CSS 422 - Software Architecture Flashcards
Software Architecture is
description of the subsystems or components of a software system and the relationships between them
A more formal description of Software Architecture is
Architecture is the fundamental organization of a system embodied in its components, their relationships to each other, and to the environment, and the principles guiding its design and evolution
What four common, recurring themes make up software architecture?
System
Structure
Environment
Stakeholder
System
A collection of componenets organized in specific ways to achieve specific functionality
Environment
A context or circumstance in which a software system is built and has a direct influence on its architecture
Structure
A set of elements that are grouped or organized together according to a guiding rule or principle
Stakeholder
A person or group of persons who has an interest or concern in a system and its success
An Architecture of a system is best represented as
structural details of the system.
Deployment architecture is strongly connect to the quality attributes of
scalabililty, performance, security, and interoperability
An architecture
Defines a structure
Picks a core set of elements
Caputures early design decisions
Manages stakeholder requiremenets
Influences the organizational structure
Is influenced by its environment
Documents the system
Conforms to a pattern
Architectural patterns
certain patterns and sets of styles for system architecture that have had success in practice
Examples of architectural patterns
client-server
pipes
filters
data-based architectures
Why create a formal Architecture
Every system already has an architecture, whether formal or not, so you can’t have a system without also having an architecture
documenting the architecture allows it to be shared
Makes changes and iterative development possible
Makes the system extensible and modifiable
Technical architect
Technical architect is concerned with the core technology used in an organization
Security architect
a Technical architect that tunes the security strategy used in applications to fit the organizations information security goals
Information architect
A technical architect that comes up with the architecture to make information available to/from applications in a way tha tfacilitates the organizations business goals
System architect
is worried about how the core system architecture maps to the software and hardware architecture and the various details of human interatctions with the components in the system
Enterprise architect
is concerned with how the different elements in an organization and their interplay is tuend twoards achieveing the goals of the organization in an efficient manner.
Software design
A blueprint of the details at the implementation level of the various subsystems and components that make up those subsystems
Modifiability
a software quality attribute
quality attribute
A measurable and testable property of a system which can be used to evaluate the performance of a system within its prescribed environment with respect to its nonfunctional aspects
nonfunctional requirements
global constraints that describe how a system should operate.
ex: things like speed, reliability, data integrity, etc.
The opposite of functional requirements like.
like: this needs to be able to access the database and pull back information on billing accounts
quality attributes
modifiability
testability
scalability
availability
security
deployability
modifiability
the ease with which changes can be made to a system and the flexibility with which the system adapts to such changes
testability
how much a software system is amenable to demonstrating its faults through testing
scalability
a system’s capacity to accommodate increasing workload on demand
performance
the amount of work accompmliished by a system using a given unit of computing resource. higher the work/unit ratio, the higher the performance
availability
the property of readiness of a software system to carry out its operations when the need arises
security
the degree of ability of a system to avoid damage to its data and logic from unauthenticated access
deployability
the degree of ease with which software can be taken from the development to the production environment
coupling
the degree of interdependence between software modules
high coupling means that the modules are closely connected and changes in one module may affect other modules
it can also be defined as the number of dependencies that modules share between them
cohesion
the degree to which elements within a module work teogether to fulfill a single, well-defined process
the degree to which the elements belong together
the aspects of modifiability
readability
modularity
reusability
maintainability
readability
the ease with which a program’s logic can be followed and understood
modularity
the software is written in well-encapsulated modules which do very specific, well documented functions
reusability
the number of parts of a system (including tools, designs, code, and others) that can be reused in other parts of the system
maintainability
the ease and efficiency with which the system can be updated and kept working in a useful state by its stakeholders
aspects of readability
well written - uses simple syntax and well known features. logic is clear and concise. module and variable names are meaningful
well documented - good comments (inline). also, represenets external documents for configuration or api usage
well formatted - follows style guidelines
“Pythonic” =
keeping up with th eZen of Python
antipatterns
coding practices that contribute to unreadable code
list the antipatterns
code with little or no comments
code that breaks best practices for the language
programming antipatterns:
– spaghetti code
– big ball of mud
– copy-paste programming
– ego programming
python antipatterns
– mixed indentation
– mixed string literal types
–overuse of functional constructs (map(), reduce(), filter(), lambda(), etc)
Ways to document code
Inline documentation
* comments
* function documentation
* module documentation
Extermal documentation
Readme, INSTALL, CHANGELOG
User manuals
* formal documents created bya adedicated person or team
Ways to document in Python
Inline comments
Function docstring (with notes about the parameters): @params, @returns
Class docstrong
Module docstring
PEP 8
Python Enhancement Proposal
A coding and style guide for Python
cohesion
how much the code within a module is aligned; i.e., a measure of how well the code is about the same thing
coupling
a measure of how much a module’s code relies on code from another module
Interface
the modules functions classes or methods that provide a connection / interface to the module
Can be thought of as the API of the module
python module variables that are not part of the module interface
If the variables are not meant to be accessed by an external program, they should be preceded by a double underscore: ___
abstraction
Abstraction is used to hide the internal functionality of the function from the users
User is familiar with that “what function does” but they don’t know “how it does.”
so the
class Square():
can let you instantiate a square = Square()
object without having to understand how to draw it, or without having to tell the computer how to draw it for each step. The square class does it for you.
The square class has abstracted away the details
abstract common services
so if module B uses services from module A, then those services are abstracted services from module A
Because Module B can’t see how those modules are built – it only knows how to use those services
Using inheritance techniques
another strategy for modifiability
If there is code that is similar, it might be easier to user inheritance to pass along the common functions instead of keeping similar functions hard-coded into each successive class
super() method in Python
Using Late Binding Techniques
another strategy for modifiability
plugin mechanisms
broker/registry lookup services
notification services
deployment time binding
using creational patterns
Unit testing
the most fundamental type of testing performed by developers.
A unit test applies the most asic unit of software code - typically, functions or class methods - by using executable assertions which check the output of ht eunit being tested against an expected outcome
python unit testing software
py.test - a full-featured, mature testing framework
nose2 - a third party unit-testing module for python
unittest - a unit testing module in the standard python library
unittest module provides these object
Test cases
test fixtures
test suites
test runners
test results
pytest
will automatically find tests with any function prefixed with the word “test_”
nose 2
a third-party unit-testing module available for python
code coverage
the degree to which the source code under test is covered by a specific test suite.
ideally test suites should aim for higher code coverage, as this would expose a larger percentage of the source code to tests and help to uncover bugs
Lines of Code (LOC)
a percentage of the subroutines (functions) covered by a test suit
how are code coverage metrics reported?
by lines of code (LOC)
doctests
Inline tests in a function , class or module documentation which add a lot of value by combining code and tests in one place without having to develop or maintain separate test suites
silent testing
you can’t see the details of the test, you only learn whether all of the tests passed - no output is produced
white box testing
testing that involves knowing the exact way something is coded.
usually done by the develpers. i.e., they know exactly how something is coded - they have access to the source code, so they know what the code should do, and can find out immediately why it has broken
Examples of Code Documentation
Inline documentation
External documentation
user Manuals
Scale vertically or scale up
When the system scales by either adding or making better use of resources inside a computer node (such as CUP or RAM)
Scale horizontally or scale out
when a system scales by adding more compute nodes to it
Scalability
The degree to which a software system is able to scale when compute resources are added
Concurrency
The amount of work that ets done simultaneously in a system
Performance of a system
a function of its software and of its hardware capabilities
performance
the degree to which a system is able to meet its throughput and/or latency requirements in terms of the number of transactions per second is known
Software Performance Engineering
develops performance models early in the SDLC and uses results from those models to modify the software design and architecture to meet performance requirements in multiple iterations
Performance Engineering Lifecycle
performance is a non-functional requirement
the PEL parallels the steps in the SDLC with every step the performance is improved
Stress testing tools
these tools are used to supply workload to the system under test - simulating peak worloads in and burst loads (of very high traffic)
Load generators
Stress testing tools
Stress testing tools examples
httpperf
ApacheBench
Loadrunner
apache Jemeter
Locust
Monitoring tools
these tools work with the application code to generate performance metrics such as the time and memory taken for functions to execute, number of functioncalls made per request-response loop and the average and peak times spent on each function
Instrumentation tools
trace metrics and track events such as exceptions
track details as the line number where the exception occurred and the timestap, and environment fo the appplication
code or application profiling tools
these tools generate statistics about functions, frequency of duration of calls, time spent on each function call
this allows the developer to find critical sections of code where the most time is spent and allowing them to optimize those sectionsP
Python standard library instrumentation and profiling modules
profile
cProfile
Performance complexity of code
How a routine or function responds to input size typically in terms of the time spent in executing the code
Big O Notation AKA:
Backhmann-landau notation
Asymptomatic notation
the letter O
the rate of growth of a function with respect to input size – also called th eorder of the function
real time
the actual wall-clock time that elapsed for an operation
user
the amount of actual CPU time spent within the process in user mode
Sys
the amount of CPU time spent executing system calls within the kernel for the program
Total CPU time =
user + sys time
amortized analysis of algorithms
takes into account both the lower and upper end of the time taken for executing algorithms and gives the programmer a realistic estimate of the average time spent
