Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Rule-based automatic software performance diagnosis and improvement
Xu J. Performance Evaluation69 (11):525-550,2012.Type:Article
Date Reviewed: May 2 2013

There are well-documented advantages to automating performance evaluation during software design rather than waiting until system testing to uncover design problems. This paper addresses improving the performance of a design by identifying incremental configuration changes, but does not compare unique and significantly different design alternatives. The focus is the extension of a new performance evaluation tool known as Performance Booster (PB).

The paper’s title and abstract are somewhat ambiguous as to the scope of the automation addressed by the PB tool. The description of the work begins with the extraction of a layered queueing network (LQN) from a unified modeling language (UML) specification, and then progresses through a series of rule-based performance analyses. Xu concludes by mapping the proposed design changes back to the UML specification. PB automates only a small part of the task. The extraction of the LQN is performed by another toolset, PUMA, and the extraction of performance measures is supported by Jess, an open-source tool. Mapping the design changes back to UML is a nontrivial manual task.

The rule-based analysis conducted by PB compares the mean response time, mean throughput, and probability a resource is busy (resource utilization value) for the design under analysis with various design changes. The improvement amounts to incremental changes to baseline mean values of the design. The author does provide detailed descriptions of the two semantic networks used in PB, the application of model parameters, the search algorithm, and the rule base. The numerous case studies provided are not described uniformly. The paper would be more informative if the author had provided a single case study with documentation from the beginning to the end of the seven-step PB information flow, rather than several disparate examples.

Reviewer:  Nancy Eickelmann Review #: CR141195 (1308-0713)
Bookmark and Share
  Featured Reviewer  
 
Computer-Aided Software Engineering (CASE) (D.2.2 ... )
 
 
Performance Measures (D.2.8 ... )
 
Would you recommend this review?
yes
no
Other reviews under "Computer-Aided Software Engineering (CASE)": Date
CASE: computer-aided software engineering
Lewis T., Van Nostrand Reinhold Co., New York, NY, 1991. Type: Book (9780442003616)
Jul 1 1991
AD/Cycle
Montgomery S., Van Nostrand Reinhold Co., New York, NY, 1991. Type: Book (9780442308254)
Sep 1 1992
The three Rs of software automation
McClure C., Prentice-Hall, Inc., Upper Saddle River, NJ, 1992. Type: Book (9780139152405)
Jul 1 1993
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy