Case PowerPoint slides

Download Report

Transcript Case PowerPoint slides

NCR CORPORATION
A Data Warehousing Solutions Provider
Presented by:
Dave Raspberry
Cheng Murray-Khoo
Eric Braun
NCR Overview

Founded in 1884 by John Patterson

Began as a cash register company, entered
the computer business in 1960

Bought in 1990 by AT&T, later divested

Purchased Teradata Corporation in the late
eighties
NCR Overview

Based on Teradata and products based on
the Intel chip, began the platform known as
the WorldMark series

Subsequently, Data Warehousing is
becoming the dominant business in NCR’s
product line

Now pursuing an across-the-board approach
to customers--large and small
How is NCR different?

Utilizes Parallel Processing and ER logical
modeling
– the only fully parallel solution on the market
featuring parallel loading, processing, and
archiving

Strategic approach based on three factors:
– Performance
– Scalability
– Ease of Setup and Support
Parallel Processing- some lessons learned

Divide the rows evenly

Optimize the SQL requests

Provide a scalable interconnect

Load and restore data

Provide for easy system administration
Scalable Data Warehouse Framework

When building a data warehouse, NCR’s
approach involves areas throughout the
organization, IT resources, the technologies, the
processes, and the businesses

SDW accommodates independent data marts or
directly building a DW with dependent data marts

Recommends beginning with a small scalable
data warehouse, focusing on one or two area.
Scalable Data Warehouse Framework

There are four dimensions of the scalable system
– The ability to input and extract data with
consistent response times
– The number of users or queries that can be run
simultaneously
– The environmental complexity of the data
model and the queries being run against the
model
– The degree of support needed to maintain
scalability
NCR’s Logical Data Model Philosophy

NCR believes star schema’s limit business
intelligence

During modeling process, prefers Third Normal
Form

Denormalizing helps the DBMS, but hurts the
quality of information

Third normal form avoids data integrity
compromises
NCR’s Logical Data Model

The four hardest things for a database to do:
–
–
–
–
Join Tables
Aggregate Data
Sort Data
Scan large amounts of data
WorldMark Servers

Four generations of large and medium scale servers

Optimized for Teradata architecture and Bynet technology

WorldMark 4800 and 5200 operate on Intel chip technology

4800 is scalable, and designed to run on applications from
50 GB to 1 TB

Can be upgraded to the WorldMark 5200

Designed for large scale data warehousing, 400GB to
100TB
Teradata Relational Database
Management System (RDBMS)

Most powerful decision support parallel relational
database.

Realistically support data warehouses in excess
500 gigabytes of user data.
WHAT MAKE NCR TERADATA SO SPECIAL?
Teradata runs on Symmetric
Multiprocessing (SMP) hardware platform
Each access module process
(AMP) as one processor.
Each AMP executes the query
functions and data
management.
Each AMP controls and
maintains a portion of database
stored on the disks, TASKS
CAN PERFORM IN PARALLEL
as assigned by VNET.
Each AMP acts as a “unit of
parallelism”.
Teradata’s Unconditional Parallelism
NCR’s Teradata patented
“shared nothing” architecture.
AMP
Teradata runs on Massively Parallel
Processing (MPP) hardware platform
Interconnect
that coordinates
and synchronizes
the activities of a
large number
of SMP nodes
Scalability - the cornerstone of
Teradata RDBMS
BYNET design can linearly scale to support up to 4096 SMP nodes
on a single system.
Database Capacity:



128 Terabytes of data
1, 024 SMP nodes
32, 768 physical processors
NCR Scalable Data Warehouse

NCR handles the world’s biggest data
warehouse for decision support without
compromising the data integrity by denormalizing the tables.

NCR is the only vendor in the market utilizing
parallel processing with ER logical modeling.
NCR’s Alliance Partners
NCR’s Strategies

Customer Privacy
– Claims to be the first solutions firm to place customer privacy
at the center of its strategy
– New database products from NCR will enable marketers to add
a customer's personal preference on how their data should be
used

Neighborhood Retailing

Customer Management Solutions

Customer Relationship Management
How Do Customers Benefit From NCR?

Retailers analyze their profitability on a product, customer, and
store-level basis

Banks use NCR's data warehousing to separate the profitable
customers from the freeloaders

Supermarkets are working on an automated pricing system for
every item in the store

In the communications industry, NCR provides much needed
information management and the ability to bring new products to
market quickly.
NCR Customer’s Complaints

Databases from Oracle, IBM and Microsoft
are quickly adding all the features that
perform sophisticated requests on
multiterabyte data-warehouses

Changing ownership and leadership

NCR’s size
Competition and Problems

Competition:
Andersen Consulting, EDS, IBM, and Unisys

Still trails Oracle, Compaq Computer, IBM and HewlettPackard in data warehouse sales

Core business in computer systems is declining by
almost 30% a year and data warehouse sales haven’t
picked up the slack

NCR relies on international markets for 60% of its revenue

Stock price has dropped nearly 50% over the last four
months
Conclusion and Future Outlook

Now dedicated to becoming the leader in data warehousing

Focused on putting together a solid strategy to help
businesses understand who their customers are and how
best to serve them

Outsources to Solectron nearly everything it used to
manufacture itself, resulting in great cost savings

Forecast six to ten acquisitions in the $20 million to $100
million range in the near future

Possess the needed technology and assets

Planning initiatives in e-commerce
NCR
Questions…?