CH 11 Lect 2

Object-Oriented Software Engineering
Using UML, Patterns, and Java

and System Testing
Chapter 11: Integration-
Integration Testing Strategy
Subsystem = ?
 The entire system is viewed as a collection of subsystems (sets
of classes) determined during the system and object design.
 The order in which the subsystems are selected for testing and
integration determines the testing strategy
 Big bang integration (Nonincremental)
 Bottom up integration
 Top down integration
 Sandwich testing
 Variations of the above
 For the selection use the system decomposition from the
System Design
Recall: OO vs. architecture vs. design pattern vs. RE – all moving targets
Bernd Bruegge & Allen H. Dutoit Object-Oriented Software Engineering: Using UML, Patterns, and Java 2
Example: Three Layer Call Hierarchy
A
Layer I
B C D Layer II
E F G Layer III
What would A, B and D look like?

Integration Testing: Big-Bang Approach
Unit Test
A
Don’t try this!
Unit Test
B
Unit Test
C
System Test
Unit Test
D
Unit Test
E
Unit Test
F
Bottom-up Integration A
Layer I
B C D Layer II
Test E G
E F Layer III
Test B, E, F
Test F
Test C Test
A, B, C, D,
E, F, G
Test D,G
Test G
Special program needed to do the testing, Test Driver:

A routine that calls a subsystem and passes a test case to it
Pros and Cons of bottom up integration testing
 Bad for functionally decomposed systems:

 Tests the most important subsystem (UI) last
 Useful for integrating the following systems
 Object-oriented systems
 real-time systems
 systems with strict performance requirements
Top-down Integration Testing
A
Layer I
B C D Layer II
E F G
Layer III
Test
Test A Test A, B, C, D A, B, C, D,
E, F, G
Layer I
Layer I + II
All Layers
Special program is needed to do the testing, Test stub :
A program or a method that simulates the activity of a missing subsystem by answering to the
calling sequence of the calling subsystem and returning back fake data.
Pros and Cons of top-down integration testing
 Test cases can be defined in terms of the functionality of the

system (functional requirements) Is this enough?
 Writing stubs can be difficult: Stubs must allow all possible
conditions to be tested.
 Possibly a very large number of stubs may be required,
especially if the lowest level of the system contains many
methods.
 One solution to avoid too many stubs: Modified top-down
testing strategy
 Test each layer of the system decomposition individually
before merging the layers
 Disadvantage of modified top-down testing: Both, stubs
and drivers are needed
Sandwich Testing Strategy
 Combines top-down strategy with bottom-up strategy

 The system is view as having three layers
 A target layer in the middle
 A layer above the target
 A layer below the target
 Testing converges at the target layer
 How do you select the target layer if there are more than 3
layers?
 Heuristic: Try to minimize the number of stubs and
drivers
Sandwich Testing Strategy A
Layer I
B C D Layer II
E F G
Test E Layer III
up
Bottom Test B, E, F
Layer Test F
Tests
Test
A, B, C, D,
Test D,G E, F, G
Test G
Test A,B,C, D
Top Test A down

Layer
Tests
Pros and Cons of Sandwich Testing
 Top and Bottom Layer Tests can be done in parallel

 Does not test the individual subsystems thoroughly before
integration
 Solution: Modified sandwich testing strategy Self reading exercise
Modified Sandwich Testing Strategy Self reading exercise
 Test in parallel:
 Middle layer with drivers and stubs
 Top layer with stubs
 Bottom layer with drivers
 Test in parallel:
 Top layer accessing middle layer (top layer replaces
drivers)
 Bottom accessed by middle layer (bottom layer replaces
stubs)
Self reading exercise
Modified Sandwich Testing Strategy Double
Test I
A
Layer I
Test B
B C D
Layer II
Test E
Triple E F G
Layer III
Triple Test I
Test I Test B, E, F
Double
Test II
Test F
Test
Test D A, B, C, D,
Double
Test D,G E, F, G
Test II
Test G
Test A,C
Test A
Test C Double
Test I
Self reading
Steps in Integration-Testing
1. Based on the integration strategy, 4. Do structural testing: Define test

select a component to be tested. cases that exercise the selected
Unit test all the classes in the component
component. 5. Execute performance tests
2.. Put selected component together; 6. Keep records of the test cases and
do any preliminary fix-up testing activities.
necessary to make the integration
test operational (drivers, stubs) 7. Repeat steps 1 to 7 until the full
system is tested.
3. Do functional testing: Define test
cases that exercise all uses cases
with the selected component The primary goal of integration
testing is to identify errors in the
What is the relationship among packages,
(current) component configuration.
classes, components, interactions, activities,
deployments and use cases?
What would be model-driven software testing?
Self reading
Which Integration Strategy should you use?
 Factors to consider  ...Top-level components are

 Amount of test harness usually important and
(stubs &drivers) big cost cannot be neglected up to the
end of testing
 Location of critical parts in
the system what?  Detection of design errors
postponed until end of
 Availability of hardware testing
 Availability of components  Top down approach
 Scheduling concerns  Test cases can be defined in
 Bottom up approach terms of functions examined
 good for object oriented  Need to maintain correctness
design methodologies of test stubs
 Test driver interfaces must  Writing stubs can be difficult
match component interfaces
 ...
System Testing
 Functional Testing
 Structure Testing
 Performance Testing
 Acceptance Testing
 Installation Testing
Impact of requirements on system testing:

 The more explicit the requirements, the easier they are to test.
 Quality of use cases determines the ease of functional testing
 Quality of subsystem decomposition determines the ease of
structure testing
 Quality of nonfunctional requirements and constraints determines
the ease of performance tests:
Self reading
Structure Testing
 Essentially the same as white box testing.

 Goal: Cover all paths in the system design
 Exercise all input and output parameters of each component.
What’s a a component again?
 Exercise all components and all calls (each component is called at

least once and every component is called by all possible callers.)
How do we know about all possible callers?
 Use conditional and iteration testing as in unit testing.
Functional Testing
Essentially
. the same as black box testing
 Goal: Test functionality of system
 Test cases are designed from the requirements analysis
document (better: user manual) and centered around
requirements and key functions (use cases)
What’s in a use case?
What do we test
.
 The system is treated as black box.
 Unit test cases can be reused, but in end user oriented new test
cases have to be developed as well.
Performance Testing
 Stress Testing  Timing testing

 Stress limits of system (maximum # of  Evaluate response times and
users, peak demands, extended time to perform a function
operation)  Environmental test
 Volume testing  Test tolerances for heat,
 Test what happens if large amounts of humidity, motion, portability
data are handled  Quality testing
 Configuration testing  Test reliability, maintain- ability
 Test the various software and & availability of the system
hardware configurations  Recovery testing
 Compatibility test  Tests system’s response to
 Test backward compatibility with presence of errors or loss of
existing systems data.
 Security testing  Human factors testing
 Try to violate security requirements  Tests user interface with user
Is this list enough? What are the 2 types of requirements?

Test Cases for Performance Testing
 Push the (integrated) system to its limits.
 Goal: Try to break the subsystem
 Test how the system behaves when overloaded.
 Can bottlenecks be identified? (First candidates for redesign in the
next iteration
 Try unusual orders of execution
 Call a receive() before send()
 Check the system’s response to large volumes of data
 If the system is supposed to handle 1000 items, try it with 1001
items.
 What is the amount of time spent in different use cases?
 Are typical cases executed in a timely fashion?
How many abnormal/exceptional cases can there be?

Acceptance Testing
 Goal: Demonstrate system is  Alpha test:

ready for operational use
 Sponsor uses the software at
 Choice of tests is made by the developer’s site.
client/sponsor  Software used in a controlled
 Many tests can be taken setting, with the developer
from integration testing always ready to fix bugs.
 Acceptance test is performed  Beta test:
by the client, not by the
developer.  Conducted at sponsor’s site
(developer is not present)
 Majority of all bugs in software is
typically found by the client after  Software gets a realistic
the system is in use, not by the workout in target environ-
developers or testers. Therefore ment
two kinds of additional tests:  Potential customer might get
discouraged
What’s ethnomethodology?
Testing has its own Life Cycle
Establish the test objectives

Design the test cases
Write the test cases
Test the test cases
Execute the tests
Evaluate the test results
Change the system
Do regression testing
Albert Einstein - We can't solve problems
by using the same kind of thinking we used
Test Team when we created them.
Professional
Tester too familiar
Programmer with code
Analyst
Test System
User Team Designer
Configuration
Management
Specialist
Summary
 Testing is still a black art, but many rules and heuristics are
available
 Testing consists of component-testing (unit testing, integration
testing) and system testing
 Design Patterns can be used for integration testing
 Testing has its own lifecycle
 And …
Additional Slides
Using the Bridge Pattern to enable early Integration
Testing
 Use the bridge pattern to provide multiple implementations

under the same interface.
 Interface to a component that is incomplete, not yet known or
unavailable during testing
Seat Interface Seat Implementation

VIP
(in Vehicle Subsystem)
Simulated
Stub Code Real Seat
Seat (SA/RT)
Bottom-up Testing Strategy
 The subsystem in the lowest layer of the call hierarchy are

tested individually
 Then the next subsystems are tested that call the previously
tested subsystems
 This is done repeatedly until all subsystems are included in the
testing
 Special program needed to do the testing, Test Driver:
 A routine that calls a subsystem and passes a test case to it
SeatDriver Seat Interface Seat Implementation

(simulates VIP) (in Vehicle Subsystem)
Simulated
Stub Code Real Seat
Bernd Bruegge & Allen H. Dutoit
Seat (SA/RT)
Object-Oriented Software Engineering: Using UML, Patterns, and Java 27
Top-down Testing Strategy
 Test the top layer or the controlling subsystem first

 Then combine all the subsystems that are called by the tested
subsystems and test the resulting collection of subsystems
 Do this until all subsystems are incorporated into the test
 Special program is needed to do the testing, Test stub :
 A program or a method that simulates the activity of a missing
subsystem by answering to the calling sequence of the calling
subsystem and returning back fake data.
Scheduling Sandwich Tests: Example of a
Dependency Chart
Unit Tests Double Tests Triple Tests SystemTests


CH 11 Lect 2

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CH 11 Lect 2

Uploaded by

Copyright:

Available Formats

Object-Oriented Software Engineering

Using UML, Patterns, and Java

What would A, B and D look like?

Special program needed to do the testing, Test Driver:

 Bad for functionally decomposed systems:

 Test cases can be defined in terms of the functionality of the

 Combines top-down strategy with bottom-up strategy

Top Test A down

 Top and Bottom Layer Tests can be done in parallel

1. Based on the integration strategy, 4. Do structural testing: Define test

What would be model-driven software testing?

 Factors to consider  ...Top-level components are

Impact of requirements on system testing:

 Essentially the same as white box testing.

 Exercise all components and all calls (each component is called at

How do we know about all possible callers?

 Use conditional and iteration testing as in unit testing.

 Stress Testing  Timing testing

Is this list enough? What are the 2 types of requirements?

How many abnormal/exceptional cases can there be?

 Goal: Demonstrate system is  Alpha test:

Establish the test objectives

 Use the bridge pattern to provide multiple implementations

Seat Interface Seat Implementation

 The subsystem in the lowest layer of the call hierarchy are

SeatDriver Seat Interface Seat Implementation

 Test the top layer or the controlling subsystem first

Unit Tests Double Tests Triple Tests SystemTests

You might also like