Subscribe by Email

Monday, June 25, 2012

How is optimization of installation testing done?

It becomes very important to carry out installation testing towards the end of the release process. It is very important to subject a software system or application to installation testing so that it can be released without any hesitation to the customers. 
One can make sure without installation testing that the software system or application will successfully install on the user’s system in the first go itself! The user might need to install the software system or application again and again which in turn might frustrate the user. You will loose customers as a consequence of this. Therefore perform installation testing becomes absolutely necessary. 
In this article we have discussed about the need for optimizing the installation testing and how to optimize it. by the optimization of the installation testing we mean modifying it further so that  some aspects of it use less resources and work more efficiently. 

What is the need to optimize?

- When you optimize a testing process you can run it more quite rapidly and that too it uses very less memory then. 
- The optimization though has been named as “optimization”, it can never produce a truly optimal system. 
- The amount of time taken by a test case of the installation can be reduced by making it consume more memory.
- In cases where the memory is limited, slower algorithms can be used.
- While optimizing the installation testing one should keep in mind that there is no such design that fits all the test cases. 

On what levels installation testing is optimized?

The installation testing process much like the other processes is optimized through the following number of levels:

Design Level: 
- This is the highest level of the above mentioned process. 
- Design is the best aspect that can be optimized to utilize the resources in a much better way.
- The implementation of the optimized design can be made more efficient by using efficient algorithms which in turn have been benefited from code of good quality. 
- It is obvious for the architecture of the test design to have an overwhelming affect on its performance.
- In some cases optimization process may also involve the following:
(a)  More elaborate algorithms
(b)  Special tricks
(c)   Special cases
(d)  Complex trade offs
- A fully optimized installation testing process becomes difficult to comprehend and may also house more errors than the un- optimized processes.

Source code Level: 
- This step involves rooting out the bad quality code from the test cases and therefore it also improves performance.
- When the malfunctioning code is removed, the obvious slow downs are avoided. 
- However, it is possible that any further optimization made to that test case may decrease its maintainability. 
- But, today there are available what are called “optimizing compilers”.

Compile Level:
- At this stage, it is ensured by the optimizing compiler that the test case is optimized to such a level at which the test case can be compiled by the compiler.

Assembly Level:
- This is the lowest level of optimization process and involves taking advantage of the full repertoire of system instructions. 
- This is the reason why most of the operating systems use the embedded systems are written in assembly code.
- Many of the test cases are compiled from a high level language to assembly language and from there it is manually optimized.

Run time: 
- Run time optimization is needed and are performed by the assembler programmers and just in time compilers which exceed the capability of static compilers.

Two types of optimizations can be performed on the installation testing process namely:
  1. Platform dependent optimization (uses specific properties of one platform) and
  2. Platform independent optimization (effective on all platforms)

No comments:

Facebook activity