Software Testing and Quality Assurance

Document Sample
Software Testing and Quality Assurance Powered By Docstoc
					Software Testing and Quality
Assurance

   Lecture 35 – Software Quality
            Assurance



                                   1
Lecture Objectives
   Reliability Block Diagrams
       Series – Parallel Systems




                                    2
Reliability Block Diagrams
   We need some way of predicting and
    evaluating the reliability of our designs.
   For system there are well established
    modeling methods:
       Reliability Block Diagrams; and
       Markov Models
        These models are stochastic in nature.


                                                 3
Reliability Block Diagrams
   Stochastic process is one in which
       There is a sequence of events in time
        where each event is part of a probability
        distribution.




                                                    4
Series – Parallel Systems
   Reliability block diagrams seek to
    decompose systems into parallel and
    serial blocks
       Where each block interacts with other
        block to effect the reliability of the system
        as a whole.




                                                    5
Series – Parallel Systems -
Example




                              6
 Series – Parallel Systems -
 Example
    A redundant pair of writers put values
     from the source onto the bus, and
    A reader that reads the values from the
     bus.

The system fails if no message is received by the reader
               in the given time interval.



                                                      7
       Series – Parallel Systems -
       Example
   A message starting
    at the source then it
    is passed to both
    writers and both
    writers pass the
    message to both
    channels.



                                     8
Series – Parallel Systems -
Example
   In the model,
       Each writer acts in serial with the bus and
       Both channels of bus act in serial with the
        reader.
   To analyze the model;


    We must calculate the reliabilities for the serial
         And parallel blocks in the model.
                                                         9
Series – Parallel Systems -
Example




                              10
Reliability and Failure Logic for
Serial Blocks
   For the system to model without failure at a
    given time T,
       Each of the components B1, B2 and B3 must
        function independently with out failure for the time
        period T.
   For system to function in the Time T, either
       Block 1 must not fail in time T;
       Block 2 must not fail in time T;
       Block 3 must not fail in time T.

                                                         11
Reliability and Failure Logic for
Serial Blocks
   For system to fail in the Time T,
       Only one of the components need to fail.
   Thus, the probability of failure is




                                               12
Reliability and Failure Logic for
Serial Blocks
   The law of additive law of probability
    gives us;




       Where F(Bi) is the probability that the first
        failure in Bi will occur before Time T.


                                                   13
Reliability Logic for Parallel
Blocks
   All blocks must fail before the system
    fails and
       Only one block needs to function for the
        system to function.
   Parallel blocks, work almost inversely to
    serial blocks.



                                                   14
Reliability Logic for Parallel
Blocks
   For a parallel block to fail in a time T,
       We need all of the blocks to fail within the
        time T and so




                                                   15
Reliability Logic for Parallel
Blocks
   For a parallel blocks to be reliable for a
    time T,
       We have must experience no failure in time
        T.




                                               16
Reliability Block Diagrams from
System Architectures -
Guidelines
   Guideline 1
       Determine what constitutes a system
        failure. This in turn determines which
        component failures causes a system to fail.
   Guideline 2
       Determine what components need to fail in
        order to cause a system failure.
       How messages, signals or data flows
                                                17
Reliability Block Diagrams from
System Architectures -
Guidelines
   Guidelines 3
       Try to ensure that each block in the
        reliability block model captures one
        function of the system.
   Guideline 4
       Try to ensure that you capture the
        parallel/serial connections from the system
        accurately .

                                                18
  Reliability Block Diagrams from
  System Architectures -
  Guidelines
     Guidelines 5
         There may be more than one mode for the
          system.
         You need to create a reliability block
          diagrams for each mode of the system.


        Windows can operate in a Safe Mode,
More reliable because it is limited to basic functionality.

                                                              19
Key points
   Reliability Block Diagrams; and Markov
    Models
   Reliability block diagrams seek to
    decompose systems into parallel and
    serial blocks
   Reliability and Failure Logic


                                        20

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:1
posted:4/17/2013
language:Unknown
pages:20