| [First! for the Web]
Tandem to fine-tune NonStop DB
---------------------------------------------------------------------------
Date: Monday, February 3, 1997
Source: PC Week
---------------------------------------------------------------------------
PC Week via Individual Inc. : To strengthen its position in the data
warehousing market, Tandem Computers Inc. next month will announce it is
enhancing its proprietary NonStop SQL/MP relational database for data
mining.
The enhancement will let users embed precalculated, fully indexed crosstabs
into relational tables, thus shortening the process of mining the data,
sources said.
Tandem will detail this technology and announce partnerships with a handful
of data mining tool vendors at an event on Feb. 11 in Boston. The vendors
will modify their tools to access and read the Tandem database, sources
said.
"Tandem is doing some interesting things for extending their system to make
it useful for a variety of data warehousing functions, and this another
step in that direction," said Herb Edelstein, president of Two Crows Corp.,
a consultancy in Potomac, Md. "NonStop SQL is one of the best databases.
... This further enhances its functionality. Its only drawback is that it
runs only on Tandem systems."
NonStop SQL/MP supports parallel queries and features an optimizer that can
choose among a variety of query techniques such as hash joins, merge joins
and accelerated groupings.
The data mining functionality would be another step in Tandem's quest to
position its database and high-end Himalaya servers as backbones for large,
mission-critical data warehouses.
A system based on Tandem's Himalaya servers can scale to more than 4,000
processors while single tables can grow to 1 terabyte in size, officials of
the Cupertino, Calif., company have said.
Juan Carlos Perez, PC Week
---------------------------------------------------------------------------
[Feedback] [Index] [Search] [Customize] [Individual, Inc.]
[Main] [Up] [< Prev Story] [Next Story >] [Calendar]
---------------------------------------------------------------------------
[Individual, Inc.]
|
| [First! for the Web]
Tandem Introduces Object Relational Data Mining Solutions and Services
for Vertical Markets; Business-Driven Offerings Target Card Marketing,
Micromerchandising, Claims Analysis and Other Key Applications
[Feedback] [Index] [Search] [Customize] [Individual, Inc.]
[Main] [Up] [Prev Story] [Next Story] [Calendar]
----------------------------------------------------------------------
Date: Wednesday, February 12, 1997
Source: Business Wire
----------------------------------------------------------------------
CUPERTINO, Calif.--(BUSINESS WIRE) via Individual Inc. -- Applying its
vertical market expertise and new Object Relational Data Mining
architecture to real-world business problems, Tandem Computers
Incorporated today launched a series of Object Relational Data Mining
solutions packages for card marketing, micromerchandising and
insurance claims analysis.
Tandem also announced new consulting services designed to allow
companies to quickly enjoy low-risk, discovery-driven decision making.
The solutions and services are based on Tandem's revolutionary new
Object Relational Data Mining architecture. This enables customers to
efficiently mine their entire database, not merely samples, for useful
patterns and trends. The result is a more effective realization of the
full business value of data.
"Object Relational Data Mining solutions add significant new
functionality to customer segmentation and predictive modeling
techniques," said Jonathan Kalman, managing director of MRJ Technology
Solutions, a leading specialty systems integrator. "Tandem is taking a
profoundly different approach by integrating its powerful database,
capable of handling an entire organization's data, with leading data
mining tools."
Delivering full value of business data
The new solutions packages will be comprised of the cross-platform
Tandem ServerWare database, appropriate integrated data mining and
other analysis tools from leading solutions partners, Tandem S-Series
massively scalable Himalaya and/or Microsoft Windows NT Server-based
hardware platforms, application and reporting templates, data models,
and Directional Consulting services. Though specially tested and
packaged, the solutions are all easily customizable. Initial solutions
include:
-- Card Marketing
Aimed at card acquirers and issuers, this solutions package applies
Object Relational Data Mining architecture and other decision support
technology to improve the effectiveness of cardholder retention and
acquisition efforts. This provides a better understanding of when
certain customers are likely to leave and why, leading to more
effective customer segmentation, increased response rates to marketing
promotions, and improved margins through targeted product development
and pricing.
-- Micromerchandising
This package enables retailers to mine immense volumes of detailed
merchandising data, resulting in improved in-stock positions, reduced
markdowns by better understanding buying patterns and trends, enhanced
promotional effectiveness, and improved store profitability through
more precise forecasting.
-- Claims Analysis
Aimed at insurance providers looking to contain underwriting costs and
improve loss ratios, this package uses Object Relational Data Mining
technology to support new product development, fraud profiling and
detection, better service provider alliances, and more exact
underwriting experience comparisons.
Immediate customer reaction to these benefits is positive. Said Juan
Verastigui, director of Claims System Development at USAA, a leading
insurance company, "Tandem's Object Relational Data Mining
architecture and the way it leverages the parallel ServerWare database
will provide USAA with the ability to derive full value from all our
claims data, and not just subsets. The resulting faster and more
complete answers to our business queries will have a very positive
effect on our bottom line."
Looking ahead, Object Relational Data Mining architecture will enable
the mining of complex data types that include voice, video and images.
Said MRJ's Jonathan Kalman, "Object Relational Data Mining solutions
provide immediate value with traditional data types, and extensibility
to meet future multimedia analysis needs."
Directional Consulting, new Object Relational Data Mining services:
Tandem's Directional Consulting services are an integral part of the
new solutions packages and are also available separately. These
services define a low-risk, high-return methodology -- proven over
many Tandem based data warehousing implementations -- for exploring
and understanding how data mining can support particular business
initiatives.
Directional Consulting services use a phased approach to having data
mining production environments up and running within 90 days. The
process begins with establishing priorities for implementation of
Object Relational Data Mining and proceeds to a "proof of concept"
phase to verify that the selected data mining solutions will meet
expectations. System design, data modeling and implementation then
follow, culminating with the establishment of a robust, scalable
operational environment that supports application evolution and
growth.
Availability
Tandem Card Marketing, Micromerchandising, and Claims Analysis
solutions will be available beginning in the first quarter of 1997.
These will be enhanced to take advantage of Object Relational Data
Mining technology in third quarter of 1997.
About Tandem
Founded in 1974, Tandem Computers Incorporated designs and delivers
technology solutions that companies rely on to compete in a business
world that runs 24 hours a day. A US$1.9 billion company headquartered
in Cupertino, California, Tandem has offices, strategic partners, and
providers in more than 50 countries around the world.
Note to Editors: Tandem, Object Relational Data Mining, NonStop,
Himalaya, ServerWare, and the Tandem logo are trademarks or registered
trademarks of Tandem Computers Incorporated in the United States
and/or other countries. Microsoft and Windows NT are either trademarks
or registered trademarks of Microsoft Corporation in the United States
and other countries. All other brand or product names are trademarks
or registered trademarks of their respective companies.
CONTACT: Tandem Computers Incorporated, Cupertino | Kristine Austin,
408/285-6645 | World Wide Web Home Page Address: http://www.tandem.com
----------------------------------------------------------------------
[Feedback] [Index] [Search] [Customize] [Individual, Inc.]
[Main] [Up] [< Prev Story] [Next Story [Calendar]
----------------------------------------------------------------------
[Individual, Inc.]
|
| [First! for the Web]
Tandem Introduces Platform-Independent NonStop SQL/MX/ First Database
Designed For Cluster Computing; Two terabyte clustered database running on
64 Pentium Pro processors proves Windows NT Server scalability for
large-scale data warehousing
---------------------------------------------------------------------------
Date: Thursday, May 15, 1997
Source: Business Wire
---------------------------------------------------------------------------
CUPERTINO, Calif.--(BUSINESS WIRE) via Individual Inc. -- Tandem(R)
Computers Incorporated today introduced its NonStop(R) SQL/MX open,
parallel database management system, part of its NonStop Software offerings
for business-critical cluster computing. The new platform-independent
database extends Tandem's database performance and rock-solid availability
to the Windows NT(R) Server platform, while adding advanced data mining,
object handling, and query optimization features.
NonStop SQL/MX runs on any Intel based Windows NT Server based system
(single symmetric multiprocessor or clustered configuration), as well as on
Tandem's NonStop Himalaya(R) servers, providing growing applications with a
well-defined and seamless growth path. When run on Windows NT Server based
clusters, NonStop SQL/MX applications inherit the linear scalability,
reliability, ease-of-management, and parallel query capabilities that have
made the Tandem NonStop platform the leading choice for decision support,
data mining, operational data stores, and business-critical online
transaction processing.
As proof of its clustering capabilities, Tandem is currently demonstrating
the ability of NonStop SQL/MX to manage a 2 terabyte database containing a
large table of 30 billion rows, divided into more than 700 partitions for
massive parallelism, and running on a 64-processor Windows NT Server
cluster. NonStop SQL/MX maintains a single database image across all the
servers in the cluster, allowing the application to be managed as if it
resides on a single server. The database also supports near-linear
scalability as new nodes are added, meaning that each new node participates
fully in the parallel processing and contributes a full server's worth of
performance as a result of NonStop SQL/MXs cluster-efficient architecture.
Windows NT Server database applications come of age
"NonStop SQL/MX marks the debut of Windows NT Server clustering as a viable
alternative for business-critical database applications," said Bill Heil,
Tandem senior vice-president and general manager of the company's
ServerWare business unit. "Cluster technology is accelerating the move
towards inexpensive commodity platforms and NonStop SQL/MX has by far the
most advanced clustering capability in the industry. It's built using the
same technology that has been driving the world's most demanding
applications for the last twenty years, and it's easy to use -- most of our
Himalaya server customers don't even know they're running a clustered
environment."
The demonstration of NonStop SQL/MX on Windows NT Server uses genuine
retail data supplied by Dayton Hudson Corporation. "Windows NT Server is a
strategic platform for Dayton Hudson, and we're very interested in the
price/performance that it can bring to large-scale data warehousing.
Tandem's work with Windows NT Server helps us push the envelope further in
that direction," said Mike Thyken, director of technical services, Dayton
Hudson Corporation.
Combined with the robust database capabilities provided by NonStop SQL/MX,
cost-effective Windows NT Server solutions will now be deployed for such
applications as Web-enabled decision support, Internet transaction
processing, micromerchandising, and financial fraud and risk management
among others. As these applications grow, NonStop SQL/MX will allow them to
seamlessly scale up to the NonStop Himalaya platform when appropriate.
NonStop SQL/MX applications will also interoperate across the Windows NT
Server and NonStop Himalaya platforms.
NonStop availability
NonStop SQL/MX takes full advantage of Tandem's clustering technologies to
support the continuous availability for database applications. This
includes supporting sub-second take over in the event of a failure in a
processor or cluster node. In these cases, critical tasks are immediately
taken over by other nodes using Tandem's patented process-pair technology.
This capability has helped make Tandem the platform of choice for such
high-availability applications as automated teller machines, stock trading,
credit card processing, directory assistance systems, cellular roaming, and
emergency 911 systems.
In addition, NonStop SQL/MX minimizes planned downtime by allowing key
database administrative tasks to be performed online. For example, database
updates and data reorganizations can be made even while transactions or
queries are occurring against the database.
Integrated Object Relational Data Mining and DataBlades
On top of its scalable, parallel architecture, NonStop SQL/MX employs a
combination of advanced technologies to catapult database performance,
reliability, and functionality even further beyond that available from
competing products, regardless of platform. These technologies include
integrated Object Relational Data Mining(TM) architecture and support for
DataBlade technology.
Data mining helps businesses discover useful patterns and trends hidden in
their data. In the future, this technology will enable the mining of
complex data types that include voice, video and images. With its Object
Relational Data Mining architecture, Tandem is tightly integrating a range
of core data mining functionality into the NonStop SQL/MX database. This
integration allows data mining algorithms to work against a customers
complete data set, and not just file extracts as is the case with
conventional data mining. This, in turn, leads to faster, more accurate,
and more complete results.
The support of DataBlade technology, meanwhile, will allow complex data
types such as voice, video, and other multimedia objects to be included in
a comprehensive query processing and data mining environment. For example,
in a medical application, users will be able to bring up not only
structured patient records (name, address, etc.), but also x-rays, charts,
and other image information.
Advanced parallel and query management technologies Other key features that
contribute to the industry's most advanced parallel database environment
include comprehensive parallel operations, a unique object-oriented data
flow architecture, an advanced data access manager, parallel transaction
management facilities, and more (see attachment for details).
Availability
NonStop SQL/MX will ship in Fall 1997. A software developers kit will be
available in June 1997.
Founded in 1974, Tandem Computers Incorporated designs and delivers
technology solutions that companies rely on to compete in a business world
that runs 24 hours a day. A US$1.9 billion company headquartered in
Cupertino, California, Tandem has offices, strategic partners, and
providers in more than 50 countries around the world.
Tandem press materials are also available on NEWSdesk International, on the
World Wide Web at http://www.newsdesk.com .
Advanced NonStop SQL/MX Technologies
- Parallel operations -- NonStop SQL/MX extends parallelism to every stage
of complex query processing. All major query tasks -- table scans, joins,
sorts, index creations, and so on -- can be divided into multiple subtasks
for parallel execution.
- Data flow architecture -- A state-of-the-art object technology, data flow
architecture provides dramatic query performance improvements through the
easy rearrangement of query operation steps in an query execution tree.
- Advanced data access manager -- NonStop SQL/MX's data access manager
eliminates the need for a distributed lock manager by performing SQL
operations (selection, projection, aggregates, join in some cases, and so
on) directly on data as it comes off the disk. This significantly improves
performance over competing databases because only relevant data is sent to
the user's program.
- MultiDimensional query technology -- NonStop SQL/MXs MultiDimensional
Access Method (MDAM) provides efficient access to data in multiple
dimensions without requiring a separate index for each dimension or
resorting to separately maintained "data cubes."
- New query optimization technology -- NonStop SQL/MX combines rule and
cost-based optimization to select the best query plan.
- Optimal lock management -- Unlike other cluster databases that require a
distributed lock manager, NonStop SQL/MX eliminates the need to send lock
requests to other nodes in the cluster by isolating locks on the node to
which the data is assigned.
- Parallel transaction management -- NonStop SQL/MX presents a single
transaction audit log image, though the log is written in parallel for
scalable transaction throughput. Reading of the log for rollback or
recovery also takes place in parallel.
NOTE TO EDITORS: Tandem, Himalaya, NonStop, and the Tandem logo are either
registered trademarks or trademarks of Tandem Computers Incorporated in the
United States and/or other countries. Windows NT is a registered trademark
or trademark of Microsoft Corporation in the United States and other
countries. UNIX is a registered trademark in the United States and other
countries, licensed exclusively through X/Open Company Ltd. All other brand
and product names are trademarks or registered trademarks of their
respective companies.
CONTACT: Tandem Computers | Kristine Austin, 408/285 6645 | World Wide Web
Home Page:http://www.tandem.com
---------------------------------------------------------------------------
[Feedback] [Index] [Search] [Customize] [Individual, Inc.]
[Main] [Up] [< Prev Story] [Next Story >] [Calendar]
---------------------------------------------------------------------------
[Individual, Inc.]
|