application patterns

Application Patterns tomi vanek

February 2014

Copyright © 2014 Accenture All rights reserved. 2

• User Interaction

• Services

• Persistence – scaling, locking, cache

• Real-time processing

• High Volume Batch Processing

Agenda

User Interaction


MVP

MVC / MVP / MV-VM

Model

View Controller Action

Model

View Presenter Action

Model

View

View-Model Action

MVC

MV-VM


Model-View-Controller (MVC)


MVC Variant: Page Flow


Page flow with asynchronous

commit of order

Simple synchronous MVC

implementation

MVC: Page Flow Example


In-Memory Server Session State


Persistent Session State


Serialized Session State in View


Client Session State in Local Storage

Services


Method invocation / Resource manipulation

• Services: – Strongly typed interface

contract.

– Broad range of aspects are standardized.

– Implemented in integration infrastructure and technologies.

– Service discoverability and lifetime management.

• But: – Limits in scalability.

– Expensive extensibility.

• Resources:

– Simple data exploration.

– Simple extensibility.

– Simple data consumer

code.

• But:

– Limited schema

definition.

– No standards for

complex functionalities /

scenarios.


Business Layer Remote Procedure Call

Presentation

Presentation Layer

Business Tier Data Tier

External

Business Module

Resources Data Layer Business Layer

External

System

Repository

Integration Gateway

Service

Service

Relational Database


Compact Business Layer with Active Records

Application Tier

Dynamic

Presentation Module

Presentation Layer Resources Domain Layer

Presentation

Data Tier

External

External

System

Integration Gateway

Relational Database


REST / WS Business Layer

Presentati

on


External

Business Module

Resources Data Layer Business Layer Presentation Layer

External

System

Repository

Integration Gateway

Service

Service

Relational Database C

on

troller


Asynchronous Business Layer

Presentation


External

Business Module


External

System

Repository Relational Database

Integration Gateway

Queue

Queue

Service

Queue

Queue

Service


Push Notification Business Layer

Presentation

Business Tier

Business Module


Subject Register

Service


Data synchronization

Persistence - Scaling


Database Scaling: Master – Slave

Internet

Load Balancer

Load Balancer

Slaves

Web Servers

Master

Write Read

Replicate


Database Scaling: Active - Passive

Internet

Load Balancer

Passive

Web Servers

Failed

Active Instance

Passive to Active

Instance

Replicate

Read / Write


Database Scaling: Cluster

Internet

Load Balancer

Virtual Database

Cluster

Controller

Virtual Database

Cluster

Controller


Database Scaling

Load Balancer

Shard Resolver

Web Servers

Product 1 .. n

Product n+1 .. o

Product o+1 .. p

Locking


Optimistic Locking


Pessimistic Locking


Read-through / Write-through Cache


Pre-fetch


Publisher

Real-time processing Integration Patterns


Reference model

34

Messaging Idioms

• Channels

• Message Construction

• Rooting

• Transformation

Messaging Design Patterns

• Pipes and Filters

• Claim Check

• Scatter-Gather

Integration Application Architecture Design Patterns


Pipes and filters


Pipeline


Claim-check


Claim-check for secure communication


Scatter-Gather with Publish-Subscribe Input Channel


Scatter-Gather with Recipient List

Scale-Out Batch High Volume Batch Processing


Batch Application Elements

Job

Launcher Job Step

Reader

Processor

Writer

Job Repository


Batch reference model


Scaling


Job with Parallel Steps


Step with parallel threads


Partitioning


Remote chunking

Discussion

application patterns

Software