Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Delivery Center Testfactory.Lenkungsausschuss Offshore Programm.
======!"§==Systems=
Enabling Offshore Test withModel-based TestingMarch 07, 2007; Test Factory, Andreas Braun
Page 1March 07, 2007; Dr. Andreas Braun
Contacts at Test Factory, T-Systems
Andreas BraunAndreas BraunAndreas BraunAndreas BraunProject Manager & Team Lead for MBT ResearchEmail: [email protected]: +49 89 1011-3461
HansHansHansHans----Peter GoorenPeter GoorenPeter GoorenPeter GoorenProject ManagerEmail: [email protected]: +49 211 6211-3370
Page 2March 07, 2007; Dr. Andreas Braun
History of Software Testing
‘‘‘‘80s/ early 80s/ early 80s/ early 80s/ early ‘‘‘‘90s: 90s: 90s: 90s: Test management and improvement of software qualityTest management and improvement of software qualityTest management and improvement of software qualityTest management and improvement of software quality
ApproachApproachApproachApproach: independent test team, test processes & methods
� more stable software systems, better software quality
� expensive, costly, no common approach
Late Late Late Late ’’’’90s: 90s: 90s: 90s: reduce cost, improve competitivenessreduce cost, improve competitivenessreduce cost, improve competitivenessreduce cost, improve competitiveness
ApproachApproachApproachApproach: IT-Offshoring
� competitive, effective, qualified workforce
� communication hurdles and breakdown, difficult knowledge transition
‘‘‘‘00: 00: 00: 00: reduce effort, systemize workreduce effort, systemize workreduce effort, systemize workreduce effort, systemize work
ApproachApproachApproachApproach: testing based on models
� Formal, complete content, improved communication
Ongoing research: Ongoing research: Ongoing research: Ongoing research: Enable offshore test, industrialize testEnable offshore test, industrialize testEnable offshore test, industrialize testEnable offshore test, industrialize test
HypothesisHypothesisHypothesisHypothesis: Model-based Testing helps to solve both issues
TTTT----SystemsSystemsSystemsSystemsTest FactoryTest FactoryTest FactoryTest Factory
Page 3March 07, 2007; Dr. Andreas Braun
Some Terms and Definitions
� Outsourcing� Near shore� Offshore� Near shoring vs. off shoring
� Captive vs. independent� Export or domestic market� Globalization
� Test, modeling, Model-based Testing
Page 4March 07, 2007; Dr. Andreas Braun
Enabling Offshore Test with Model-based TestingAgenda
Part 1: Offshoring Governance and ManagementPart 1: Offshoring Governance and ManagementPart 1: Offshoring Governance and ManagementPart 1: Offshoring Governance and Management� Introduction, Vision, Strategy� Governance and Management Approach
Part 2: ApproachPart 2: ApproachPart 2: ApproachPart 2: Approach� Masterplan, Iterative Approach� MBT Pilots, Goals and Measures
Part 3: ModelPart 3: ModelPart 3: ModelPart 3: Model----based Testing based Testing based Testing based Testing � The Model-based Testing approach revisited� Modeling-based Testing in CRM-T� An example from CRM-T
Bachelor/ MasterBachelor/ MasterBachelor/ MasterBachelor/ Master’’’’s/ Diploma Thesess/ Diploma Thesess/ Diploma Thesess/ Diploma ThesesPart 4: OutlookPart 4: OutlookPart 4: OutlookPart 4: Outlook
� Offshore Test Process� Organization and Roles� Offshore Suitability Metrics
Page 5March 07, 2007; Dr. Andreas Braun
Part 1---Offshoring Governance and ManagementAgenda
� Introduction to Offshoring and Outsourcing� Reminder: Modeling in Testing & the Offshoring Test Framework� Offshore Capability and Related OSM Level� Legacy Situation and Goals
Page 6March 07, 2007; Dr. Andreas Braun
Job Migration: Is Offshoring a Threat to the Local Job Situation?US IT Jobs 1999/2003 (BLS)
VarianceVarianceVarianceVariance20032003200320031999199919991999Role Role Role Role
2922
70
257
156
245
97
486
481
293
410
403
+8,71%2688Total
+16,67%60Hardware engineers
-8,54%281Computer systems managers ********
+59,18%98Network & data communications analysts
+19,51%205Network and systems admin
-3,96%101Database administrators ********
+13,55%428Computer systems analysts
+3,89%463Computer support
+40,19%209Software Engineering systems
+41,87%289Software engineering applications
-23,82%529Programmers *********
Page 7March 07, 2007; Dr. Andreas Braun
Job Migration: Is Prevention of Offshoring a Threat?A recent (near shoring) example from the automotive industries.
� Schefenacker produces almost 30% of rear-view mirrors in the automotive industry worldwide
� Schefenacker has no direct end-customer access
� Hence, Schefenacker is a supplier and lives on outsourcing
� Schefenacker utilized its market position to increase pressure on manufacturers
Problem SituationProblem SituationProblem SituationProblem Situation
� Production cost to high
� No window of opportunities for movement (no price hedging, resource cost increases, market pressure)
� They missed the entry point into outsourcing/ near shoring due to management mistakes and due to labor union pressure
� As a consequence and due to time pressure, mistakes in outsourcing (Slovenia) led to additional loss
� Schefenacker is now in for insolvency, covenant breach not reached
TodayTodayTodayToday
� As a consequence to the power struggle, Schefenacker later lost such customers
� Consequence: to be taken over by financial investors and hedge fonds (Permira)
� Loss of production at DaimlerChrysler
Lessons learntLessons learntLessons learntLessons learnt
� Evading offshoring not promising
� Market pressure will increasingly reduce window of opportunities
� Cost
� Time
Consequences:Consequences:Consequences:Consequences:
� Offshoring inevitable
� Offshoring must be productive and chargeable
� Early definition needed, experience needed
Our Vision and Strategy:Our Vision and Strategy:Our Vision and Strategy:Our Vision and Strategy:
� No staff replacement, but:
� free local resources for new projects
� Reduce cost of work/ cost of production
� Retain control
� No pure offshoring (“Fire & Forget”)
� Leverage mix of skills & cost
� Automation first!
Source: FAZ, October 31st, 2006Source: FAZ, October 31st, 2006Source: FAZ, October 31st, 2006Source: FAZ, October 31st, 2006
News from Friday, 16.02.2007 Schefenacker: Financial restructuring plan drawn up Senior lenders and the main shareholder of automotive supplier Schefenacker (Schwaikheim / Germany; www.schefenacker.com), Alfred Schefenacker, have agreed on a financial restructuring plan that foresees nearly 75% of the mirror specialist passing into the hands of British hedge funds. The company’s current debt of EUR 450m will be replaced by a new financial package that includes credit facilities worth EUR 305m, along with fresh capital of EUR 55m. The lighting business, accounting for 22% of total sales, is to be divested. �
Page 8March 07, 2007; Dr. Andreas Braun
Reminder: Modeling in Testing & the Offshoring Framework The Test Factory Offshoring Framework consists of five layers defining major activities of Offshore Management and Operations.
OffshoreGovernance
Offshore Management
Offshore Test Process
Tools & Methods
Operative ManagementProgram Management
Strategic Management
Project Management
Project Execution
VisionStrategy
Page 9March 07, 2007; Dr. Andreas Braun
Offshore Capability and Related OSM Level.To improve the offshore capability, the SIT aims at moving towards a blended workforce model (“Cross-border Collaborative Test”) with shared responsibilities.
Offs
hore Cap
ability/
Offs
hore Cap
ability/
Offs
hore Cap
ability/
Offs
hore Cap
ability/
Com
plex
ity of S
ervice
Delivery
Com
plex
ity of S
ervice
Delivery
Com
plex
ity of S
ervice
Delivery
Com
plex
ity of S
ervice
Delivery
ResponsibilityResponsibilityResponsibilityResponsibility RemoteLocal
Low
High
Onsite Onsite Onsite Onsite StaffStaffStaffStaff
Onsite Onsite Onsite Onsite ProjectProjectProjectProject
Onsite Onsite Onsite Onsite ––––Offshore Offshore Offshore Offshore projectprojectprojectproject
Pure Pure Pure Pure OffshoreOffshoreOffshoreOffshore
OffshoreOffshoreOffshoreOffshoreTest FactoryTest FactoryTest FactoryTest Factory
CrossCrossCrossCross----border border border border Collaborative Collaborative Collaborative Collaborative
TestTestTestTest
•Traditional approachTraditional approachTraditional approachTraditional approach•Staff augmentationStaff augmentationStaff augmentationStaff augmentation•Performance Performance Performance Performance
management: management: management: management: individual feedbackindividual feedbackindividual feedbackindividual feedback
•Offshore team works Offshore team works Offshore team works Offshore team works onononon----sitesitesitesite
•WellWellWellWell----defined scopedefined scopedefined scopedefined scope
•Team takes Team takes Team takes Team takes ownershipownershipownershipownership
•Client servicing, Client servicing, Client servicing, Client servicing, project management project management project management project management onshoreonshoreonshoreonshore
•ApplicationApplicationApplicationApplicationtesting offshoretesting offshoretesting offshoretesting offshore
•Better flexibility, Better flexibility, Better flexibility, Better flexibility, scalability, stabilityscalability, stabilityscalability, stabilityscalability, stability
�Rarely adoptedRarely adoptedRarely adoptedRarely adopted
�No onNo onNo onNo on----site presencesite presencesite presencesite presence
�High riskHigh riskHigh riskHigh risk
•Vendor setsVendor setsVendor setsVendor sets----up center up center up center up center offshore dedicated to offshore dedicated to offshore dedicated to offshore dedicated to client needsclient needsclient needsclient needs
• In return, client In return, client In return, client In return, client commits to buy a commits to buy a commits to buy a commits to buy a minimum amount of minimum amount of minimum amount of minimum amount of services over an services over an services over an services over an extended timeextended timeextended timeextended time
•Support complements Support complements Support complements Support complements clientclientclientclient‘‘‘‘s internal s internal s internal s internal standards and standards and standards and standards and proceduresproceduresproceduresprocedures
•Complementary Complementary Complementary Complementary resources from resources from resources from resources from multiples countries multiples countries multiples countries multiples countries pooled on a common pooled on a common pooled on a common pooled on a common delivery platformdelivery platformdelivery platformdelivery platform
•HighlyHighlyHighlyHighly----structured structured structured structured partnerships in the partnerships in the partnerships in the partnerships in the different countriesdifferent countriesdifferent countriesdifferent countries
• Leverage capabilities Leverage capabilities Leverage capabilities Leverage capabilities of each center to of each center to of each center to of each center to create a portfolio of create a portfolio of create a portfolio of create a portfolio of highhighhighhigh----quality, costquality, costquality, costquality, cost----effective solutionseffective solutionseffective solutionseffective solutions
•Performance Performance Performance Performance management: management: management: management: SLAsSLAsSLAsSLAs
Level 0 Level 1 Level 2 Level 3 Level 4 Level 5
Legacy Situation and GoalsThe Legacy Offshoring-Model is a mixture between Provision of Services and Assignment of Work Packages (Craft).
� Integration of offshore partners hampered
� “Local heroes”
Impact on Current OffshoringImpact on Current OffshoringImpact on Current OffshoringImpact on Current Offshoring
Standardized Notation and Tools, Standardized Notation and Tools, Standardized Notation and Tools, Standardized Notation and Tools, Formal Methods Formal Methods Formal Methods Formal Methods ����
� Industrial processes �UML, U2TP, TTCN-3�Model-based Testing�Artifact-based Onshore-Offshore Communication
�
Difficult transition over organizational boundaries and project contexts
�
Tools, Methods, and ProcessTools, Methods, and ProcessTools, Methods, and ProcessTools, Methods, and Process� Lack of formal methods and
notation� No standardized methods� Craftsmanship
GoalGoalGoalGoalResulting ProblemsResulting ProblemsResulting ProblemsResulting ProblemsAsAsAsAs----is Situationis Situationis Situationis Situation
Knowledge Sharing and Reuse Knowledge Sharing and Reuse Knowledge Sharing and Reuse Knowledge Sharing and Reuse �����Models encapsulate domain knowledge
�Model repositories enable reuse
�Costly training upfront�Many inquiries�Much overhead due to coaching�Low productivity
�
Person-dependent expertiseDifficult knowledge sharing and transfer
�
Tacit KnowledgeTacit KnowledgeTacit KnowledgeTacit Knowledge� Architecture� Organization� Business processes� Sources for information,
stakeholders� No up to date documentation
Defined communication paths Defined communication paths Defined communication paths Defined communication paths �����Onsite and Offshore-coordinators
� Local stakeholders and experts on both sides
� Language encapsulation
�Overhead of cost and time�Burden on local experts
�Translations are error proneMisunderstandings
�
Language BarriersLanguage BarriersLanguage BarriersLanguage Barriers� German documentation/ GUI� Jargon, Buzzwords� Translation to English� No artifact sharing
Leverage Balanced Teams Leverage Balanced Teams Leverage Balanced Teams Leverage Balanced Teams �����Blended workforce model� Increased responsibility�Critical mass
�Charging by the hour�No responsibility transfer to offshore partners
�No workload balancing, no resource sharing
�
No work packaging (entities)Difficult responsibility transfer and integration of offshore ressourcesReactive workstyle
�
Monolithic Project TeamMonolithic Project TeamMonolithic Project TeamMonolithic Project Team� Tight coherence and coupling of
team members� Strong intra-team bindings� Weak extra-team bindings� Tight informal communication� No artifact sharing
Page 11March 07, 2007; Dr. Andreas Braun
Any Questions?
Page 12March 07, 2007; Dr. Andreas Braun
Part 3---Update on Model-based Testing Agenda
� Model-Based Testing (MBT)� Definition� Activities
� MBT in CRM-T� Activities and Artifacts� Process and Iterations� Example out of CRM-T� Terminology� Results� Evaluation
� Summary & Conclusion
Page 13March 07, 2007; Dr. Andreas Braun
The Model-Based Testing Theme Song © 1999, Harry Robinson
(to the tune of “I'd Like to Teach the World to Sing”)
Pete Seeger once said that any good movement needs a theme song, so (with apologies to all) here is my submission for the model-based testing theme song, complete with harmony section:
I’d like to teach the world to model Increase code quality I’d like to find bugs in the spec And keep my weekends free (“… see my family…” )
I’d like to teach the world to model And teach monkeys to test So they can find the bugs at night While I’m home getting rest ( “…that sounds good to me…” )
I’d like to teach the world to model And teach monkeys to think So they could find bugs early on And not just on the brink (“ … of delivery… “)
I’d like to teach the world to model For perfect quality I’d like to automate their tests And ship stuff defect-free ( “ … that’s how it should be …” )
Page 14March 07, 2007; Dr. Andreas Braun
Update on Model-based Testing in CRM-T Definition & Benefits
� MBT is software testing in which test cases are derived in whole or in part from a model that describes some (usually functional) aspects of the system under test (SUT). [Wikipedia]
� MBT main activities:� Build Model
− Different Modeling Methods (e.g. Finite state machines, UML Diagrams)
� Test Generation� Test Execution
� Adopt MBT in CRM-T by:� Tailoring MBT activities according to our needs� Keeping it compatible with traditional test systems
Page 15March 07, 2007; Dr. Andreas Braun
Model-based ApproachProject phases and corresponding toolset.
Model-based Approach
Mapping the test model to TTCNTTCNTTCNTTCN----3333
Test Process
UMLUMLUMLUML Models used in:Requirements, Analysis; Modelling of domain knowledge (func., techn.); Test case generation
U2TPU2TPU2TPU2TP for test case specification(structure and behavior)
UMLUMLUMLUML for depicting the process flow
Analysis
Manual Execution
Automation
Evaluation
Design
Execution
Page 16March 07, 2007; Dr. Andreas Braun
Test ModelTest ModelTest ModelTest Model
Test StructureTest StructureTest StructureTest Structure
Test ControlTest ControlTest ControlTest Control
DeploymentDeploymentDeploymentDeploymentDiagramDiagramDiagramDiagram Test Case BehaviorTest Case BehaviorTest Case BehaviorTest Case Behavior
Test ConfigurationTest ConfigurationTest ConfigurationTest Configuration
ClassClassClassClassDiagramDiagramDiagramDiagram
Test PackageTest PackageTest PackageTest Package
InteractionInteractionInteractionInteractionDiagramDiagramDiagramDiagram
Modeling Categories: Specification vs. Test Model.Functional requirements are depicted in the UML. They build up the basis for test modeling with MBT.
System ModelSystem ModelSystem ModelSystem Model
Use CaseUse CaseUse CaseUse CaseDiagramDiagramDiagramDiagram
ActivityActivityActivityActivityDiagramDiagramDiagramDiagram
PlainPlainPlainPlainTextTextTextText
DescriptionDescriptionDescriptionDescription
Component Component Component Component /Deployment/Deployment/Deployment/Deployment
DiagramDiagramDiagramDiagram
InteractionInteractionInteractionInteractionDiagramDiagramDiagramDiagram
Page 17March 07, 2007; Dr. Andreas Braun
Advantages of the Methodology and Techniques selected.
� Models � Reduce tacit knowledge � knowledge is embedded in artifacts� Abstraction copes with complexity � improves clarity and comprehensibility� Improve communication� Easy to modify and maintain updates
� Model-based Testing� Combines testing with (semi-) formal methods� Allows usage early in the software life cycle � Higher test coverage, better results � the formal is approach less error-
prone
� Toolset� Same syntax for UML and U2TP
− Consistent specification− One single modeling tool− Reduced training and coaching effort
� Systematic mapping of U2TP to TTCN-3
Page 18March 07, 2007; Dr. Andreas Braun
Test Assignment
Deliver Test Basis
Estimation, Planning
Analyze Test Basis
System Modeling****
Test Case Generation
Test ModelingAllocate Test Data
Automated Test Execution
Log Results
Test Report
Quality AssuranceAuditCustomer Report
Manual Test Execution
Assignm
ent
Assignm
ent
Assignm
ent
Assignm
ent
Ana
lysis
Ana
lysis
Ana
lysis
Ana
lysis
Des
ign
Des
ign
Des
ign
Des
ign
Exec
ution
Exec
ution
Exec
ution
Exec
ution
Test Closu
reTe
st Closu
reTe
st Closu
reTe
st Closu
re
Defect?Defect?Defect?Defect?NoNoNoNo
YesYesYesYes
Model-based Testing Process: Overview.
Defect Management Proc.
Page 19March 07, 2007; Dr. Andreas Braun
Assignm
ent
Assignm
ent
Assignm
ent
Assignm
ent
Ana
lysis
Ana
lysis
Ana
lysis
Ana
lysis
**** If no models are delivered in the test basis only.If no models are delivered in the test basis only.If no models are delivered in the test basis only.If no models are delivered in the test basis only.
Test Assignment
Deliver Test Basis
Estimation, Planning
Analyze Test Basis
System Modeling****
Model-based Testing Process: System Modeling.
CreateUse Case Diagrams
CreateDeployment Diagrams
CreateActivity Diagrams
System ModelingSystem ModelingSystem ModelingSystem Modeling
Page 20March 07, 2007; Dr. Andreas Braun
Test Case Generation
Assignm
ent
Assignm
ent
Assignm
ent
Assignm
ent
Ana
lysis
Ana
lysis
Ana
lysis
Ana
lysis
Test Assignment
Deliver Test Basis
Estimation, Planning
Analyze Test Basis
System Modeling**** IdentifyValid Paths
Verify Path Coverage and Completeness
GenerateTest Cases
Combine with additional paths/
merge paths
Test Case GenerationTest Case GenerationTest Case GenerationTest Case Generation
Complete?Complete?Complete?Complete?
NoNoNoNoYesYesYesYes
Model-based Testing Process: Test Case Generation.
Identify TestObjectives
Page 21March 07, 2007; Dr. Andreas Braun
Assignm
ent
Assignm
ent
Assignm
ent
Assignm
ent
Ana
lysis
Ana
lysis
Ana
lysis
Ana
lysis
Des
ign
Des
ign
Des
ign
Des
ign
CreateTest Suites
ModelTest Configuration
ModelTest Architecture
ModelTest Cases
Test ModelingTest ModelingTest ModelingTest Modeling
Model
Test Control
Test Assignment
Deliver Test Basis
Estimation, Planning
Analyze Test Basis
System Modeling****
Model-based Testing Process: Test Modeling.
Test Case Generation
Test ModelingAllocate Test Data
Page 22March 07, 2007; Dr. Andreas Braun
ManualTest Execution
Assignm
ent
Assignm
ent
Assignm
ent
Assignm
ent
Ana
lysis
Ana
lysis
Ana
lysis
Ana
lysis
Des
ign
Des
ign
Des
ign
Des
ign
Exec
ution
Exec
ution
Exec
ution
Exec
ution
ModelExecution Flow of Events
PrepareTest Data
Create Test SuiteDocument
ManuelleManuelleManuelleManuelle TestdurchfTestdurchfTestdurchfTestdurchfüüüührunghrunghrunghrung
ExecuteTest Suite
Test Assignment
Deliver Test Basis
Estimation, Planning
Analyze Test Basis
System Modeling****
Test Case Generation
Test ModelingAllocate Test Data
Model-based Testing Process: Manual Test Execution.
Page 23March 07, 2007; Dr. Andreas Braun
Testfallgenerierung
Test ModelingAllocateTest Data
Automated Test Execution
Assignm
ent
Assignm
ent
Assignm
ent
Assignm
ent
Ana
lysis
Ana
lysis
Ana
lysis
Ana
lysis
Des
ign
Des
ign
Des
ign
Des
ign
Exec
ution
Exec
ution
Exec
ution
Exec
ution
TTCN-3 Mapping
Create Templates for Test Data
Install/ AdaptSUT
Automated Test ExecutionAutomated Test ExecutionAutomated Test ExecutionAutomated Test Execution
ExecuteTest Suite
Test Assignment
Deliver Test Basis
Estimation, Planning
Analyze Test Basis
System Modeling****
Model-based Testing Process: Automated Test Execution.
Page 24March 07, 2007; Dr. Andreas Braun
Test Assignment
Deliver Test Basis
Estimation, Planning
Analyze Test Basis
System Modeling****
Test Case Generation
Test ModelingAllocate Test Data
Automated Test Execution
Log Results
Test Report
Quality AssuranceAuditCustomer Report
Manual Test Execution
Assignm
ent
Assignm
ent
Assignm
ent
Assignm
ent
Ana
lysis
Ana
lysis
Ana
lysis
Ana
lysis
Des
ign
Des
ign
Des
ign
Des
ign
Exec
ution
Exec
ution
Exec
ution
Exec
ution
Test Closu
reTe
st Closu
reTe
st Closu
reTe
st Closu
re
Defect?Defect?Defect?Defect?NoNoNoNo
YesYesYesYes
Model-based Testing Process: Overview.
Defect Management Proc.
Page 25March 07, 2007; Dr. Andreas Braun
Abstract
Test Suite
Update on Model-based Testing in CRM-TTerminology
� Test Suite: A high level concept grouping together a collection of test cases related to what they are intended to test.
� Scope varies from organization to organization.� Abstract and Executable (Testfallkette)
� Abstract:� Abstract view on prerequisites (communication, time and data)
� Executable (“Testfallkette”): A counterpart to the abstract test suite� Ready to be executed.� Concrete prerequisite states or steps and system configuration.
AbstractAbstractAbstractAbstract
Test SuiteTest SuiteTest SuiteTest Suite
ExecutableExecutableExecutableExecutable
T ran
sla te Te
s ts
T ran
sla te Te
s ts
T ran
sla te Te
s ts
T ran
sla te Te
s ts
Models
Test Cases
INPUT EXPECTED
OUTPUT
ExecutableTest Suite
(“Testfallkette”)
Page 26March 07, 2007; Dr. Andreas Braun
Any Questions?
Page 27March 07, 2007; Dr. Andreas Braun
Bachelor/ Master’s/ Diploma Theses@ T-Systems Test Factory
1. Design and Implementation of an Import/Export Filter for U2TP1. Design and Implementation of an Import/Export Filter for U2TP1. Design and Implementation of an Import/Export Filter for U2TP1. Design and Implementation of an Import/Export Filter for U2TP Test Suites in a Test Suites in a Test Suites in a Test Suites in a test management system.test management system.test management system.test management system.� Learning of the test process and quality features (ISO-Standards),
U2TP and UML 2.0� Build a Meta-model for the UML Testing Profile for SparxX Enterprise
Architect (EA).� Definition of a Taxonomy for Test Specification.� Mapping of SparxX EA XMI outputs in a testing data-model (Test Work
Bench, TWB).The goal is to design and implement an executable Prototype. Keywords/know-how: UML, U2TP, UML Profile, XMI, Java APIs, SQL DB.
Page 28March 07, 2007; Dr. Andreas Braun
Bachelor/ Master’s/ Diploma Theses@ T-Systems Test Factory
2. Definition of 2. Definition of 2. Definition of 2. Definition of modelingmodelingmodelingmodeling guidelines and processes for Modelguidelines and processes for Modelguidelines and processes for Modelguidelines and processes for Model----based Testing based Testing based Testing based Testing (MBT).(MBT).(MBT).(MBT).
� Learning of the test processes and the quality features (ISO Standards), UML and MBT.
� Analyse the existing MBT processes of Testfactory T-Systems� Learning of cross project modeling examples� Derivation of Modeling guidelines� Identification of description possibilities for modeling guidelines
Goal is the definition of a framework for Model-based Testing.
Page 29March 07, 2007; Dr. Andreas Braun
Bachelor/ Master’s/ Diploma Theses@ T-Systems Test Factory
3. Evaluation of 3. Evaluation of 3. Evaluation of 3. Evaluation of ModelingModelingModelingModeling and Specification methods for different Test and Specification methods for different Test and Specification methods for different Test and Specification methods for different Test characteristics and requirements.characteristics and requirements.characteristics and requirements.characteristics and requirements.
� Learning of the test processes and the quality features (ISO Standards)� Analysis of specifications of different test types and quality features.� Implementation of case studies and evaluation of concrete examples.� Evaluation of the possibility of automated test generation.� Scientific assessment of results.� Formulation of recommendations and suggestions.
Goal is the development of an overview of modeling guidelines for different test types and requirements.
Page 30March 07, 2007; Dr. Andreas Braun
Bachelor/ Master’s/ Diploma Theses@ T-Systems Test Factory
4. From Use Cases to Activity Diagrams:4. From Use Cases to Activity Diagrams:4. From Use Cases to Activity Diagrams:4. From Use Cases to Activity Diagrams:� Convert textual-use case specifications into activity diagrams� Creation of templates for textual-use cases� Annotating use cases with stereotypes in such a way that the test
cases are clearly delimited and then having the tool automatically parse those annotated documents to convert them into XML metadata interchange (XMI) files for later input into Sysiphus[1].
Goal is the development of a plugin in Sysiphus which allows test designers to semiautomatically convert their existing textual-use case specifications into activity diagrams.[1] Sysiphus: http://sysiphus.in.tum.de
Page 31March 07, 2007; Dr. Andreas Braun
Part 4---Outlook Agenda
� Phases in the Offshore Project� Organizational Model
� Organigram� Roles & Expectations in the Project Domain� Roles & Expectations in the Offshore Domain
Page 32March 07, 2007; Dr. Andreas Braun
Phases in the Offshore ProjectThe Offshoring (Sub-) Project is organized accordingly to the Rational Unified Process (RUP)
ElaborationTest architecture� Spec. cycles� Int./ ext. reviews� Updates (CRs)� Final spec (TWB)
TransitionTest Closure
� Auditing� Test reporting� Test closure� Process
improvement
Baseline Test End Criteria
InceptionPlanning & Setup� Statement of work� Analysis� Estimation� Review� Kick-off
ConstructionTest control� Test execution� Steering� Defect management � Reporting
Draft !
Testanalyse
und -design
Testrealisierung
und -durchführung
Testauswertung
und -bericht
Testabschluss
Testplanung
und -steuerung
Testaktivitäten
Organizational ModelThe Offshoring and Project Domain are orthogonal and fulfill specific, defined assignments.
Draft !
Team Lead
Team Member
SPOCOnsite
Coordinator
Functional Lead
Assignment Owner
Status to: Functional Control by: Discipl. Reporting to:
Onsite: Offshore: (Dotted lines denote lose coupling)
Offshore D
omain
Project D
omain
Offshore ProgramManager
Rules:� Onshore-Offshore
communication is strictly artifact-based.
� Fixed resource allocation
Project Manager
Project Assignment:� Fulfill client expectations� Face to the client� Achieve project goals� Meet budget, deadlines� Delivery quality work� Define offshore work
packages/ assignments� Meet project’s offshore
capability
Offshore Assignment:� Meet offshoring strategy� Optimize chargeability� Improve method� Deliver quality work� Meet deadlines
Offshore Manager
OffshoreLiaison Manager
Program
D
omain
Scope, Q
ualityT
ime, B
udget
Program Manager
AD
SF
TF
Page 34March 07, 2007; Dr. Andreas Braun
CostCostCostCost�Non- profitable�Low profit/ break even�Clear profit margin
�Profit Margin (Lucrative)
�Core (Onsite)Core (Onsite)Core (Onsite)Core (Onsite)�Non-core (Offshore)
�Mission criticalMission criticalMission criticalMission critical�Non-critical
�High, i.e. dedicated individualsHigh, i.e. dedicated individualsHigh, i.e. dedicated individualsHigh, i.e. dedicated individuals�Medium, i.e. intra-company�Low, i.e. external
Value Value Value Value (starts with worst)
-200 to +80
�Confidentiality
StrategicStrategicStrategicStrategic(VETO)(VETO)(VETO)(VETO) �Criticality
�Core business
ScoringScoringScoringScoringCriteriaCriteriaCriteriaCriteriaCategoryCategoryCategoryCategory
Criteria Catalogue: Strategic Dimension Draft !
�Bad�Normal�Good
�Software quality/ defect analysis
�Floating�Normal�Stable
�Scope stability
�Pure Application�E2E
�Business process/ application test
�Bad/ instable�Limited availability/ stability�Full performing/ stable
�Infrastructure performance and stability
�Not available�Procurement�Available
�Availability of technical resourcesTechnology/Technology/Technology/Technology/AdministrationAdministrationAdministrationAdministration
-130 to +130
�Short, <3 months�Mid, 3 to 6 months �Long, > 6 months
�Duration
AssignmentAssignmentAssignmentAssignment
�Adhoc�Planned
�Forerun
�Undefined�Vague�Defined
�Scope Definition
�Small, 1,2 FTEs�Medium, 3,4 FTEs�Large, 5+ FTEs
�Size
�Process section (partial)�Full process (end to end)
�(Test) Process Coverage
�Native (German)�English
�Deliverable Language
�Native (German)�English
�High�Medium�Low
Value Value Value Value (starts with worst)
�Complexity
Work PackageWork PackageWork PackageWork Package
�Spec. & Doc Language
ScoringScoringScoringScoringCriteriaCriteriaCriteriaCriteriaCategory Category Category Category
Criteria Catalogue: Operations Dimension Draft !
Page 36March 07, 2007; Dr. Andreas Braun
-100 to +100
�Offshoring transparent�Offshoring basically accepted�Offshoring supported
�Expectations(Type/ Form/ Quality of Deliverables)
StakeholdersStakeholdersStakeholdersStakeholders�High interaction�Medium interaction�Low interaction
�Communication Interfaces
�Not available�Partly available�Available
�Availability of skilled resources
ResourcesResourcesResourcesResources
�Long, expensive�Brief, cheap
�Training upfront
�Tacit�Well defined�Not needed
�Tacit�Well defined�Formal Specifications
�Tacit�Well defined�Formal Specifications
Value Value Value Value (starts with worst)
�Contextual/ Cultural
�Application
�Business Process
CriteriaCriteriaCriteriaCriteria
KnowledgeKnowledgeKnowledgeKnowledge
ScoringScoringScoringScoringCategory Category Category Category
Criteria Catalogue: People Dimension Draft !
Page 37March 07, 2007; Dr. Andreas Braun
Offshore Assessment (Calculation Sheet) SIT Summary
Offshore Assessment
Assignment name: Testdata (reguläre Anforderungen)
EKI Archivierung, EMK, KAN
CRM_Sat ohne Archivierung
DMS INDIV KVS OM: neue Produkte ohne Business-Kritikalität
OM neu (Prozessänderungen)
TIMB
By: allDate: 23.11.2006 23.11.2006
Review by:
Strategic Dimension: 5 -10 0 -95 0 5 5 5 -95 5Operations Dimension: 30 -10 0 -5 -15 -15 0 25 -5 -35
People Dimension: -5 -20 -30 -30 -20 -15 -5 0 -20 -20
Total Score: 30 -40 -30 -130 -35 -25 0 30 -120 -50
Page 38March 07, 2007; Dr. Andreas Braun
Any Questions?
Page 39March 07, 2007; Dr. Andreas Braun
Samples
MBTMBTMBTMBT
ExampleExampleExampleExample
Page 40March 07, 2007; Dr. Andreas Braun
MBT Example CRM-T:Offshore Specification Process Flow
Use Case DiagramActivity DiagramUse Case Description
Testbase
AnalysisAbstract Test Suite
Initial Models
Final Models
Design
Review
Page 41March 07, 2007; Dr. Andreas Braun
MBT Example CRM-T: Test Basis 1/4Use-Case Diagram: 26U01 Agent in CTI und CRM-T anmelden
Page 42March 07, 2007; Dr. Andreas Braun
MBT Example CRM-T: Test Basis 2/4Use-Case Description: 26U01 Agent in CTI und CRMT anmelden
Fehlermeldung derWinset-Applikation
Anmeldevorgang imStartup-Tool
Siebelstart mit CTI ?
nein
ja
Anmeldung an ACD
ACD Anmeldungerfolgreich? nein
ja
Winset-Applikationstarten
"Login" Siebel ServerBusiness Service starten
Ist ACD Telefonein Teleset ?
ja
nein
Fehlermeldung derWinset-Applikation
Anmeldevorgang imStartup-Tool
Siebelstart mit CTI ?
nein
ja
Anmeldung an ACD
ACD Anmeldungerfolgreich? nein
ja
Winset-Applikationstarten
"Login" Siebel ServerBusiness Service starten
Ist ACD Telefonein Teleset ?
ja
nein
MBT Example CRM-T: Test Basis 3/4Activity Diagram: Use-Case 26U01 Agent in CTIund CRMT anmelden
Siebel wird mit CTIgestartet
Siebel wird ohne CTIgestartet
Siebelstart mit CTI ?
nein ja
Rückmeldung anStartup-Tool
Fehlermeldung
Zuweisung inCRM-C
erfolgreich?nein
"Login" Siebel ServerBusiness Service
Teleset / Konfigurations-einträge löschen
Ist Anwenderein CTI Agent?
nein
jaFehlermeldung
Ist ACD Telefonein Teleset ?
nein
ja
ggf. Teleset/Konfigurationseinträge anlegenbzw. aktualisieren
ja
Siebel wird mit CTIgestartet
Rückmeldung anStartup-Tool
Rückmeldung anStartup-Tool
Agent über CTI anACD anmelden
Page 44March 07, 2007; Dr. Andreas Braun
MBT Example CRM-T: Test Basis 4/4Use-Case: 26U01 Agent in CTI und CRMT anmeldenStandard and Alternative Workflows
Work Package 26 - Requirement Traceability MatrixSr No. Use Case
NumberUse case name Test Case
NumberTest Case Name Description Priority Flow of Events Validation
1 26U01 Agent Login in CTI and CRMT TC001 26B01_CRMT_26U01_LoginWithTeleset Start Siebel with CTI, ACD Telephone is Teleset.If user is CTI agent update the configuration entries. And since ACD telephone is teleset Login to ACD via CTI.
1 Precondition: 1) The telephone type is teleset2) User is a CTI Agent
Steps:1) Login to the startup tool2) Start Siebel with CTI3) If applicable, create or update teleset/configuration entries
Status = Not Available
Agent logs into ACD via CTI
Siebel Started with CTI
2 26U01 Agent Login in CTI and CRMT TC002 26B01_CRMT_26U01_LoginWithWinset Start Siebel with CTI, ACD Telephone is not teleset so starts Winset application and assign configuration entries.If allocation is succesful and ACD is not teleset then gives feedback to Start Up tool as Siebel started with CTI.
1 Precondition: 1) The telephone type is winset2) User is a CTI Agent
Steps:1) Login to the startup tool2) Start Siebel with CTI3) Login to ACD4) Start Winset5) If applicable, create or update teleset/configuration entries
Status = Not Available
Siebel Started with CTI
3 26U01 Agent Login in CTI and CRMT TC003 26B01_CRMT_26U01_LoginWithoutCTI Login to Siebel without CTI 3 Precondition: 1) The telephone type is teleset2) User is a CTI Agent
Steps:1) Login to the startup tool2) Start Siebel without CTI3)If necessary delete teleset/ configuration entries4)Feedback to startup tool as Siebel started without CTI
Status = Not Available
MBT Example CRM-T: Test Case GenerationAbstract Test Suites
s d B C 1 _ C R M -T_ L o g i n _ W i th W i n s e t
« T e st C o m p o n e n t »
U S E R
« T e st C o m p o n e n t »
S ta r tu p To o l
« S U T »
C R M -T
« T e st C o m p o n e n t »
C R M -C
« T e st C o m p o n e n t »
C TI m i d d l e w a r e
« V a l i d a t i o n A c t i o n »
P a s s( : : )
se l e c t (si e b e l st a rt w i t h C T I )
N o : = i s A C D t e l e p h o n e t e l e se t
A C D L o g i n
Y e s: = A C d L o g i n S u c c e ssfu l l
S ta rt a p p l i c a t i o n (W i n se t )
S i e b e l L o g i n
S t a rt w i t h C T I
S t a rt
C T I st a rt e d
l o a d c o m p o n e n t s
y e s:= C T I a g e n t
C re a t e C o n f i g u ra t i o n e n t ry
A ssi g n C o n f i g u ra t i o n
A ssi g n m e n t su c c e ssf u l
N o : = i s A C D T e l e se t
R e g i st e r A g e n t to A C D
R e g i st ra t i o n su c c e ssf u l
C T I st a rt e d
C R M -T w i t h C T I st a rt e d(S t a t e = n o t re a d y )
se tS t a t e (S t a t e = R e a d y )
MBT Example CRM-T: Test ModelingTest Case 001: 26B01_CRMT_26U01_LoginWithWinset
Page 47March 07, 2007; Dr. Andreas Braun
sd BC1_CRM-T_ Login_WithoutCTI
«TestComponent»
User
«SUT»
CRMT
«TestComponent»
Startup Tool
«ValidationAction»
Pass(::)
Login
Yes:= Login Successful
Start without CTI
Login Without CTI
Delete Configuration entries
Started Without CTI
Siebel is started withoutCTI(State=not available)
setState(State=Ready)
MBT Example CRM-T: Test ModelingTest Case 002: 26B01_CRMT_26U01_LoginWithoutCTI
s d BC1 _ CRM -T_ Log in_ WithTe le s e t
« T e stCo m p o n e n t»
US E R
« T e stCo m p o n e n t»
S ta rtup Too l
« S UT »
S ie be l CRM -T
« T e stCo m p o n e n t»
CRM -C
« T e stCo m p o n e n t»
CTI M iddle w a re
« V a l i d a ti o n A cti o n »
P a s s(: :)
se l e c t(si e b e l sta rt w i th CT I)
ye s:= i s A CD te l e p h o n e te l e se t
S ie b e l L o g i n
S ta rt wi thCT I
S ta rt
CT I sta rte d
L o a d Co m p o n e n ts
Y e s:= CT I a g e n t
Cre a te Co n fi g u ra ti o n e n try
A ssi g n Co n fi g u ra ti o n
A ssig n m e n t S u cce ssfu l l
Y e s:= i sA CDT e l e se t
Re g i ste r a g e n t to A CD
Re g istra t i o n S u cce ssfu l l
CT I S ta rte d
CRM -T w i th CT Ista rte d (S ta te = n o ta va i la b le )
se tS ta te (sta te =re a d y)
MBT Example CRM-T: Test ModelingTest Case 003: 26B01_CRMT_26U01_LoginWithTeleset
Page 49March 07, 2007; Dr. Andreas Braun
Backup
(Backup)(Backup)(Backup)(Backup)
Part 2Part 2Part 2Part 2
Page 50March 07, 2007; Dr. Andreas Braun
Part 2---Case Studies, Pilots, and CRM-TAgenda
� Goals and Measures � Masterplan� Pilot’s Approach� Pilot Results
� ETC� üVBT� TS India
� Summary & Lessons Learnt� Open Issues� Next Steps
Page 51March 07, 2007; Dr. Andreas Braun
Model-Based Testing Pilots: Goals and Measures The MBT Pilots evaluate the measures suggested in the Testfactory Offshoring Framework in real-world projects.
� Prepare a decision memo (Go, No Go, Go But) by end of October
�Gain insight:Gain insight:Gain insight:Gain insight:� Feasibility of MBT� Increase chargeability� Share knowledge
� Enable knowledge transfer� Transfer knowledge from heads/ experts into models� Training and coaching� Involve Indian teams (communication, newsletter, …)
�Prevent tacit knowledge and dismantle information barriers:Prevent tacit knowledge and dismantle information barriers:Prevent tacit knowledge and dismantle information barriers:Prevent tacit knowledge and dismantle information barriers:� Knowledge transfer� Artifact-based communication� Creation of knowledge repositories
� Open and standardized Methods� Modeling tools…� UML, U2TP, TTCN-3
�Common methods and tools:Common methods and tools:Common methods and tools:Common methods and tools:� Best-practice methodology� Standardized and industrial tools & methods
� Evaluate and Improve our current Offshore Test process� Encapsulate Test Assignments with necessary
knowledge for execution� Increase responsibility in India
�Transition to industrialized processes:Transition to industrialized processes:Transition to industrialized processes:Transition to industrialized processes:� Prepare for service level agreements (SLAs)� Allow for workload balancing� Re-use
Measures in the MBT Pilots:Measures in the MBT Pilots:Measures in the MBT Pilots:Measures in the MBT Pilots:Offshore Governance Goals:Offshore Governance Goals:Offshore Governance Goals:Offshore Governance Goals:
Page 52March 07, 2007; Dr. Andreas Braun
Masterplan: Iterations to Develop and Evaluate the Offshoring FrameworkThe Offshoring Framework Initiative started in December 2005 and improves the method and management approach in several iterations. To date, two iterations have been completed, the third iteration (CRM-T) has just been started and implements many recent improvements.
122005
012006
022006
052006
042006
032006
062006
072006
102006
112006
082006
092006
122006
012007
022007
032007
042007
Offshoring GovernanceKöhne, Geimer, (Schütte), Schmitt, Braun
TU-München
Diploma ThesisSakout, Braun
2 Case Studies ÜVBT:Sakout, Digambar
6 Pilots ÜVBT, ETC:Pawlyta, Vaughan, Unland, Morawetz, Kolarik, Meyer-Holtkamp, Klaus, Poller, Staudenmaier.
3 Work Packages CRM-T:CTI, KKM/ BMM, (Testdata)
Offshoring CoordinationKolarik, Paul
CoachingAl-Jadiri Research, Support
Al-Jadiri
????
Improvements:•Template, Taxonomy•Handbook•Training TU-München•TTCN-3 dropped
MBT TrainingProf. Brügge, Al-Jadiri
Improvements:•TWB integration (first attempt as attachments)•Improved taxonomy (test suite preconditions)•Onsite/ offshore coordinators•Test modelling in India, Test architecture in GER•Definition and Tracking of Work Packages
IEEE-Presentation, RioBraun, Brügge, Sakout
Offshore-Phase:Execution, Training
Page 53March 07, 2007; Dr. Andreas Braun
ChargeabilityChargeabilityChargeabilityChargeabilityup to KW40up to KW40up to KW40up to KW40
Model-Based Testing Pilots ÜVBT+ETCWe select six subprojects to evaluate different alternatives of MBT in two large-scale testing projects.
ÜÜÜÜVBT: DMSVBT: DMSVBT: DMSVBT: DMS
AnalysisAnalysisAnalysisAnalysis
Spec.Spec.Spec.Spec.DesignDesignDesignDesign
TestTestTestTestDesignDesignDesignDesign
ExecutionExecutionExecutionExecution
ÜÜÜÜVBT: Customer VBT: Customer VBT: Customer VBT: Customer MgmntMgmntMgmntMgmnt....
ÜÜÜÜVBT: OM NewVBT: OM NewVBT: OM NewVBT: OM New ÜÜÜÜVBT: VBT: VBT: VBT: iKDBiKDBiKDBiKDB
MUCMUCMUCMUC
MUCMUCMUCMUC
MUCMUCMUCMUC MUCMUCMUCMUC MUCMUCMUCMUC
IndiaIndiaIndiaIndia
IndiaIndiaIndiaIndiaTranslationTranslationTranslationTranslation
PlainPlainPlainPlain----Test Spec.Test Spec.Test Spec.Test Spec.ExecutionExecutionExecutionExecution
MUCMUCMUCMUC
MUCMUCMUCMUC
IndiaIndiaIndiaIndia
MUCMUCMUCMUC
IndiaIndiaIndiaIndia
IndiaIndiaIndiaIndia
Germany
India
ETC:ETC:ETC:ETC:KONZ+KONZ+KONZ+KONZ+ÜÜÜÜWS WS WS WS
MUCMUCMUCMUC
MUCMUCMUCMUC
IndiaIndiaIndiaIndia
MUCMUCMUCMUC
ETC: ETC: ETC: ETC: CRMCRMCRMCRM
MUCMUCMUCMUC
MUCMUCMUCMUC
IndiaIndiaIndiaIndia
MUCMUCMUCMUC
Henning Pavel Katharina Andreas Rosemary Dirk
100%
100% 100%
100% 100%
100%
MUCMUCMUCMUCIndiaIndiaIndiaIndia
Deepak**Deepak**Deepak**Deepak** 70%70%70%70%Bhakti **Bhakti **Bhakti **Bhakti ** 75%75%75%75%
Saylee**Saylee**Saylee**Saylee** 95% 95% 95% 95% Awadhoota*Awadhoota*Awadhoota*Awadhoota* 90% 90% 90% 90%
Teststart in Indien in ETCTeststart in Indien in ETCTeststart in Indien in ETCTeststart in Indien in ETCplanmplanmplanmplanmäßäßäßäßig ab KW37ig ab KW37ig ab KW37ig ab KW37
**** Awadhoota Awadhoota Awadhoota Awadhoota worksworksworksworks OnshoreOnshoreOnshoreOnshore******** BlockingsBlockingsBlockingsBlockings werewerewerewere usedusedusedused forforforfor testdatatestdatatestdatatestdata generationgenerationgenerationgeneration....
Page 54March 07, 2007; Dr. Andreas Braun
MBT Pilots Results in ETCThe application of MBT in ETC was successful and led to a high recognition both onsite and offshore. However, due to the project character of ETC, MBT is not applicable as the single method without changing major boundary conditions (client access, project language).
Indian colleagues report the feeling of higher productivity and less overhead.
+ higher productivity
The models are easy to understand and execute
+ easy to understand and execute
MBT highly approved by Indian collegues+ high approval of MBT
Due to the limited scope in ETC, no re-use applicable.Existing test cases need to be translated to UML.
Translation of specification to models takes approx. 1 day/ test case. ETC has > 1000 test cases in the current release.
-- Effort for modeling
English specification is welcomed in India, but basically two languages remain in ETC.
The ETC project has no direct client access. The boundary conditions are set and not easily to change.
Client requires German textual specifications and reporting for review purposes.
-- UML and English not client acceptable for ETC
Artifacts are transferable, tacit knowledge reduced, packaging.
Information is structured and formalized, less inquiries and misunderstandings.
+ Communication with offshore partner is improved
More defects found early in the specification stage
Better structure, complete description, and test data definition enforced by the process and templates.
+ Specification improved
Page 55March 07, 2007; Dr. Andreas Braun
MBT Pilots Results in üVBTThe application of MBT was successful and productive in a real-world project. However, due to the different character of the selected subprojects, the results varied. In summary, MBT is applicable in future ÜVBT projects and will likely continue to improve productivity, chargeability as well as the commitment of onsite and offshore staff.
+ test execution and coaching effort is reduced
Defects are found in earlier stages of testing, the test basis is analyzed more precisely.
+ the quality of the specification is clearly improved
-- make MBT a best practice process
Subsequent versions of connected models generate version conflicts.
-- versioning and maintenance of models open
Needs to be addressed and supported by external advisors (TUM)
Project-dependent guidelines for level of detail are missing.
-- missing or incomplete modeling guidelines
In progress-- Inadequate Tools integration (TWB)
+ well-defined roles & responsibilites
Both onsite and offshore, long-term perspectives.+ improved motivation
+ improved communication
Divide and conquer+ Decoupling of offshore resources
Easy tracking of results, forecast+ Test Suites become work packages
Knowledge repository in models in India created.+ knowledge sharing
For instance, CRM-C models reused in CRM-T, OM models reused.
+ re-use of models and knowledge
Page 56March 07, 2007; Dr. Andreas Braun
MBT Pilots Results in IndiaThe survey included 14 participants. Six indicators are considered for the evaluation and are displayed below. Overall results of each indicator reflect positive impression. However, there are training and coaching needs.
Learning & Implementation Productivity
General Approval
Motivation
Communication EfficiencyQuality of Testing
• Method and Tools simple to learn• For some, Modeling required extraeffort
• More knowledge leads to quicker execution
• work packages allows parallel working
• Modeling makes Testing more interesting and innovative
• Method experienced high acceptance• Rated as highly promising
• Clear communication channels• Less time required to solve an issue• Error localization quicker
• More detailed specifications• More detailed reporting possible
Page 57March 07, 2007; Dr. Andreas Braun
Summary of Results and Lessons LearntThe MBT approach is practical in real-world projects and instantly improves various aspects of both onshore and offshore testing.
Industrialized ProcessProcessProcessProcess Improved specifications Improved test design and architecture Reuse
KnowledgeKnowledgeKnowledgeKnowledge Sharing Encapsulated domain knowledge Prevents tacit knowledge
Organization & PeoplePeoplePeoplePeople Allows for improved organizational forms Increased satisfaction and commitment Reduces overhead
Page 58March 07, 2007; Dr. Andreas Braun
Open IssuesThe MBT approach is practical in real-world projects and instantly improves various aspects of both onshore and offshore testing.
� Depending on the project type, the impact and benefit varies� Offshore Suitability Catalogue!
� Still several open issues and lose ends to be further investigated� Project-specific tailoring and guidelines are missing
� Next steps, continuous improvement� The process is afar from being a “Best Practice”
Page 59March 07, 2007; Dr. Andreas Braun
Any Questions?