Diverse Environments

1
Diverse Environments

• 60s Mainframe
• One large processor, many users
• Minimal data redundancy, single point of security
  and control
• 70s Cluster of identical centrally located
  processors
• Invisible load balancing
• 80s PCs
• Need to link PCs together
• LAN, WAN, Internet
• distributed computing

2
Distributed IS

• Network of subsystems located at, and adapted to,
  specific areas of need
• Some subsystems need to interact with other
  subsystems
• Some subsystems need to share files or data
  processing facilities
• Some subsystems are relatively self contained
• The drive to distribute
• Cost of desktop computing is decreasing (real and
  nominal)
• Increased communication bandwidth and cheap
  storage
• Local demand for special apps is increasing
• WWW provides natural connectivity

3
Pros and Cons of Distributed Computing

• Increased adaptability and responsiveness to
  local needs
• Increased availability of resources due to
  redundant systems
• Closer alignment with organizational structure
• Improved incremental growth and local scalability
• Increased end user empowerment
• Location independence support development of
  virtual teams
• Independence from single vendor solutions

• Increased difficulties in controlling information
  resources
• Potential for massive data redundancy
• Increased complexity and cost
• Increased network management resources
• Increased difficulty in testing and detecting
  failures

4
Move the Data or the Applications?

• Distributed Applications
• File-Server (an extra hard drive)
• Distributed Data
• Client-Server (an extra processor)

Server
Work station
Request
Entire File
Server
Work station
Request
Data Requested
5
Local Connectivity Diagram
BuyingAgents
Singapore Warehouse
Specific Location
Omaha Warehouse
Inventory Control
London Warehouse
Mobile Location
Material Suppliers
External Location
Payables Management
Connections dont have directions
6
Connection Topologies

• More cable
• Server dies, network dies
• Server determines throughput

• Low complexity
• Node failure OK

• Star
• Ring (Token)
• Bus

• Dies if single cut
• Relatively high install cost

• Easy to synchronize connections
• Can span wider differences
• Generally high band width

• All or part of network can be disabled if theres
  a cut

• Relatively cheap to install

7
Managing Distributed Data
8
Designing for Collaboration

• The system should support individual work as well
  as collaboration.A successful collaborative
  system must
• allow end users convenient access to each others
  work
• The system must afford mutual intelligibility.
• The ability of one user to understand the work
  stored by another.
• System must support simultaneous, fine
  granularity access.
• The finer the granularity of the locked portion
  of the database, the greater the potential
  productivity.
• Collaborative wearable computers.

9
Designing for ERPs

• An ERP is designed to solve the problem of
  information islands by providing data storage and
  processing mechanisms that serves the needs of
  all departments and functions within the
  organization.
• Implementation Considerations
• In order to shorten ROI periods use BPR in
  parallel
• Identify the implementation Scenarios
• Comprehensive Implementation Scenario (processes)
• Compact Implementation Scenario (technology)
• Off-the-shelf vs. Best-In-Class ERP
• ERP vendors provide best-in-class applications
  for specific business requirements and integrate
  them.
• There is no such thing as a minor ERP project

10
Designing for Intranets

• An intranet centralizes the business process in
  an easily accessible, platform-independent
  virtual space.
• Focusing on Processes rather than Departments
• Common characteristics shared by successful
  intranets
• They focus on tasks, rather than documents for
  simple data capture, and aim to integrate those
  tasks into distinct processes.
• They encourage collaboration by creating shared
  and familiar spaces
• Intranet Design Myths
• Intranets are cheap.
• Build it and they will come.
• Intranets are for REALLY BIG organizations.
• Intranets require an Internet connection and are
  not secure.
• Intranets are low maintenance applications.
• Intranets are an IS thing.

11
Data Warehouse
85 queries
15 queries
Users
Enter/Input
12
Designing for the Data Warehouse

• The data warehouse is a collection of integrated,
  subject-oriented databases designed to support
  the DSS function, where each unit of data is
  nonvolatile and relevant to some moment in time.
  (Inmon, 1992)
• Two implicit assumptions
• The DW is physically separated from all other
  operational systems
• DWs hold aggregated data and transactional data
  for management separate from those used for
  online transaction processing.
• Online Analytical Processing (OLAP)
• Typology
• Virtual DW, Central DW, Distributed DW

13
The Sins of Data Warehouse Design

• If you build it, they will come.
• Omission of an architectural framework.
• Underestimating the importance of documenting all
  assumptions and potential conflicts.
• Abuse of methodology and tools.
• Abuse of the data warehouse life cycle - it never
  ends!
• Ignoring concerns about the resolution of data
  conflicts.
• Failure to demonstrate the mistakes made during
  the first DW project.