Enterprise Client/Server Technology

Massively Parallel Processing for Business

John Zamick, Ray Warren, John O'Sullivan

Publisher: Thomson, 1995, 299 pages

ISBN: 1-85032-146-9

Keywords: Networks, IT Architecture

Last modified: June 10, 2021, 10:42 a.m.

Businesses are now beginning to realize the power of intensive data analysis and its strategic importance in supporting their commercial objectives. Since solutions based on more traditional sequential computer architectures are simply not able to deliver the power required for this level of analysis, the commercial world is turning to a computer technology called Massively Parallel Processing (MPP).

MPP promises this data analysis power and much more. As the processing engine of enterprise data applications, MPP will help drive forward client/server computing into the corporate date center. Already, some far-sighted organizations such as American Express, Dow Jones and Prudential Bache have implemented MPP systems to help derive strategic value from the mountains of data that they gather.

This book sets out to demystify MPP and convince readers of its relevance for data-intensive business applications, including Executive Information Systems for decision support, reengineered business processes, on-line transaction processing, and more. It is intended for managers in business, marketing and information technology who are involved in making strategic decisions about appropriate technology platforms. It will also be an important guide for systems and database administrators requiring some technical background in MPP to aid them in supporting the very large and demanding database applications that will be critical to business viability and success in the late 1990s.

  1. The business forces driving fundamental changes in large-scale computing technologies
    • Transformation to the knowledge society
    • The history of large computers in the commercial environment
    • The inforamtion explosion
    • New business drivers — 'informating' vs automating
    • A shifting emphasis — concentration on effectiveness rather than efficiency
    • New-enterprise computer users — the rise of marketing, purchasing, engineering etc
    • Business process re-engineering — IT supporting leaner, meaner enterprises
    • Globalization and market fragmentation
    • Innovation and time to market as the new keys to success
    • Knowledge as the basis for competitive advantage
    • The need for the wildest possible availability of shared knowledge
    • The Impact on computer architectures and databases
  2. Commercial MPP usage in a client/server world
    • MPP and RDBMS: a most suitable marriage
    • 'Parallel query' is the 'killer' application for MPP
  3. User profiles and case studies
    • Everybody processes information
    • Generic profile of the early adopter of MPP
    • The business sectors leading the adoption of MPP
    • Illustrative case studies:
      • Insurance: John Alden Life Insurance Company
      • Retail leisure: Bass Taverns
      • Retail: British Shoe Corporation
      • Manufacturing: National Semiconductor Corp
  4. The position on parallelism of the RDBMS ISVs
    • The RDBMS marketplace in the future
    • Some suggested purchasing heuristics
    • Oracle (Oracle Parallel Server)
    • IBM (DB/2/6000)
    • Informix (On-Line Dynamic Server)
    • Sybase (System 10)
    • Tandem (Non-Stop SQL)
  5. Supply-side technological forces driving MPP adoption
    • The perennial need for more and cheaper computer power
    • In the data centre — only a little better
    • The hidden discontinuation in the power requirements of large scale systems
    • Business requirements for 100% availability of computer systems
    • Other factors mitigating against solutions based on traditional technology for large-scale systems
    • MPP — overwhelming endorsement by hardware vendors
  6. Converging technologies enabling MPP uptake
    • Client/server computing comes of age
    • Emergent lessons from a decade of client/server experiment
    • The five pillars of server computing in the commercial client/server environment
    • UNIX — now acquiring true functionality for data centres
    • The commodity microprocessor
    • RDBMS as the de facto standard for enterprise databases
    • Everyday parallelism
  7. The implications for client/server architectures of the communications revolution
    • The fundamental significance of changing communications capabilities
    • Distributed computing — the motive power of the communications revolution
    • The implications for communications of the client/server model
    • Third-generation communications — the revolution of switched digital data (parallel communications)
    • 'Rightsizing': rightsizing plus
  8. Parallel processing: some conceptual fundamentals
    • Fundamentals of parallel processing
    • Latency parallelism (speed)
    • Throughput parallelism (capacity/size)
  9. The important approaches to parallel hardware architectures
    • Functional multiprocessing (FMP)
    • Shared-memory parallel processing (moderately parallel)
    • Distributed.memory parallel processing (massively parallel)
    • Moderately versus massively parallel processing
    • SMP's probably future
  10. Critical issues for MPP hardware architectures and systems in RDBMS environment
    • The pre-eminence of the network: the key to the machine!
    • Scalability of the network
    • Availability and fault tolerance of the network
    • Usability of the network
    • Node-processor arechitecture
    • Support from open-systems commodity operating systems
    • The traditional 'holy grails' of MPP — perhaps no longer canonical in today's commercial setting
  11. The RDBMS in an MPP environment
    • RDBMS parallel processing in perspective
    • The similarities between RDBMS and operating systems
    • The three main elements of parallel RDBMS processing
    • Parallel SQL — query decomposition for parallel execution
    • Parallel administration — scalable database administration
  12. Adoption strategies for the commercial user
    • MPP: can it be ignored?
    • Commercial MPP usage over the next three years
    • The MPP server as data warehouse for the enterprise
    • The second-generation data warehouse
    • Migration, redevelopment or new development?
    • Coexistence or replacement?
    • MPP RDBMS: not as difficult as it may seem
    • Exceptional candidate applications for parallel-processing solutions
    • Server parallelism: a responsibility of vendors
  13. The Future
    • New markets and uses — the future prominence of 'personal computing'
    • Tomorrow's consumer computing servers — orders of magnitude bigger than 'commercial' computing today
    • Important likely MPP applications of the near to medium term
    • The implications for business today


Enterprise Client/Server Technology

Reviewed by Roland Buresund

Good ******* (7 out of 10)

Last modified: May 21, 2007, 3:04 a.m.

Massively Parallel Processing for Business is the subtitle and is a very good description of the contents. I would recommend it to any CIO that needs to understand what parallel technology can do for him/her.


There are currently no comments

New Comment


required (not published)