependency trees

-
Associate adaptation actions to basic SBA metrics and properties

-
Identify adaptation strategy based on the effects of adaptation actions











1.
Kazhamiakin
,
Wetzstein
,
Karastoyanova
,
Pistore
,
Leymann
. SBA Adaptation
based on quality factor
analysis

. MONA+ Workshop, ICSOC/
ServiceWave

Conference 2009


Quality modeling
for analysis and
adaptation
Identification of
Adaptation
Strategies
Process
adaptation
Analysis of
influencing quality
factors
Metrics
Model
Adaptation
Requirements
Adaptation
Actions
Model
Adaptation
Actions
R.Kazhamiakin / WP
-
JRA
-
1.2


Month 24 Review Meeting

©

S
-
Cube


34
/#

Cross
-
layer Adaptation and Monitoring

Coordinated Adaptation


Results


Framework for self
-
supervising processes

-
Generic adaptation framework with uniform and
extendable language for adaptation coordination

-
Encoding of recovery actions

-
Encoding of coordinated strategies

-
Wide range of adaptation actions

-
Service
-
level: ignore, halt, retry, rebind

-
Process
-
level: change parameters, change
partner, restore, recovery
subprocess

-
Pluggable run
-
time architecture to
accommodate different adaptation types

-
AOP techniques for extending process engine
functionalities




1.
Baresi
, Guinea, Pasquale. Integrated and Composable Supervision of BPEL processes. ICSOC
Conference 2008

2.
Baresi
, Guinea. Self
-
supervising BPEL processes.
PoliMi

TR 74.2009

www.s
-
cube
-
network.eu

End
-
to
-
End

Quality

Provision



Adaptation & Monitoring

(SAM)

Business

Process

Mgt. (BPM)

Composition &

Coordination

(SCC)

Infra
-

structure

(SI)

Engineering & Design

(SED)

©

S
-
Cube


3/#

Adaptation & Monitoring

(SAM)

Quality Definition, Negotiation & Assurance

(SQDNA)

Motivational

„Scenario“

A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

response

time: 1 s

response

time: 2 s

response

time: 1 s

2. Pay

Renewal Fee

3. Update Vehicle
Record

4.a. E
-
Mail
Confirmation

1. Identify
Vehicle

[valid]

[not valid]

ePay

response time: 2 s

cost: 1.50


SecurePay

response time: 3 s

cost: 1


Yahoo

response time: 1.5 s

GMail

response time: 2 s

4.b. Mail Validation
Sticker





= service invocation/activity

= third
-
party service



= internal service



= alternative binding

RenewalHandler

TODO:

show

what

end
-
to
-
end
means
,

Negotiation

and
assurance

WP Vision

Key
Challenges

...Definition
[Mo1
-
45]

...Assurance
[Mo18

45]

...Negotiation
[Mo10

45]

Devise novel principles, techniques & methods for Quality...

D



End
-
to
-
End Quality Reference


Model (completed in Y1)



Rich and Extensible Quality


Definition Language

N



Exploiting HCI


knowledge for


automatic quality


contract establishment



Proactive SLA


negotiation and


agreement (from Y3)

A



Run
-
time Quality


Assurance Techniques



Quality Prediction


Techniques to Support


Proactive Adaptation

Key Results Achieved


Quality Reference Model


Problems


Understanding quality attributes across SOA layers


Solution


Quality Reference Model


S
-
Cube Publications

1.
NNN


A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

©

S
-
Cube


39
/<
Mx
>

A. Metzger / WP
-
JRA
-
1.3


Month 12 Review Meeting, Essen, April 2009

Key Results Achieved


Knowledge Modelling


S
-
Cube Quality Reference Model (QRM)


Model containing 77 quality attributes in 10 categories



䑥汩v敲慢汥e䍄
-
䩒A
-
ㄮ㌮2 孍漠ㄲ1㨠
“Quality reference model for SBAs”



















D

Quality
(77)
Security
(9)
Data
-
related
(7)
Quality of (Use)
Context
(9)
Other
(1)
Usability
(11)
Dependa
-
bility
(20)
Network
-
&
SI
-
related
(7)
Perfor
-
mance
(6)
Config
. &
Mgt.
(4)
Cost
(3)
©

S
-
Cube


9/18

SLOs
accessible

via

KM!!!

A. Metzger / WP
-
JRA
-
1.3


Month 12 Review Meeting, Essen, April 2009

Key Results Achieved

Knowledge Modelling


S
-
Cube Quality Reference Model (QRM)


Approach

-
Data collection
: describe most important quality models in disciplines

-
Quality attributes analysis
: identify relevant attributes

-
Consolidation:
synthesize S
-
Cube QRM from quality attributes / models


Models Analyzed

-
ISO Software Quality Model

-
UML
-
Based Quality Models

-
Statically Inferred QoS Attributes Model

-
Design by Contracts Models

-
Functional Quality in Service Composition Model

-
Service Networks and KPIs Model

-
Grid Quality Model

SOC

Softw.

Eng.

BPM

Grid

D

©

S
-
Cube


10/18

A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

©

S
-
Cube


42
/<Max>

Exploiting HCI knowledge for automatic
quality contract establishment


Problems


Need for automated service contract negotiation (time can be critical factor)


User interaction and experience impacts on negotiation


Quality modelling languages offer limited capabilities to support automated negotiation

-
Lack of formalization (hinders automation)

-
Negotiation
-
related concepts missing; e.g.,
negotiatable

vs. non
-
negotiatable

attributes

-
Missing shared terminology between provider and consumers


Solution


Quality meta
-
model (QMM) encompassing the concepts for a rich, extensible, and
semantically enriched quality definition language


Automated negotiation techniques based on QMM concepts


Considering codified UI aspects in QMM and negotiation techniques


S
-
Cube Publications

1.
Marco
Comuzzi

and Barbara Pernici. A Framework for
QoS
-
Based Web service Contracting. In ACM Transactions on
the Web, 3(3), 2009

2.
Marco
Comuzzi
,
Kyriakos

Kritikos

and
Pierluigi

Plebani
. Semantic
-
aware Service Quality Negotiation. In
ServiceWave

2008

3.
Marco
Comuzzi
,
Kyriakos

Kritikos

and
Pierluigi

Plebani
. A semantic based framework for supporting negotiation in
Service Oriented Architectures. In Proceedings IEEE CEC 2009

4.
Kyriakos

Kritikos

and
Dimitris

Plexousakis
. Mixed
-
Integer Programming for
QoS
-
Based Web Service Matchmaking. In
IEEE Transactions on Services Computing, 2(2), 2009

N

D

A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

©

S
-
Cube


43
/<Max>

Automatic quality contract establishment

Foundation: Quality Definition


S
-
Cube Quality Meta Model (Excerpt)
[1, 2, 3, CD
-
JRA
-
1.3.3]


Based on Quality Reference Model [CD
-
JRA
-
1.3.2],

built from quality attributes relevant for each of the

layers
(


䩒A
-
㈮2Ⱐ䩒A
-
㈮2Ⱐ䩒A
-
㈮2)


Augmented by negotiation
-
related
concepts


Considering UI aspects; e.g., user models
(


䩒A
-
ㄮ1)

N

D

Quality selection model:
Specifies the
importance of each quality attribute for the
requester and how the best service should be
selected.

Requestor’s

Requirements

towards QoS

A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

©

S
-
Cube


44
/<Max>

Automatic quality contract establishment

Foundation: Quality Definition


S
-
Cube Quality Meta Model (Excerpt)
[1, 2, 3, CD
-
JRA
-
1.3.3]


Negotiation
-
related concepts





N

D

Negotiable:

provider can fix the QoS value at execution time
(e.g., response time)

Non
-
negotiable:

QoS value is pre
-
determined at execution
time (e.g., reputation)

A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

©

S
-
Cube


45
/<Max>

Automatic quality contract establishment

Results: Negotiation Techniques


Framework for quality (
QoS
) contracting
[1, CD
-
JRA
-
1.3.3]


(1) Phase 1: Matchmaking

-
a) Service offer must cover the
requirements

on
non
-
negotiable

QoS

dimensions

-
b) Service offer must cover, at least partially, the
requirements

on
negotiable

QoS

dimensions expressed

-
c) Price associated to minimum quality profile


budget
B

of service requestor


(2) Phase 2: Selection

-
Providers ranked by bidding function

-
Maximize requirement coverage (maintaining price below
B
)

-
Penalize services that only partially cover the requirements (utility function)

-
Provider of service associated to lowest bid
L

is selected

-
Extra
budget
EB

=
B


L


(3) Phase 3: “Actual” Negotiation (if EB > 0)

-
Select for each negotiable
QoS

dimension a single
QoS

value

-
Assuming
EB

should be spent (user model)

-
Increase of
QoS

levels (order relation) based on priority of
QoS

attributes
(quality selection model)

N

D

A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

©

S
-
Cube


46
/<Max>

Run
-
time Quality Assurance

Techniques


Problems


Design
-
time QA is not enough due to dynamic adaptation, context changes and open
nature of SBAs


Monitoring at run
-
time only checks “arbitrary” applications in operation (no systematic
coverage)


Solution


Understanding when in the life
-
cycle run
-
time verification should be performed


Exploiting consolidated design
-
time QA techniques (here:
verification
) during run
-
time


S
-
Cube Publications

1.
Domenico Bianculli, Carlo Ghezzi and Cesare Pautasso. Embedding Continuous Lifelong Verification in Service Life
Cycles. In Proceedings PESOS @ ICSE 2009

2.
Domenico Bianculli and Carlo Ghezzi. SAVVY
-
WS at a glance: supporting verifiable dynamic service compositions.
In Proceedings ARAMIS @ ASE 2008

3.
Domenico Bianculli, Carlo Ghezzi, Paola Spoletini, Luciano Baresi and Sam Guinea. A Guided Tour through
SAVVY
-
WS: a Methodology for Specifying and Validating Web Service Compositions. In Proceedings Advances in
Software Engineering 2008

4.
Andreas Gehlert, Antonio Bucchiarone, Raman Kazhamiakin, Andreas Metzger, Marco Pistore, Klaus Pohl:
Exploiting Assumption
-
Based Verification for the Adaptation of Service
-
Based Applications. In Proceedings SOAP
@ SAC 2010

5.
Brice Morin, Olivier Barais, Grégory Nain, and Jean
-
Marc Jézéquel. Taming dynamically adaptive systems with
models and aspects. In Proceedings ICSE 2009.

A

A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

©

S
-
Cube


47
/<Max>


Life
-
cycle Model

[1, CD
-
JRA
-
1.3.4]


QA

oriented life cycle

layered on existing iterative life cycle

(



-
䩒A
-
ㄮ1)


Due to
open nature
,

SBS need to "continuously" assert

properties that have a “lifelong” validity

-
E.g., there is no guarantee that a service
implementation eventually fulfils the contract
promised (e.g., SLA)

-
E.g., during design
-
time QA, it is not
possible to model the behaviour of the
underlying distributed infrastructure (e.g.,
Internet)


Existing QA techniques applied

at each stage of the service life cycle


Combining different techniques can

improve the overall quality of QA

Run
-
time Quality Assurance Techniques

Results

A

A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

©

S
-
Cube


48
/<Max>

Run
-
time Quality Assurance Techniques

Results

A


Run
-
time Verification: Basic Approach

[2, 3, 4]


Activities during Design
-
Time:

-
Specify service composition (workflow):
W
(


䩒A
-
ㄮ1Ⱐ䩒A
-
㈮2)

-
Assume properties of the outside world / context:
A
(


乥杯瑩慴楯温

-
e.g.,
QoS

of services as stipulated in SLAs

-
Formalize requirements towards workflow:
R
(


䩒A
-
ㄮ1Ⱐ䩒A
-
㈮2)

-
Check (e.g., using model checker) that workflow meets requirements

-
W
,
A
|
--

R ?


W

R

A

W, A
|
--

R ?

X

A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

©

S
-
Cube


49
/<Max>

Run
-
time Quality Assurance Techniques

Results

A


Run
-
time Verification: Basic Approach

[2, 3, 4, CD
-
JRA
-
1.3.4]

X

M


A ?


W
,
A
’,
M

|
--

R ?


M


Activities during Run
-
time
:

-
(1) Monitor assumptions
(


䩒A
-
ㄮ1
)
:
M

-
(2) Check violation of assumptions:
M


A ?

-
If violated: (3) check if requirements are still met

-
based on past monitoring data
M

+ assumptions on
“future” invocations
A


-
W
,
A
’,
M

|
--

R


-
If requirements are not met: (4) adapt SBA (e.g.,
replace services)
(


䩒A
-
ㄮ1
)


A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

©

S
-
Cube


50
/<Max>

Quality Prediction Techniques to

Support Proactive Adaptation


Problems


Reactive adaptation has significant shortcomings


Quality prediction needed to enable pro
-
active
adaptation
(


䩒J
-
ㄮ1)


Unnecessary pro
-
active adaptations need to be avoided (as they can be costly)


Solution


Exploiting synergies between monitoring and
online testing


Determining (and fostering) confidence of failure prediction


S
-
Cube Publications

1.
Julia
Hielscher
, Raman Kazhamiakin, Andreas Metzger and Marco Pistore. A Framework for Proactive Self
-
Adaptation of Service
-
based Applications Based on Online Testing. In
ServiceWave

2008, Nr. (5377), Springer, 10
-
13 December 2008.

2.
Andreas Gehlert, Julia
Hielscher
,
Olha

Danylevych

and Dimka Karastoyanova. Online Testing, Requirements
Engineering and Service Faults as Drivers for Adapting Service Compositions. In Dimka Karastoyanova, Raman
Kazhamiakin, Andreas Metzger and Marco Pistore editors, Proceedings of the International Workshop on Service
Monitoring, Adaptation and Beyond (MONA+ 2008), December 13, 2008, Madrid, Spain, Pages 39
--
50, 2008.

3.
Andreas Gehlert, Andreas Metzger, Dimka Karastoyanova, Raman Kazhamiakin, Klaus Pohl, Frank Leymann,
Marco Pistore. Adaptation of Service
-
Based Systems based on Requirements Engineering and Online Testing.
Internet of Services Book (S.
Dustdar
, Ed.)


to be published
-

4.
Andreas Metzger, Osama Sammodi, Klaus Pohl. Towards Pro
-
Active Adaptation with Confidence


Augmenting
Monitoring with Online Testing. SEAMS Workshop @ ICSE 2010

5.
Philipp
Leitner
,
Branimir

Wetzstein
,
Florian

Rosenberg, Anton
Michlmayr
,
Schahram

Dustdar
, and Frank Leymann.
Runtime Prediction of Service Level Agreement Violations for Composite Services. In Proceedings Non
-
Functional
Properties and SLA Management @ ICSOC 2009

A

A. Metzger / WP
-
JRA
-
1.3


Global Meeting, Pisa, Mar. 2010

©

S
-
Cube


51
/<Max>

Quality Prediction Techniques to

Support Proactive Adaptation


PROSA
-
Framework

[1, 2, 3, CD
-
JRA
-
1.3.4]


Framework for exploiting online testing
for

pro
-
active adaptation

-
Integration of online testing with
monitoring

-
Integration of pro
-
active, corrective
adaptation with pro
-
active,
perfective adaptation


Failure Prediction with Confidence

[4, CD
-
JRA
-
1.3.4]


Determining whether failure prediction
is of expected confidence


Initiating online tests to collect data
points for required confidence


Invalidating data points in cases of
adaptation

Service
-
based
Application

Decide on

Adaptation

monitor

adapt

predict

Online

Testing

A





Case studies: a methodological
approach






Towards validation: Case studies


Goals:


validate research results with realistic scenarios


develop research challenges derived from case studies




A methodology for case study documentation has been
developed




Case studies


S
-
Cube proposal


An approach to describe case studies derived from NEXOF
-
RA and
enriched with other elements from the RE literature


The identification of a set of case studies which the approach is
applied on





Case studies


Directly from S
-
CUBE


Wine production (Donnafugata)


Automotive supply chain (360Fresh and IBM)


Derived from NEXOF
-
RA


E
-
health diagnostic workflow (Siemens/Thales)


Traffic management (Siemens)


E
-
government (TIS and Engineering)



For a complete description of the case studies analyzed in S
-
Cube, please refer to
the deliverable CD
-
IA
-
2.2.2 and for scenarios CD
-
IA
-
2.2.4

© S
-
Cube



The proposed case study description
approach


Business goals: express the main purposes of some system in the terms
of the business domain in which the system will live or currently lives


Domain assumptions and constraints: report properties of the domain or
restrictions on the design of the system architecture



Domain description: phenomena occurring in the world together with the
laws that regulate such a world


Abstract scenario description: a way to describe world phenomena


P. Plebani
-

IE4SOC Opening
-

Stockholm 23/11/2009

© S
-
Cube



Business goals and domain
assumptions/constraints


Business goals and domain assumption/constraints rely on
the same elements:


Description


Rationale


Involved stakeholders


Conflicts


Supporting material


Priority

P. Plebani
-

IE4SOC Opening
-

Stockholm 23/11/2009

© S
-
Cube



Domain description


Purpose


Study and describe phenomena in the world as well as shared
phenomena


Content


Glossary


Relationships among the main terms

-
Through class diagrams, semantic networks, E/R diagrams, …


Boundaries between the world and the machine

-
Context diagrams

P. Plebani
-

IE4SOC Opening
-

Stockholm 23/11/2009

© S
-
Cube



Scenarios description


Purpose: to describe possible situations and interactions
between the world and the machine


Structure of description


Involved actors


Detailed operational description


Problems and challenges


Non
-
functional requirements and constraints


Accompanying material

-
sequence and activity diagrams

-
(sub)use case diagrams


From scenarios,
abstract scenarios

are derived (template)

P. Plebani
-

IE4SOC Opening
-

Stockholm 23/11/2009

© S
-
Cube



Case study description life
-
cycle


The four elements NOT necessarily have to be defined
sequentially


Goals

assumptions

domain

scenario


They can be defined iteratively


Some rules:


All the terms used in the description have to be put in a glossary


All identified actors have to appear in the context diagram (and vice
versa)


From each scenario there exist at least one related business goal and
vice versa


Scenarios are not overlapping


Goals are not overlapping

P. Plebani
-

IE4SOC Opening
-

Stockholm 23/11/2009

© S
-
Cube



Barbara Pernici


Month 24 Review Meeting, April 2010

Coverage of life cycle

www.s
-
cube
-
network.eu

P. Plebani
-

IE4SOC Opening
-

Stockholm 23/11/2009

© S
-
Cube

Classifying & Comparing case studies



Case study meta
-
data


Used to index case studies in the repository for facilitating
search mechanisms


Meta
-
data:


Source


Real vs. Realistic


Abstract


Available solutions


Licensing




P. Plebani
-

IE4SOC Opening
-

Stockholm 23/11/2009

© S
-
Cube



Comparison dimensions


S
-
Cube


Description of business situations and presence of agile service
networks


Need for negotiating, establishing, monitoring, enforcing
QoS


Need for service consumers with various different characteristics


Need for distributed infrastructures


Need for highly distributed service compositions


Highly changing requirements at various levels (from business to
infrastructure)


Others


Security


Reliability




P. Plebani
-

IE4SOC Opening
-

Stockholm 23/11/2009

© S
-
Cube




SEEKING NEW
CASE STUDIES

(OR SCENARIOS FOR EXISTING ONES)!