Evolution of Information Technology

Infrastructure and ArchitectureBA 572 - Week 1 – Part 1

Sources: HBR 397 – 118, “Intranets and Middleware” Dr. James Coakley (Oregon State University) MIS Textbook by Turban, Rainer & Potter (Chapter 5) Mr. Sakthi Angappamudali (The Standard) Mr. Lee Martin (Hitachi Consulting) Dr. V.T. Raja (Oregon State University)

BA572 Week 1 (Part 1) Outline

IT Infrastructure vs. IT Architecture Evolution of IT Infrastructure and Architecture

Major eras of the computer industry

Terminology/Acronyms Centralized/Decentralized/Distributed Computing

TPS, MIS, DSS, ES, Middleware, OOP, DW, OLAP, Data Mining etc.

Comment on Performance Metrics

Definitions

Information Technology (IT) Infrastructure: physical facilities, services and management that support computing resources Information Technology

HardwareSoftwareDatabaseTelecommunications & Networks

IT personnel

Definitions

Information Systems (IS) Architecture: the “plan” that aligns IT infrastructure with business needs Help people effectively fulfill their information

needs Note that the term “Information Architecture” is

also being used to describe process of designing web sites

Adapted from "Intranets and Middleware", HBR 397-118.

S1

19601980

1990

Ab

ilit

y t

o f

ill in

form

ati

on

need

s

2000

PC/LAN

Client/Server db

db

db

Distributed

db

dbdb dbdb

Web Services

Mainframe

Evolution of Information Technology Infrastructure

Data Processing Era

IT Infrastructure (host-centric processing) Hardware: Mainframe with text-based terminals Software: Independent functional applications

Served one purpose Data Storage: independent “files” for each

functional application Telecommunications: Limited support of

distributed operations IT Personnel: technically oriented

Mainframe

IS Architecture:Transaction Processing System (TPS)

Emerged in the early days of IS Collect, store, and process transactions

Source documents are basis for input Perform routine, repetitive tasks Found in all functions of an organization If they fail, the whole organization may suffer Automate “highly structured” decision processes

Payroll

Mainframe

IS Architecture: Management Information System (MIS)

Convert/use TPS data to support monitoring Alert managers to problems or opportunities Provide periodic and routine reports

e.g., summary reports, exception reports, comparison reports

Provide structured information to support decision makingResulted in “Information overload”

Mainframe

IS Architecture: Centralized Corporate Structure

Executive

Operational

Managerial

InboundLogistics

Purchasing

RawMaterials

Production FinishedGoods

OutboundLogistics

Sales

Functional Transaction Processing System

Management Information System

Mainframe

Micro-Computing Era

IT Infrastructure (PC environment) Hardware: PCs (low cost compared to

mainframe) Software: Individual PC applications Data storage: Individual files linked to apps Telecommunications: low-speed LANs IT Personnel: technically oriented & mainframe

biased

PC/LAN

IS Architecture:Decision Support Systems

Proliferation of desktop applications Why?

TPS/MIS were not providing information needed to support decisions

“End-user” developmentUndocumented spreadsheet models

Proliferation of localized data storage

PC/LAN

db

db

db

db

IS Architecture

Executive

Operational

Managerial

InboundLogistics

Purchasing

RawMaterials

Production FinishedGoods

OutboundLogistics

Sales

Functional Transaction Processing System

Management Information System

Desktop DecisionSupport System

PC/LAN

Client/Server Era IT Infrastructure (distributed computing

environment) Hardware: PCs and Specialized Servers Software: Facilitating Data storage: Distributed Relational database and

centralized warehouse Telecommunications: high-speed LANs Network: Client/Server IT Personnel: technically skilled, business oriented

Information Systems architecture? Share applications and data within and across functional

areas

Client/Server

db

Facilitating Software Systems

Office automation IT for “office” employees

Document tracking, communication, scheduling, etc.

Client/Server

db

Facilitating Software Systems (cont’d)

Decision Support SystemsProvide information to support “semi-structured”

decision makingEffectiveness focus

Expert SystemsKnowledge-base integrated with DSSMost are “rule-based” systems that process facts, not

numbersCredit evaluationCisco/DELL tech support

Client/Server

db

Database Approaches

Centralized All data in one location

Promotes maintenance and securitySubject to single point of failure

Client/Server

db

Database Approaches

Distributed data management Get data closer to applications Replicated

Complete copies in multiple locationsSignificant overhead

PartitionedEach location has portion of database

Data management becomes an issue Complex Concurrency Control

db

db

db

Distributed

db

db

Transactions used to interact with a relational “client-server” database For each transaction, OLTP typically deals with

a small number of rows from the tablesThe transactions are typically highly

structured, repetitive and have predetermined outcomes

E.g., orders, changing customer address, etc.

Online Transaction Processing

db

db

db

Distributed

db

db

Client/Server Systems

Executive

Operational

Managerial

InboundLogistics

Purchasing

RawMaterials

Production FinishedGoods

OutboundLogistics

Sales

Functional Transaction Processing System

Client/Server System

db db db db db

db

Network Era (Distributed Computing)

IT Infrastructure (distributed computing environment) Hardware: PCs and high-end Servers Software: Enabling, enterprise-wide Data storage: Distributed Relational Database Telecommunications: high-speed WAN Network: Middleware IT Personnel: still technical, but business

awareness

db dbdbdb

Distributed Computing

Middleware

Introduction of Middleware

Software that makes it possible for systems on different platforms to communicate with each other. Allows applications to talk to each other

Consistent Application Program Interface (API)Code application to talk to middleware, not

underlying resourcesUpgrade/modify underlying resources without

needing to modify applications

db dbdbdb

Distributed Computing

Middleware

Object Request Broker (ORB)

ORB involves synchronous communication and location/platform transparency.

ORB uses object-oriented programming methods.

db dbdbdb

Distributed Computing

Middleware

ORB (cont’d)

ORB architecture:

ORB

ClientRemote Service

locate service

activate service

establish connection

communicate

db dbdbdb

Distributed Computing

Middleware

File Sharing

Napster:

ORB

RequestStored Files

locate service

activate service

establish connection

communicate

db dbdbdb

Distributed Computing

Middleware

Peer-to-Peer File Sharing

Kazaa:

Request

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

db dbdbdb

Distributed Computing

Middleware

Advantages of ORB Middleware

Anonymous interaction among applications Integrate new client/server applications with

existing legacy, mission-critical applicationsEasier development environment

Reduce cost Improve time-to-market of applications

Enables distributed data environmentEnables dynamic web applications

db dbdbdb

Distributed Computing

Middleware

Disadvantages of ORB Middleware

Switching costs are high Upgrade from previous “Middleware” solutions

Requires high technical expertise Tend to outsource Lengthy deployment time

db dbdbdb

Distributed Computing

Middleware

Unresolved Issues with ORB

SecurityScalability

Related to network capacityRapidly changing technologies

db dbdbdb

Distributed Computing

Middleware

DBMS Applications

With advent of high-speed, distributed architectures expanded our use of database beyond capturing and storing transaction data Knowledge Discovery

Process of extracting useful knowledge from volumes of data

Supported by:Massive data collection (Data Warehouse/Data Marts)Multiprocessor computingOn-line Analytical Processing (OLAP)/Data mining

db dbdbdb

Distributed Computing

Middleware

Data Warehouse

Collection of data in support of decision making process that is: Subject-oriented: organized by entity, not application Integrated: stored in one place, even though it originated

from a variety of sourcesCrosses functional boundaries of an organization

Time-variant: represents a snapshot at one point in time Nonvolatile: data is read-only Typically very large

db dbdbdb

Distributed Computing

Middleware

Data Warehouse

Large repository of detailed and summary data used to support the strategic decision making process for the enterprise

Stores current and historical data (internal and external) Integrates data from organization’s disparate information

systems used by functional units Involve hundreds of gigabytes, and terabytes of data Run on very powerful computers Expensive

•Design•Mapping

•Design•Mapping

OLTP Systems

FunctionalIS

External Data

• OLTP - Raw DetailNo/Minimal History

•DW-Integrated•Scrubbed

•History•Summaries

• Targeted• Specialized (OLAP)

OLTP, DW and DM - Data Characteristics

•Extract•Scrub•Transform

•Extract•Scrub•Transform

CentralRepository

•Load•Index•Aggregation

•Load•Index•Aggregation

DataWarehouse

Data Mart

•Replication•Data Set Distribution

•Replication•Data Set Distribution

End UserWorkstations

Data Warehousing Process

Multidimensional Database (cont’d)

Data marts Scaled-down version of a data warehouse that

focuses on a specific areae.g., a department, a business process

db dbdbdb

Distributed Computing

Middleware

Individual Architected Data Marts

Common Logical Subject Area ERD

Common Business Dimensions

Common Business Rules

Common Business Metrics

Glossary

SalesDistribution

Product

Marketing Customer Accounts

Finance Operations and Inventory

Vendors

An Incremental Approach

ArchitectedArchitectedEnterpriseEnterpriseFoundationFoundation

SalesDistribution

Product

Marketing Customer Accounts

Finance Operationsand Inventory

Vendors

Enterprise Data Warehouse

The Eventual Result

Multidimensional Database

OLTP not good when doing analysis of data – poor performance

OLAP – on-line analytical processing

db dbdbdb

Distributed Computing

Middleware

On-line Transaction Processing (OLTP) and On-line Analytical Processing (OLAP)

OLTP: Immediate processing/analysis and handling of multiple concurrent transactions from customers/users Example:

OLAP: Capability for manipulating and analyzing large volumes of data from multiple perspectives (multidimensional analysis) Example:

“Slice and Dice” an OLAP Cube

Advantages of OLAP

All hierarchical or aggregated values can be pre-calculated in the cube rather than accessing the Warehouse Major reduction in query time

Each cube makes “business sense” Not normalized data structures

db dbdbdb

Distributed Computing

Middleware

Massive Data Analysis

Data mining Provides a means to extract patterns and

relationshipsExample: Analyze sales data to identify products that

may be attractive to a customerAmazon.com buyer suggestions

Two capabilitiesAutomated prediction of trends and behaviorsAutomated discovery of previously unknown patterns

db dbdbdb

Distributed Computing

Middleware

Data Mining

Some Benefits: Market Segmentation Fraud Detection Market Basket Analysis Trend Analysis

Business Intelligence

BI/Analytics software (suite): Used to collect, store, analyze and present

sufficient and accurate information in a timely manner and in a usable form

Includes OLAP, data mining, statistical analysis Has a positive impact on business strategy, and

operations Addresses analysis paralysis caused due to

information overload?

Business Intelligence

Enterprise BI Suites and Platforms

The Decision Making Roadmap

TransactionSystems

DecisionSupport Systems

ExecutiveInformationSystems?

Business Planning

RUNRUN MANAGEMANAGE GROWGROW

UsersUsers Knowledge BrokersKnowledge Brokers ManagementManagement

• Operational• Functional• Current• Detailed

• Multi-Dimensional

• History• Summary

• Analyze What If Scenarios

• History• Detailed

DataData InformationInformation

KnowledgeKnowledge

VisionVisionActionsActions

Network Enabling Software

Supplier Customer

Enterprise Wide Systems

Enterprise Wide Systems

Enterprise Wide Systems

Supply ChainManagement

Customer Relationship Management

db dbdbdb

Distributed Computing

Middleware

IT Infrastructure (Web-enabled) Hardware: Low-end PC with Browser, high-end

Servers Software: Web extensions Database: Distributed Relational Network: Use IP-based standards Telecommunications: broadband IT Personnel: Business analysts, technical

specialties

Internet Era

Business use of the Internet:Electronic Commerce

E-business: Subset of e-commerce Transactions between

business partners

Individual EnterpriseSupplier/ CustomerInternet

Intranet

Extranet

B2C: InternetB2B: ExtranetB2E: Intranet

Web-based Solutions Early attempts to incorporate WWW into inter-

organizational systems Static, state-less web pages

Complicated navigation Not “connected” to underlying data

Page not dynamically updated when data changes

Dynamic and interactive web applications connected to enterprise database(s)

Web 2.0 http://www.oreillynet.com/pub/a/oreilly/tim/news/2005/0

9/30/what-is-web-20.html http://en.wikipedia.org/wiki/Web_2.0

Web Services

Standards are evolvingSecurity?Do web services 'solve' interoperability

between applications? Need ERP?

db dbdb

Web Services

Comment on Performance Metrics

How does IT add value and how much value? TCO/ROI Tangible vs. Intangible Impacts

What is(are) purpose(s) of IT applications? Automate Facilitate/Informate Enable business strategy/significant competitive

advantage Alignment of IT and Business Strategy