The Design of a CASE Environment Architecture and the Performance Evaluation of Database Designs for Software Documents

Luis Miguel

EECS Department
University of California, Berkeley
Technical Report No. UCB/CSD-92-690
June 1992

http://www2.eecs.berkeley.edu/Pubs/TechRpts/1992/CSD-92-690.pdf

CASE environments are the foundation on which software engineering can implement the policies and methodologies needed to efficiently produce the software systems of the future. Our dissertation focuses on CASE data management, in particular how to provide the powerful servics demanded by CASE without sacrificing performance, and minimizing the storage space. Our approach has three main steps:

i) Find out what the requirements are.
ii) Design a system architecture with the requirements in mind.
iii) Choose the database designs for the CASE data structures with the requirements in mind.

In the area of requirements, we compile, analyze, and classify the CASE data management requirements. We also compile, analyze, and classify the different options for data manager, and determine which options satisfy which requirements.

In the system architecture area we analyze the main CASE environment architectures, and propose a new architecture that satisfies the data management requirements discussed earlier. We illustrate the architecture with a hypothetical CASE environment, TULUM.

In the area of CASE database designs we identify and classify the important CASE data structure parameters and the important CASE database schema dimensions. Using this framework we design and conduct a series of experiments to measure the space and efficiency of database designs that vary in 3 dimensions: granularity, data representation, and the number of relations.

Our achievements include a comprehensive list and classification of CASE Database requirements, as well as the design of an architecture that satisfies those requirements efficiently. We also provide a data structure analysis framework that allows database designers to make informed database schema choices. The results of our experiments show that data storage and retrieval have three important phases and that each has to be tuned to prevent bottlenecks, that increasing granularity provides substantial space and speed gains which rapidly level off, that data representation has an enormous performance impact, and that we need to minimize the number of relations and the number of tuples to maximize performance.

Advisor: C. V. Ramamoorthy


BibTeX citation:

@phdthesis{Miguel:CSD-92-690,
    Author = {Miguel, Luis},
    Title = {The Design of a CASE Environment Architecture and the Performance Evaluation of Database Designs for Software Documents},
    School = {EECS Department, University of California, Berkeley},
    Year = {1992},
    Month = {Jun},
    URL = {http://www2.eecs.berkeley.edu/Pubs/TechRpts/1992/6270.html},
    Number = {UCB/CSD-92-690},
    Abstract = {CASE environments are the foundation on which software engineering can implement the policies and methodologies needed to efficiently produce the software systems of the future. Our dissertation focuses on CASE data management, in particular how to provide the powerful servics demanded by CASE without sacrificing performance, and minimizing the storage space. Our approach has three main steps:   <p>i) Find out what the requirements are.   <br />ii) Design a system architecture with the requirements in mind.   <br />iii) Choose the database designs for the CASE data structures with the requirements in mind. <p>In the area of requirements, we compile, analyze, and classify the CASE data management requirements. We also compile, analyze, and classify the different options for data manager, and determine which options satisfy which requirements. <p>In the system architecture area we analyze the main CASE environment architectures, and propose a new architecture that satisfies the data management requirements discussed earlier. We illustrate the architecture with a hypothetical CASE environment, TULUM. <p>In the area of CASE database designs we identify and classify the important CASE data structure parameters and the important CASE database schema dimensions. Using this framework we design and conduct a series of experiments to measure the space and efficiency of database designs that vary in 3 dimensions: granularity, data representation, and the number of relations. <p>Our achievements include a comprehensive list and classification of CASE Database requirements, as well as the design of an architecture that satisfies those requirements efficiently. We also provide a data structure analysis framework that allows database designers to make informed database schema choices. The results of our experiments show that data storage and retrieval have three important phases and that each has to be tuned to prevent bottlenecks, that increasing granularity provides substantial space and speed gains which rapidly level off, that data representation has an enormous performance impact, and that we need to minimize the number of relations and the number of tuples to maximize performance.}
}

EndNote citation:

%0 Thesis
%A Miguel, Luis
%T The Design of a CASE Environment Architecture and the Performance Evaluation of Database Designs for Software Documents
%I EECS Department, University of California, Berkeley
%D 1992
%@ UCB/CSD-92-690
%U http://www2.eecs.berkeley.edu/Pubs/TechRpts/1992/6270.html
%F Miguel:CSD-92-690