Implementation Architecture Lecture 20- 22 1. Implementation View (1) “ how the system is built...

1

Implementation Architecture

Lecture 20- 22

2

Implementation View (1) Focuses on “how the system is built” Which technological elements are needed to

implement the system: Software packages (req. Development tools), Libraries (implementing behavior e.g. “Cout”), Frameworks (Generic functionality e.g. “JUnit”), Classes , ...

Addresses non-runtime quality attributes: Configurability (e.g. Platform), Testability (e.g. Mapping of FRs), Reusability

3

Implementation View (2) Again, comprised of components and

connectors

Here, components and connectors reflect “Software entities and their relationships at

the level of source and binary code”

Typically a number of implementation models can be developed

Each model focuses on the concurrent subsystems/processes from the execution view.

4

A view focussing on…

As, implementation architecture focuses on the “built-time” structure of the system.

Build-time structure: Implementation modules Off-the-shelf technology Build-time configuration

Detailed activity Call sequences i.e. Sequence Diagram

5

Components in implementation architecture (1)

Two types of components:

Application components And infrastructure components

6

Components in implementation architecture (2)

Application components: “Are responsible for implementing domain-

level responsibilities”

These are responsibilities found in a detailed conceptual architecture

Application components might be realized as: Source packages, And files

7

Components in implementation architecture (3) Infrastructure components:

“Are needed to make the system run but are not related to the application functionality”

e.g. LAN Connection Handler in Distributive system is a typical infrastructure component (Just use to connect local systems, so they can

communicate when they require)

8

Components in implementation architecture (4)

Often an infrastructure component acts as a "container" for application components A container component provides an execution

environment for the contained components (application components)

Typically, the container executes within a process and creates threads for application components

e.g. a Web application server which runs multiple applications from multiple users, each of them in their own threads.

9

Component stereotypes in implementation view

Applicationcomponents

Containers

Infrastructurecomponent

10

Connectors in implementation architecture (1)

In implementation architecture connectors represent: “a "uses" relation”

The arrow depicts the direction of this relation

The nature of communication is depicted through the connector styles

11

Connectors in implementation architecture (2) API call:

A component calls a method in another component (possibly only if both components are in the same process)

Callback: The caller passes a reference of an object to the

“callee” (called component). The “callee” invokes a method (desired computation) on that object later and return the results to caller.

12

Connectors in implementation architecture (3) Network protocol:

Needed when implementation components reside in different physical layer. Components need to agree on a common protocol or use a standardized protocol

OS signals: Communication between processes running on

the same machine

13

Connectors in implementation architecture (4)

14

Connectors in implementation architecture (5) In some cases, we represent ports as:

“endpoints of connectors between components”

Ports are used to identify a particular interface

For example: A component might be quite complex but it

provides a simple interface for communication e.g. the standard Java library provides an API

(Application Programming Interface).

15

Simple “two-tier” web app

Apache

Apache module API

mod_php

Applicationcomponents

MySQL

DataModel/s

HTTP

Containers

Infrastructurecomponent

Ports

16

Conceptual v/s implementation

ConceptualArchitecture

ImplementationArchitecture

Component

Connector

Domain-level responsibilities

Implementation module

Flow of information

“Uses” relationship

17

Implementation architecture design Find application components

Find infrastructure components

Interface design

Behavior design and verification

18

Application components (1) 1-to-1 mapping: of conceptual components onto

application components is typically not possible.

Some conceptual components will become infrastructure components. e.g. persistent storage (databases) are typically

infrastructure components

Some conceptual components are spread over a number of application components Conceptual components have complex responsibilities

e.g. Application layer (Request acceptor , Request Handler e.t.c)

Application components (2)

A number of conceptual components can be mapped onto a single application component

E.g. small number and simple responsibilities

MPS system is such an example: All UI components map onto one application

component (i.e. HTML UI)

19

20

Infrastructure components

Off-the-shelf components (e.g. Third party services)

Frameworks (e.g. application framework) Servers (web, application, database, file) Generic clients (browser)

In MPS system two infrastructure components: 1. Browser, 2. Application server

21

Interface design For all application and infrastructure

components we need to define “interfaces” (ports)

Helps in clarifying the “way to connect” the components

Some interfaces are also standardized e.g. HTTP Connectivity, SQL Connectivity

UI: Combination of HTML/HTTP It is standardized!

22

Behavior design Now we need to go into details Use-case maps are not enough anymore We need to investigate behavior at the

operation level Thus, we need a sequence diagram

23

MPS: Implementation Architecture

24

Non-runtime quality attributes Since implementation view addresses build

structure It is the right place to consider non-runtime quality

attributes, e.g. Maintainability, Extensibility, Reusability, ...

We can use a mechanism similar to use-case maps Impact-maps:

“try to investigate what parts of the system need to change if "something" happens”

25

Impact-maps Map 1: new external system

interface to external system needs to be changed

Map 2: new application application component needs to be changed

Map 3: new UI UI component needs to be changed

26

Implementation arch. summary

What you need to submit at end of Implementation Architecture:

Detailed implementation architecture with app, infrastructure and interfaces

Sequence diagram of application interfaces

27

Prototype To show that the architectural solution is

feasible, we implement prototypes: For each identified application component we

provide implementation Deploy it within the infrastructure components

Test it and check: Correctness, Functionality, Quality-attributes

28

Dimensions of prototypes (1) Horizontal Prototype:

A common term for a “user interface” prototype is the horizontal prototype.

It provides a broad view of an entire system or subsystem, focusing on user interaction more than low-level system functionality, such as database access.

Horizontal prototypes are useful for: Confirmation of user interface requirements and system

scope Develop preliminary estimates of development time,

cost and effort.

29

Dimensions of prototypes (2) Vertical Prototype:

“A vertical prototype is a more complete elaboration of a single subsystem or function”

It is useful for obtaining detailed requirements for a given function, with the following benefits: Refinement database design Clarifies complex requirements by drilling down to

actual system functionality

30

Technical prototypes (1) Throwaway or Rapid Prototyping refers to:

“creation of a model that will eventually be discarded rather than becoming part of the final delivered software”

A simple working model of the system is constructed to visually show the users what their requirements may look like when they are implemented into a finished system.

Also called close-ended prototyping.

31

Technical prototypes (2)

A “throw-away” The prototype code is not part of the delivered

system No concern for irrelevant quality attributes

(performance, robustness)

Purely to gain knowledge (or confidence): Test a new version of a commercial

component Verify that a set of components work together Examine performance trade-offs Verify that a proposed architecture is sound

32

Technical prototypes (3) Evolutionary Prototyping:

“the main goal is to build a very healthy prototype in a planned manner and constantly refine it.”

The reason for this is that the Evolutionary prototype, when built, forms the heart of the new system, and the improvements and further requirements will be built.

When developing a system using Evolutionary Prototyping, the system is continually refined and rebuilt.

33

Technical prototypes (4) This technique allows the development team

to: Add features, Or make changes

Evolutionary Prototypes have an advantage that they are functional systems.

They may be used on temporary basis until the final system is delivered.