Designing User Interfaces Spring 1999

1
SE 767-NT Software Performance Engineering Robert
Oshana Lecture 13 For more
information, please contact NTU Tape
Orders NTU Media Services [phone]
oshana_at_airmail.net
tapeorders_at_ntu.edu
2
Where we are

• Introduction
• SPE Quick View
• SPE and the UML
• Software Execution Models
• Web applications and other distributed systems
• System execution models
• SPE data collection
• Software measurement and instrumentation
• Performance oriented design
• Performance patterns
• Performance anti-patterns
• Implementation solutions
• Web applications
• Embedded and real-time systems
• The SPE process
• Implementing SPE

3
SPE data collection

• Chapter 7

4
Introduction

• Until now we assumed we had the necessary data
  for our models
• Hardest part of SPE is getting the data !
• Need to collect data early enough in the life
  cycle

5
SPE data requirements

• Key performance scenarios
• Performance objectives
• Execution environment
• Software resource requirements
• Computer resource requirements

6
Key performance scenarios

• Scenarios make it possible to isolate and extract
  the processing steps that are associated with a
  particular execution of a use case
• First identify critical use cases
• Performance affected by large tasks
• Estimate workload intensity

7
Performance objectives

• This specifies quantitative criteria for
  evaluating the performance characteristics of the
  scenarios in the system being developed
• Hard to know when you have gotten there

8
Quantitative performance objective benefits

• Basis for selecting architecture
• Ensure design choices are consistent and
  compatible
• Improve development effectiveness
• Enable selection of HW configuration
• Early warning when planned configurations are
  inadequate

9
Performance objectives

• Performance objective is not just a number, its
  a set of numbers
• Expressed in different ways
• Constraints on resource usage are limitations on
  amounts of various resources use by a scenario
• Must be specific and measurable

10
Execution environment

• This consists of
• HW and network configurations
• OS, middleware, other SW support
• Database usage
• Other SW that shares computer system
• Scenario performance clearly depends on execution
  environment

11
Software resource requirements

• Processing steps in a scenario are described in
  terms of SW resources used
• Depend on the execution environment
• Depend on the type of SW you are constructing and
  interfacing with

12
(No Transcript)
13
Computer resource requirements

• Used to map software resource requirements onto
  the amount of service they require from key
  devices in the HW configuration
• CPU time
• I/O
• Messages
• Processing overhead

14
(No Transcript)
15
Data gathering issues

• Do not need precise information early in the life
  cycle
• Start with approximations
• Document performance specifications as you
  collect them
• Define standards for including them in a CASE tool

16
SE 767-NT Software Performance Engineering Robert
Oshana End of lecture For
more information, please contact NTU Tape
Orders NTU Media Services [phone]
oshana_at_airmail.net
tapeorders_at_ntu.edu