Software Testing and Quality Assurance

1
Software Testing and Quality Assurance

• Lecture 40 Software Quality Assurance

2
Lecture Objectives

• Software Performance Engineering
• Software Performance
• Performance in Requirements Engineering

3
Software Performance Engineering

• Confusion with efficiency.
• Efficiency the degree to which a program
  performs its functions within the minimum
  consumption of resources.
• Efficiency implies trying to get the best from
  the resources available.

4
Software Performance Engineering

• Performance the degree to which a program or
  system accomplishes its functions within given
  constraints such
• Speed, accuracy or memory usage.
• We are concerned with specifying, and engineering
  systems to meet their speed, accuracy etc. usage
  constraints.

Making programs more efficient may help to
achieve the performance requirements.
5
Performance Myths

• Hardware is fast and inexpensive there are many
  factors influencing performance.
• Simply doubling the speed of processor does not
  take into account bottlenecks in design.
• You can tune software later, if necessary
  Empirical studies suggests that performance
  requires
• Good design and a good choice of algorithms,
• As well as fine tuning of code.

6
Software Performance

• Performance is effected on several levels.
• System Architecture where performance is
  affected by the decomposition of the system into
  modules
• Detailed Design affected by the selection of
  algorithms and data structures.
• System Software affected by the operating
  system, load balancing etc.
• Hardware affected by the physical speed of the
  processor and I/O devices.

7
Performance in Requirements Engineering

• Performance goals depends on the type of
  resources that we need to contain.
• If resource is time, then the goals are stated in
  terms of constraints on time.
• Workload Specification develop a workload
  profile of the functions that are most likely to
  occur in practice,
• Their frequency or request rate.

8
Performance in Requirements Engineering

• Performance Scenarios Identify major components
  from architectural design or use cases
• Order in which components will need to be
  executed to carry out system functions.
• Execution Environment determine the platforms,
  such as CPU, memory etc on which the system
  execute
• Anticipated platform loads associated.

9
Performance in Requirements Engineering

• Resource Usage resources required by the system
  and their usage by the system.
• RAM, disk, network, etc.

10
Performance Goals

• Try to specifying performance goals which are
  quantitative and measurable.
• For example, time
• Response time Time required to respond to an
  input, event etc. For example,
• database may require lt 1 sec for a specified
  transaction type and lt10 sec for a specific kind
  of update.

11
Performance Goals

• Throughput number of operations or processing
  steps per unit time. For example,
• The system must output gt 25 frames/second,
• Complete gt 100 updates/ hour.
• Constraints system performance is specified in
  terms of constraints on time or resources. For
  example,
• New system must be gt 10 faster than previous
  version.
• System must perform lt 2 disk access/ update.

12
Key points

• Difference between Efficiency and
  Performance
• Performance Myths
• Performance is effected on several levels.