Software design and embedded system tools. » Software Engineering

Why Is Load Testing Useful

admin — Sun, 25 Jul 2010 15:50:29 +0000

Load testing is the process of putting increasing demand on a system or software until the system or software breaks or fails to function. The purpose of load testing is to test the system against unusually high usage patterns to determine how the system or software will react.

One example of the need for this type of testing is a web server hosting a web site that is seasonal or receives peak levels of web traffic during highly publicized special events. Another example would be a link to a small web site from a very popular site, that then doubles or triples the number of concurrent visitors to the smaller web site. Load testing in each example could determine ahead of time whether each site could handle the increased traffic without the hosting server failing to serve web pages to visitors during the peak load.

Software load testing

Software programs can be load tested and can be particularly useful if they are network capable as most modern software packages are today. Testing is usually performed by the software company itself, rather than individual users or business users unless determining maximum capacity is critical to the enterprise. Ordinarily however, software will come with a set of benchmarks, stated maximum concurrent users and other statistics that are useful in determining suitability to purpose, often removing the need for individual testing.

With common software packages, like a word processor application or graphics editor application, these can be tested by repeated attempts to load a very large document. Will the document load successfully? Or will a suitable error message be displayed to confirm that the document is beyond the capacity of the application to handle it? Will the application simply crash with or without warning? In the case of other applications like a database that is setup to only handle 100 concurrent users, like some versions of Microsoft Access database application, what happens when the 101st user is attempting to connect to the database? Different software makers will handle the software limits in various ways and it can be helpful to be aware of these. In addition, web applications need to be written to display appropriate messages to the web user if maximum capacity has been reached.

Web server and web application load testing

The World Wide Web (the Internet) is run by networks, and web servers are utilized to offer web site functionality to web users. A lot of the software used on these networks and servers today was designed for the Internet, but quite a few have been inherited and was not specifically designed for it.

Early versions of Microsoft Access are a case in point. The 100 concurrent user limit was fine for most enterprises using a network version of the package. As long as each user quickly accessed the database, ran their database query, saved the results and then disconnected from the relational database management system (RDBMS), hitting the concurrent user limit could be avoided in many cases. Depending on the uses for the database, 1,000 users across an enterprise could have access to the RDBMS as long as no more than 100 employees did so during the same few seconds a database query was being ran.

Some web hosting companies still offer earlier versions of Access for free with their hosting packages and this can create problems for web developers that create a web application, utilizing Access as the primary RDBMS, never anticipating the web site becoming anything but modest success. The concurrent user limit could be quickly reached if traffic grew beyond expected levels and the web site would become inaccessible for many users.

Load testing in the real world

Load testing is useful to determine the maximum usable load for an application, be it standalone, networked or web-based. This can help when attempting to predict outcomes should demand exceed estimates by the wide margin. Software packages will usually come with some internally tested load suggestions but it is useful to conduct your own tests when the application is critical to business operations.

Thank you,
Jitkasem Pintaya

Scrum and Extreme Programming

admin — Thu, 03 Jun 2010 13:39:15 +0000

Agile Process (Scrum and Extreme Programming)
The Agile Process or software development refers to a set of software development methods which are based on iterative development. In this process, the solutions and requirements both evolve mutual collaboration between cross functional teams. These teams are self-organizing in nature.
The Agile software development approach usually promotes a regimented kind of project management process which encourages:
* Frequent adaptation and inspection
* Self-organization and accountability
* A leadership philosophy which promotes teamwork
* A business approach which bring into line the development with customer needs and company goals
* And a group of best engineering practices having an intention to allow for rapid delivery of good-quality software.
Extreme Programming (EP)
It is a software development methodology with an intention to enhance software responsiveness and quality to the volatile requirements of customers. Being a type of Agile process, it promotes frequent releases in small development cycles. This introduces checkpoints and improves the productivity in a way that the new requirements from customers can be adopted.
The advantages of Extreme Programming are:
* Unit testing of all code
* Avoiding programming of features until needed
* Programming in pairs or carrying out extensive code review
* Clarity and simplicity in code
* Volatile customer requirements better understood
* A flag management structure
* Frequent communication between the programmers and even with the customer
It has been noted however that there are certain drawbacks with Extreme programming. They are:
* No documented compromises of user conflicts
* Unstable requirements
* Lack of overall design document or specification
* Incorporates inadequate software design
* Necessitates meetings at recurrent intervals at huge expense to customers
* Can enlarge the risk of scope creep due to the lack of thorough requirements documentation
* Requires excess of cultural change to adopt

Scrum
Scrum is an incremental, iterative framework for agile software development and project management. The word “Scrum” is not really an acronym. However, many companies using this methodology spell it with capital letters. Initially Scrum methodology was focused for management of software development projects, but in recent times it can be used to run general program/project management approach or software maintenance teams.
Scrum, which contains sets of predefined roles and practices, is a process skeleton. Main roles in this method are:
1. Product Owner – Represents Stakeholders
2. Scrum Master – The one responsible for maintaining the processes
3. Team – A cross functional group of about 6-8 people who do actual design, testing, implementation, etc.
Each of the iteration is called a “sprint”, typical time frame for which is normally about two to four weeks. The length of each sprint is decided by the team. The product “backlog” pushes the set of features into a spring. These features are prioritized set of higher level requirements for the task to be carried out. Based on this product “backlog”, the team determines how many of the items can be completed in the next sprint. Once the sprint begins, nobody is allowed to alter the sprint backlog, which means that the set of requirements are frozen. On successful completion of a sprint, the team demonstrates the usage of that particular software.
This methodology should be encouraged in organizations since the major advantage of using the Scrum is that it enables the creation of teams which are highly self-organizing in nature. This is achieved by encouraging verbal communication amongst the team members, co-location of all the team members and disciplines which are involved for the project.

Differences between Scrum and Extreme Programming(EP) :
1. The time span for iterative sprints is different in both approaches.
2. Changes are not allowed by the Scrum teams during their sprints. Whereas Extreme Programming teams have to be much more agreeable to changes.
3. Work is done by EP teams in strict priority order. Whereas in case of Scrum, the product owner prioritizes the set of activities.
4. EP does prescribe some engineering practices; Scrum does not.

Thank you

Incoming search terms for the article:

Performance Testing and Regression Testing

admin — Wed, 02 Jun 2010 13:29:51 +0000

Performance Testing and Regression Testing

Software testing or Testing is an examination conducted to present the stakeholders with data regarding the quality of the service or product under test. This type of software testing as well provides in insight on the independent views and objectives of the software to permit the business to understand all the risks involved during the implementation of the software.
Based on the testing method employed, Software testing can be put into practice at any point of time in the development process. Although, most of the efforts for testing occur after the requirements are defined and the process for coding has been completed.

This article describes two types of testing methodologies.

1. Performance Testing
Performance testing is performed to determine how quick a system or sub-system executes under a particular amount of workload. It can also provide to verify and validate other quality traits of the system, such as reliability, scalability and resource usage.

In the field of Software Engineering, performance testing comes under the testing category which is carried out to how fast a particular aspect of the system under observation performs, given a fixed amount of workload. This sort of testing is a subset of Performance Engineering (up-and-coming computer science practices build to get performance into the architecture and design of a system, prior to the actual effort of coding). There are various purposes which performance testing can serve:
It can compare given two systems and finds out which one is performing better
It can demonstrate whether or not the system meets the required performance criteria.
It can calculate which parts of the system under observation cause the system to perform badly.

2. Regression Testing
This type of testing is any software testing which seeks to find out software errors by partly re-testing a modified program. The intention of this type of testing is to provide a kind of assurance that no additional errors were added during the process of fixing the existing problems. Regression testing is generally used to proficiently examine the system by systematically choosing the appropriate suite of tests which are required to sufficiently cove all of the affected changes.
Widespread methods of regression testing comprise of rerunning earlier run tests and scrutinizing whether previously fixed faults have re-emerged. One of the chief reasons for performing regression testing is that it’s often tremendously difficult for a programmer to outline out how a change in one part of the software will reverberate in other parts of the software.

Ad hoc testing
This term is commonly used to represent software testing without any documentation or planning. The tests are made to run only once, unless a defect is discovered. This type of testing is more of a part of exploratory testing. Here, the tester seeks to find errors by any means that seems appropriate to him/her.

Unit Testing
This type of testing is a software verification and validation method where a programmer tests if individual units of system are fit for utilization. A unit is the least testable part of an application. In procedural programming, a unit might be an individual procedure or function.

Advantages

The advantage of performing Performance and Regression testing are:
Reusable: Can reuse tests on diverse versions of an application, even if the user interface alters
Repeatable: Can check how the software responds under repeated execution of the identical operations.
Programmable: Can program complicated tests that bring out concealed information from the application.
Cost Reduction: Cost is reduced since the amounts of resources for regression test are reduced.
Reliable: Tests carry out precisely the equivalent operations each time they are run, thereby eradicating human error.
Comprehensive: Can build a suite of tests that wraps every feature in the application.
Better Quality Software: Can run additional tests in fewer time with less resources
Fast: Automated Tools run tests considerably faster as compared to the human users.

Incoming search terms for the article:

GIT vs Perforce vs Clearcase vs CVS

admin — Tue, 01 Jun 2010 14:38:33 +0000

GIT

This distributed revision control system has an emphasis on speed. It was originally designed and developed by Linus Torvalds for the development of Linux kernel. Every working directory of Git is a full-fledged repository with full revision tracking capabilities and complete history, not dependent on a central server or network access. GIT is a free software which is distributed under the terms of version 2 of the GNU General Public License.

Perforce

This is a proprietary and commercial revision control system developed by Perforce Software Inc. Perforce follows a client/server system. Here the server has the responsibility of managing a central database and a main repository of file versions. Developers can work on files in their local client workspace. After altering the changes, they can submit the changes to the main server. Client and server communicate via TCP/IP.

Clearcase

Rational ClearCase is a tool for configuration management of source code and other software assets. This tool is developed by IBM (Rational Software Division). ClearCase is the basis of version control for many mid size businesses, and has the potential to handle projects with thousands of developers.

CVS

This is a free revision control system. This system was developed by Dick Grune as a series of shell scripts during July 1986. CVS turned out to be popular in the open source software space and is released under General Public License of GNU. CVS also uses Client/Server architecture. Here, the clients connect to the server and can “check out” a version of file onto their local disk. Later after modifying the contents, they can “check in” the changes.

Incoming search terms for the article:

Types of Revision Control Software

admin — Mon, 31 May 2010 13:00:11 +0000

Revision Control Software
Revision Control software are used implement Revision Control on single or multiple sites by organizations. Revision Control, also known as Software Configuration Management (SCM) or Source Control or Version Control, is the management or control of changes to programs, documents and other sorts of information data. This type of setup or mechanism is more often used in software development where a set of people may alter the contents of the same file. The changes made are generally identified by a letter code or number which is called as a revision number or simply revision. For instance, an initial revision for a file is “revision 1”. If this file is altered by any developer, its revision becomes “revision 2”, and so on. A developer can perform many operations on these revisions. These revisions of files or directories can be merged, restored or compared. Software tools for revision management are crucial for the organization where development is simultaneously in progress on multi-developer sites.

Types of Revision Control Software
The various possible approaches for revision control software are:
Local Only: In this approach, it is necessary for all the developers to use the same computer system. Examples:
a. Revision Control System (RCS)
b. Source Code Control System (SCCS)

Distributed Model: In this approach, each of the developer has his/her own setup or local repository. And at some predefined time, all of the changes from various developers are shared or synced between repositories as a different step. Examples:
c. Open Source:
i. Aegis
ii. Codeville
iii. DCVS
iv. GIT
d. Proprietary:
i. Sun WorkShop Team Ware
ii. BitKeeper

Client-Server Model: In this approach, all the developers have to share a single repository. Examples:
e. Open Source:
i. Concurrent Versions System (CVS)
ii. CVSNT
iii. OpenCVS
f. Proprietary:
i. Perforce
ii. Clearcase
iii. Visual SourceSafe

Incoming search terms for the article:

User Acceptance Testing.

admin — Tue, 10 Nov 2009 13:51:01 +0000

UAT (User Acceptance Testing)

UAT is a sort of testing that usually done at the last step of software development process (before release the software to the market).UAT is done by user (the person who will use this software or person who possess requirement of software).It is the way to increase reliance and ensure that the software was designed correctly based on requirements.

Before doing UAT your software need to pass Unit Testing,Integration Testing,System testing and must solved the defects found in each of these testing phases (in practical, the defects relate to technical or logic of software should be fixed before sending to UAT).The defects concern with appearance or color of software are acceptable during doing UAT.

Test Case is also need in UAT because it help doing UAT more effective. Otherwise the test case should be designed on the perspective of Bunsiness Flow and Business Logic that based on requirement. The tester will emphasize on testing the software like they are real user.

Procedure for doing UAT.
One important factor you must focus on while you doing UAT is Test Environment.The test environment should be similar to Production Enviroment as much as possible.

Basic step of doing UAT are as follow…

-Make the testing plan for UAT.
-Design the test cases for UAT.
-Set up testing team for operate UAT in accordance with test cases.
-Operate UAT.
-Record Bug or Defect found during operate UAT and hold a meeting in order to summarize anything we have to modify.
-Fix bug or defect found in previous step and then do Regression Testing.
-Finish UAT, sign off.

UAT is classified as Black box testing.The test cases for UAT normally designed according
to requirements of user (after finish requirements definition phase). Business Analysis
team or System Analysis team are also provide the inputs used for creating the test cases
and it should be described using simple language which easy to understand because the
tester may be the person other than programmer or tester and they don’t know the
technical phrase.

Visit these site for more information about UAT (User Acceptance Testing)
-Wikipedia
-eHow
-Exforsys INC

Reference for this article:
-http://www.welovebug.com/user-acceptance-testing/what-is-uat/#more-1503

Thank you,
Jitkasem Pintaya.

Incoming search terms for the article:

How to do white box testing

admin — Tue, 13 Oct 2009 15:27:12 +0000

The article for today is about white box testing.I had written article about white box testing in this blog but that time it was only introduction of white box testing.This article concern with how to do white box testing.

The procedure of White box testing is normally saparated into 3 topic as follow

1.Statement Coverage
It is the testing to verify that all instructions or code must be executed at least one time.We can specify one or more than one test case for doing statement coverage testing (there may be a chance that one test case is not cover all the code we need to test).Below is an example demonstate how to do statement coverage testing.

Suppose you have a piece of code to test like this

Code1

If you determin ‘C’ as the test case then the coverage reach 100% (pass through all line of code) and only one test case is sufficient for doing Statement Coverage for this code.

Advantage of Statement Coverage.
1.To ensure that all source code can be reached at least for one time.
2.It is suitable for making Performance profiling because it can be applied directly to object code without processing source code.

Disadvantage of Statement Coverage.
1.It is insensitive to logical operators such as OR (||),AND (&&) etc.
2.If your source code have a lot of IF-ELSEIF-ELSE statements or contain consecutive switch labels then the test case used for doing Statement Coverage will be increased respectively.

2. Branch Coverage (Condition judgment Coverage)
It is the testing to verify that all branches and instructions must be executed at least one time (Concentrate on testing to cover every branches or decision points in the code).It is the easiest way to perform testing of path or workflow of program.

Please consider the code below and its explanation to understand the operation of Branch Coverage testing.

code2

In the above code we need only two test case to cover all the branches (also all instructions too), they are…
Test case #1 : Input n1 = 1 , n2 = 1
Test case #2 : Input n1 = 2 , n2 = 0

Advantage of Branch Coverage.
1.To validate all branches in the code can be reached and ensure that no branches lead to abnormal of the program’s operation.
2.It eliminate problems that occur with Statement Coverage testing.

Disadvantage of Branch Coverage.
1.There may be other condition that can be used for decision making.For the example, statement “if((b == TRUE) || functionA())”, in this statement we can define only test case with “B = TRUE” and this test case is cover this branch but functionA() isn’t tested.

3.Condition Coverage/Multiple Condition Coverage

It is the testing to verify that all condition expression within each branch will be tested (the true and the false condition of each sub-expression within the decision branch must be tested at least one time ).
It is more intensive testing than Branch Coverage and Statement Coverage.Multiple Condition Coverage has detail of testing more than Condition Coverage.Its test cases are larger than test case of Condition
Coverage because it ensures that all possible combination of conditions of sub-expression within each branch must be tested.

Please consider the code below for Condition Coverage testing.

code3

If you want to test above code by using Condition Coverage you must specify two test
cases to cover all sub-expressions in this branch.The test cases are as follow…

Test case #1: Input n1 = 60 , n2 = 70
Test case #2: Input n1 = 40 , n2 = 110

If you want to test above code with Multiple Condition Coverage you must specify four
test cases to cover all combination of possible condition of sub-expressions in this branch and the test cases look like these

Test case #1: Input n1 = 60 , n2 = 70
Test case #2: Input n1 = 60 , n2 = 110
Test case #3: Input n1 = 40 , n2 = 70
Test case #4: Input n1 = 40 , n2 = 110

Advantage of Condition/Multiple Condition coverage.
1.It is very thorough testing and the bugs are normally found by this kind of testing.

Disadvantage..
1.If the decision branch contain lots of sub-expressions or has very complex boolean expressions , the tester will have to define a large number of test cases.

Thank you,
jitkasem pintaya.

Reference site:
1.MindBlaze Blog
2.Bullseye
3.The paper from utwente.nl by Arnold Zanderink

Incoming search terms for the article:

CASE tool overview.

admin — Sat, 05 Sep 2009 14:34:32 +0000

CASE tool (Computer-Aided Software Engineering)

CASE tool is the software that support system development or facilitate in system development.It
has an ability in building various kinds of document during analysis and design the system.For the example such as building the UML diagram,the screen interface (Graphical User Interface),form and report etc…
Furthermore some CASE tool can automatically generate source code too.In system maintenance can use CASE tool to make the operation smoothly,quickly and correctly.CASE tool also help you to decrease the time in document improvement or source code improvement.So at the present time CASE tools have an important role in system development and system maintenance.

Formerly CASE tool was separated into two groups as follow…
1.Upper CASE tool is the software that used in the step of analysis and design the system.This CASE tool make various diagram such as DFD (Data flow diagram) . DFD is used for design the component of the system such as the input of system , the mechanism of system , the result of the system etc.Other diagrams such as ER diagram (Entity-relationship diagram),layout (the screen that used for bring the data into the system or screen of the report). The example of CASE tool in this group are ERwin and Visual UML.

2.Lower CASE tool is the software that can build the source code or program automatically from condition or determinate procedure such as we can write the program for manage the vacation of our employee from the specific workflow or we can write the database structure command by select from the menu.The example of CASE tool in the group are Ecore Diagram Editor and dzine.

3.Integrated CASE tool is software that fulfill the usability from system analysis,system design to the system implementation (writing the source code) and system testing.It is the software that combine Upper CASE tool and Lower CASE tool into one software.The example of CASE tool in this group is Rational Rose from IBM.

The story of CASE tool is not finish by this article.In the next article I will write about the review of CASE tools that are well-known nowadays such as MagicDraw, Visual Paradigm,Poseidon UML and the suggestion about Commercial CASE tools and Open Source CASE tools software list.

for more information please visit these sites
1.UML Modeling Tools
2.Wikipedia
3.CASE tool index

Thank you,
Admin

Incoming search terms for the article:

Software testing part II (White Box and Black Box testing)

admin — Wed, 19 Aug 2009 16:16:57 +0000

White Box testing
Another name of White box testing is Glass-Box testing or Structural Testing.
It is the test in the system.The tester have to specify the test case that will be used
for test the system.They must have an ability in programming and well understand in
software testing theory.Their task it to assign the test case to every line, every
statement or every path of the program being tested. You can compare white box
testing with the measurement of electronics circuit (In measurement of electronics
circuit,you have to measure the electric current and voltage in every node existing
in your circuit).

White box testing can be applied for testing in Unit testing,Integration testing
or System testing but in commonly the tester apply White box testing in Unit testing phase
of software development.

The advantage of White box testing.
1.The process of White box testing can increase te quality of source code and can make
the pieces of source code work more efficiency because it is the testing in the system
(In the function and in every branches of the code)

2.In White box testing we can look for an error by give an input to the system and monitor
the output from the system.Otherwise it also can look for the internal error that existing
in the mentioned system too.

The disadvantage of White box testing.
1.This kind of testing require high skill in software coding to assure the high-quality of
source code.

Black Box testing.
Another name of Black-Box testing is Functional Testing.It is the testing
that disregard internal mechanism of the system (don’t test the internal component
of the system such as source code or function).This testing method is focus on the
output that come out from the system after we sent the input to it (Output that
come out after the system responded to the input data).

In Black-Box Testing, the tester will not look for the code to test, so the code
is considered to be “Black Box” that we can’t see the content inside the box.The
tester know only they must send the input to the “Black Box” and then it will release
the output to the tester.

The tester usually use the requirement specification document (Requirement
Knowledge) to build the test case, so they know what is the outcome that the system
will send after they send the corresponding input to it.

Now I can summarize the black box testing as follow…
– It is the testing that disregard the source code or command
in the program.
– It is testing of function of system base on the requirement documents.
– The tester must assume for output that will come out from the system
after sending the various input pattern to the system (The output must match up with
the input).

Now I can summarize White Box Testing as follow…
-It is the testing that examine the structure or the work-flow of the program .
-The tester must build the specific test case used for testing in each specific condition.
-The test case must contain the case that can be executed normally and abnormally.
-Try to execute every statement in the function at least 1 time.

In the next article I will write about Test first method (Software testing part III).
You can see more information about Black box and White box testing at these websites below
1.Software Testing Fundamentals
2.Testing and Code Coverage

Reference site: lenovoblogs

Thank you,
Jitkasem Pintaya.

Incoming search terms for the article:

Software testing part I.

admin — Wed, 12 Aug 2009 07:29:51 +0000

There are no commercial software released to the market without testing.
Testing is an important process in software development.The big or
well-known software company place importance on software testing equal to
software implementation.
At present there are many type of software testing method but I summarize the
widely used testing method as following…

1.Unit test
Unit test is the testing functional
(Or testing in detail of each function in the system) to ensure
that each individual function in the system can work well.
Unit testing is very important in software development because
it is the method for finding the functional problems (It can decrease
the time in finding bugs in the posterior process).
2.Integration test
Integration test is the testing that take each unit to
combine altogether and then test the coherence between each unit to
ensure that the communication between each unit in the system can
be done correctly.
3.System test
System test is the testing of the whole system to ensure that
the totally work-flow of the system is correct.It is also
an important test that have to do in software development process
(Normally system test is done as the last test of all
testing procedure).
4.Acceptance Test
Acceptance test is the testing of customer or user who will
use the developed software.It is the test for satisfaction of the
user or customer.This testing is done after System test.

The important testing that you must emphasize on them are
Unit testing and Integration testing because many defect or bugs
can be found by these two testing method.

Testing is not necessary to be done by tester but developer
can also do the testing.Normally the Unit testing is done by
developer because the majority of this testing use the test case
that depend on the implementation, so nobody can do this testing
better than developer.However the real tester have a responsibility
for doing the Integration test and System test because…

– These two kinds of testing must do in long time.

– To prevent the bias if the developer is also the tester at the
same time.(The developer who write the code maybe
rather sure that the code they wrote is correct and will
ignore some minor part of code that have to test)

– The real tester has more time to design the test document
and test case. He/She must know clearly in the target system
that will be tested,so He/She can do the testing better than
developer.

This is part I of software testing story.In the next part (part II)
I will write about black box and white box testing,test first method and
how to design the test plan.

Thank you very much,
Jitkasem Pintaya.