SOFTWARE TESTING
ISAS SEMINAR ON
Presented by Jaikumar Sudharshan
CONTENTS
1. 2. 3. . #.
Introduction Testing objectives Testing principles !lassi"ication o" testing $lack bo% testing
CONTENTS
1. 2. 3. . #.
&hite bo% testing Per"ormance testing 'eliability testing Security testing Testing automation
CONTENTS
1. 2. 3. . #.
(vailable Tools (nd Techni)ues Testing level Testing cycle Summary *ueries
INTRODUCTION
So"t+are testing, Testing a so"t+are "or its !orrectness !ompleteness Security.
INTRODUCTION
S*( - So"t+are *uality (ssurance is testing a so"t+are product completely in a broader aspect. So"t+are testing is a part o" S*(.
TESTING OBJECTIVES
Testing
. e%ecuting a program to "ind an
error *uality o" a test case, probability o" "inding an as/yet undiscovered error. ( success"ul test uncovers a lot o" undiscovered error
TESTING PRINCIPLES
(ll tests should be 0ade according to customer re)uirements planned long be"ore testing begins begin in/the/small and progress to+ard in/ the/large 1%haustive testing is not possible
Classification of testin
It is broadly classi"ied based on point o" vie+ a test engineer takes as $lack/bo% testing &hite/bo% testing
BLAC! BO" TESTING
2nly
inputs and output "unctionality are visible Inputs are given and outputs are compared against speci"ication. 3o implementation details o" the code are considered.
4ata driven testing.
BLAC! BO" TESTING
Targets, To "ind incorrect or missing "unctions inter"ace errors per"ormance errors
W#ITE BO" TESTING
!ontrary
to black/bo% testing. structure and code are visible. Implementation o" the so"t+are is given importance programming language5 logic5 and styles are considered.
W#ITE BO" TESTING
Targets, To check !ontrol/"lo+ 6ooping data/"lo+ all the nodes or paths are covered or traversed at least once.
PERFORMANCE TESTING
so"t+are
should not take in"inite time or in"inite resource to e%ecute. P1'72'0(3!1 $89S - process that causes system per"ormance to degrade. 8sage o" +orkload to check so"t+are per"ormance.
PERFORMANCE TESTING
Targets, To check !P8 cycles. 4isk space. 4isk access operations. 0emory usage.
RELIABILIT$ TESTING
probability
o" "ailure/"ree operation o" a
so"t+are. 'obustness testing / degree to +hich it can "unction correctly in the presence o" e%ceptional inputs.
RELIABILIT$ TESTING
Targets, To check "or robustness problems such as machine crashes. process hangs. abnormal termination.
SECURIT$ TESTING
3eed "or security testing, 7la+s in so"t+are can be e%ploited by hackers. &ith the development o" the Internet5 so"t+are security problems are becoming even more severe.
SECURIT$ TESTING
Targets, To check security violations. validating the e""ectiveness o" security measures. Simulated security attacks to "ind security vulnerabilities.
TESTING AUTOMATION
The need, So"t+are testing can be very costly. (utomation is a good +ay to cut do+n time and cost. In order to automate the testing process +e have to generate the test cases to check "or complete correctness.
TESTING AUTOMATION
Today
+e still don:t have a "ull/scale system that has achieved this goal. signi"icant amount o" human intervention is still needed in testing. The degree o" automation remains at the automated test script level.
AVAILABLE TOOLS AND TEC#NI%UES
( lot o" testing tools are available. some o" them are Mothora : mutation testing tool developed at Purdue university. !an create and e%ecute test cases5 measure ade)uacy 4etermine input - output correctness5 remove bugs
3. . #.
AVAILABLE TOOLS AND TEC#NI%UES
2. 3. #.
NuMega's Boundschecker, National's Purify 'untime checking and debugging aids. !heck and protect against memory leaks and pointer problems. Ballista COTS Software Ro ustness Testing !arness , 7ull scale automated robustness testing tool.
TESTING LEVELS
"cce#tance
testing - system re)uirement/ per"ormed a"ter testing and be"ore implementation phase. "l#ha testing / testing by an independent test team at developers: site. Beta testing $ eta %ersions released outside the co&#any so&eti&es to o#en #u lic to increase feed ack'
TESTING C$CLE
SUMMAR$
Testing
is a destructive activity. Testing cannot sho+ the absence o" de"ects. Testing ; 3<= o" total e""ort.
SUMMAR$
(ll
tests should be traceable to customer re)uirements. &hite/bo% testing. $lack/bo% testing
SUMMAR$
So"t+are
)uality5 reliability and security are tightly coupled. 'obustness5 per"ormance5 reliability testing.
SUMMAR$
>Testing automation >Testing cycle >Testing levels
%UERIES
