The following are two solution blueprints by an additional attempted start-up to license Geneva. In this instance, the platform was to be ported to the Intel Itanium processor, putting a very powerful data analysis tool right under the desk of the analyst.
The vision was far sighted, anticipating what happened with cloud and data analytics. But it was ravaged by the technology collapse post Internet bubble.
More about this time can be read in Balancing Act: A Practical Approach to Business Event Based Insights, Chapter 53. Spin-offs and Chapter 1: Introduction
INTELLIGENCE ENGINE FROM PwC CONSULTING AND DIGINEER™
Consolidating massive amounts of data from multiple sources to gain an integrated view
- OI5 DataEngine™ from Digineer™
- Enterprise Hardware Platform
- Scaleable Intel® Architecture servers
Advanced Transaction Analysis for Business and Government
Crucial data that can help businesses and government operate more efficiently often lies buried within an overwhelming mass of transaction details scattered among numerous computers. Extracting this data using most data mining solutions presents many difficulties, since these solutions lack the flexibility, scope, and sheer processing power necessary to convert uncorrelated data into useful information. The Intelligence Engine solution combines technology expertise from PricewaterhouseCoopers (PwC Consulting), Digineer™, and Intel to help solve this issue. The Intelligence Engine solution spans data repositories across the world, using query-processing techniques to generate data intelligence from large numbers of events and transactions.
FACING THE CHALLENGE
Without the ability to look deep into operational processes, large-scale businesses are hindered in their decision-making and organizational goals––essentially running blind. Where data exists, it often resides on disparate servers in a variety of formats, resisting easy access or useful analysis. Collecting, consolidating, and correlating this data requires tools that can be scaled for massive volumes of data, and adapted easily to expose a wide range of information. Challenges to this end include:
- Huge volumes of transaction data––Many conventional data warehousing and data mining solutions collapse under the weight of millions of daily event-level transactions. A robust solution must marry intelligent query processing with sufficient processing power to meet high levels of activity.
- Data scattered throughout many repositories––Large organizations frequently store data in a variety of formats in data repositories spread throughout the country––or the world. Finding the key data—the proverbial needle in a haystack—demands a solution scaled to the magnitude of the task.
- Need to quickly spot trends and patterns––The accelerating pace of modern business makes it necessary to detect and respond to trends and patterns in days rather than weeks. Knowledge-discovery tools must be flexible and adaptable for this purpose.
- Evolving data requirements––Creating data models and report formats can be a very time-consuming and expensive proposition, and updating them can also require excessive time and effort. Tools to reduce this design effort and the associated maintenance help ensure that changing data requirements can be accommodated in minimal time.
MEETING THE CHALLENGE
Being able to accurately detect emerging patterns and trends in daily operations is invaluable to any organization seeking to remain competitive under the pressures of rapid change. The Intelligence Engine solution incorporates Digineer™ OI5 DataEngine™ deployed on Intel® Itanium® processor family-based servers to provide fast and flexible analytical data handling across multiple systems. By providing consolidated views of dispersed data in a variety of formats, this solution lets large organizations make fully informed decisions within their industry by assimilating data from every group or department––regardless of platform. The OI5 DataEngine™ successfully bridges dissimilar data structures and geographically separated servers to turn raw data into useful knowledge, providing a robust solution to help guide leaders of large organizations.
To accelerate the process of collecting and consolidating data, Digineer’s toolset simplifies query-processing setup and report generation. PricewaterhouseCoopers has successfully deployed this solution in a wide range of industries, including retail, financial, healthcare, and government.
The Intelligence Engine solution lets any large organization benefit from improved visibility into the heavy volume of daily transactions and events specific to their business––whether retail sales tallies per product for a nationwide distributor, the weather conditions compared to yields at a juice producer’s orange orchards, or quality control data tracked by a steel manufacturer looking for ways to improve processes.
Clients adopting this solution enjoy improved operations management through rapid analysis of transaction data. They also gain deeper insights into key processes within their organization that may have been inaccessible in the past because of the difficulty in collecting and correlating the relevant data.
FEATURES : BENEFITS
Supports multiple data formats: The OI5 Data Engine™ provides support for a wide range of data formats, allowing flexible handling of data stored in repositories across many different platforms.
Designed for faster, high-volume transaction processing: The Intelligence Engine solution combines the intelligent pre-processing capabilities of the OI5 DataEngine™ with the architecture benefits of the Intel® Itanium® processor family. The result is a solution that scales well for the extremely large volumes of data involved in transaction processing––producing rapid responses to complex data queries.
Adapts easily to changing data requirements: The solution includes tools that simplify query design and report generation, encouraging companies to create new models for data collection and analysis, while meeting the core data extraction requirements for the organization.
Advanced Business Intelligence and Analytics Solution
Retailers, financial institutions, healthcare companies, and government agencies are collecting more data than ever before. This includes sales data from stores, warehouses, e-Commerce outlets, and catalogs, as well as clickstream data from Web sites. In addition, there is data gathered from back office sources––merchandising, buying, accounting, marketing, order processing, and much more. To pull together information from all these different sources and integrate, process, and analyze it, companies require advanced business intelligence and analytical solutions powered by a robust, high-performance infrastructure.
The Intelligence Engine solution addresses this very need, by combining technological expertise from PricewaterhouseCoopers (PwC Consulting), Digineer™, and Intel®. The query-processing capabilities of the Digineer™ OI5 DataEngine™ combined with the processing power of the Intel® Itanium® processor family, enables PwC Consulting to provide companies with the ability to quickly consolidate and integrate massive amounts of data from multiple sources––giving key decision makers a single view of vital business information.
Optimized to run on Intel® Itanium® processor family-based platforms, the Digineer™ OI5 DataEngine™ delivers a reliable, high-performance mechanism for processing and querying vast quantities of raw, event-level data through a software algorithm that employs intelligent pre-processing for improved efficiency. The Explicitly Parallel Instruction Computing (EPIC) technology and 64-bit addressing capabilities found in Intel® Itanium® architecture, deliver new levels of performance when applied to data mining and knowledge discovery operations. Powerful Intel® Itanium® processor familybased servers provide the massive computing power needed to run and consolidate memory-intensive databases across channels. In addition, integrating different processes and organizations is easier and more cost effective with a modular, open environment based on industry standards and Intel® Architecture.
THE BUSINESS CHALLENGE
Today’s competitive advantage will go to companies that make the right decisions at the right time. To succeed, these companies must be able to quickly consolidate and integrate massive amounts of data from multiple sources, into a single view of their business that can increase marketing and operational efficiency.
Turning Data into Dollars, a May 2001 Forrester report, indicates that executives charged with setting the strategic direction of their organization understand the value of business intelligence and analytics, yet are faced with the following challenges:
- Mountains of data create processing bottlenecks—The sheer volume of data collected–– often involving terabyte-scale databases––creates the potential for processing bottlenecks in those systems not designed to effectively cope with this quantity of information. The massive data quantities produce such a heavy processing load that many existing solutions rapidly reach throughput limitations. Without sufficient processing power and highly optimized analytical tools, organizations must often relinquish access to important subsets of their data, or must add expensive hardware and software to increase system performance.
- Analyzing information from scores of data repositories presents a challenge— Organizations cannot easily access and analyze their operational information because of the difficulty in extracting large quantities of data from multiple data repositories spread throughout the organization. Existing solutions lack the flexibility and robustness for effective data access and extraction.
- Inflexible models fail to support evolving data streams—Organizations spend significant amounts of development time creating data models, collecting data, defining reports, and constructing consolidated data warehouses. In highly dynamic environments, these systems may rapidly lose relevancy and require ongoing adjustments at the core software or hardware level. Updating these core data model components to correspond with evolving knowledge discovery requirements necessitates a significant investment in time and effort.
- New analytical views must be developed quickly and on demand—Different departments and groups within an organization have unique needs for the information mined from available data, and these needs change frequently, often on an urgent basis. A practical solution must have the capability of easily providing new analytical views from the data to respond to immediate needs from diverse groups.
- Extracting key information is inordinately difficult—The ability to drill down through individual transactions and extract the useful patterns and trends is critical to knowledge discovery and analysis. Most existing systems do not adequately provide this capability when there are large quantities of data stored on multiple systems, or in one large system where data access is cumbersome.
The retail environment faces several challenges that can take advantage of a business intelligence solution like Intelligence Engine. For example, in a typical retail environment, data gathered from stores, Web sites, catalogs, as well as suppliers and warehouses provide valuable kernels of knowledge that can increase profitability. Consolidating and analyzing massive amounts of memory-intensive databases requires a robust infrastructure. As a result, many retailers mine only a subset of their available information. To recognize buying patterns over time, massive volumes of sales transactions must be analyzed and correlated by sales channel, region, individual store, product, and customer preference. Within the typical retail environment, this data resides in a number of separate systems.
THE SOLUTION OVERVIEW
The Intel® Architecture-based Intelligence Engine solution provides companies with a fast, affordable, and flexible way to consolidate, integrate, and analyze massive, memoryintensive data across multiple systems.
The Digineer™ OI5 DataEngine™, operating on a platform powered by the Intel® Itanium® processor family, performs high-speed analytical processing of massive amounts of event-level data. The data engine can simultaneously write large numbers of data extracts to downstream data warehouses, online and offline reporting tools, and real-time analytics applications. The 64-bit addressing capabilities of the Intel® Itanium® architecture provides the robust computing power needed to run and consolidate memory-intensive databases across channels. The extraordinary performance of the Intel® Itanium® processor family, coupled with the EPIC features, provide an architecture fully capable of adding capacity as needed, and providing heightened performance.
Historically, retailers have addressed the scale and performance problem by creating large data stores using conventional technology. As the volume of data that must be captured and processed increases––often exponentially––processing the massive volumes of data creates a bottleneck and prevents timely access to the detailed information required to run the business. Once this critical point has been reached, retailers must choose between adding expensive hardware to boost performance, or sacrificing the breadth and depth of captured data.
The OI5 DataEngine™ offers an attractive alternative to the scale and performance problem. Retailers use the solution to analyze key retail metrics, such as customer profitability, recency, frequency, volume analytics, supplier performance, and enterprise financials. The OI5 DataEngine™ derives these analytics from a collection of customer relationship management (CRM), supply chain management (SCM), and point-of-sale (POS) databases–– even if the data is distributed among several different machines. The performance provided in these kinds of implementations costs much less than equivalent systems.
The Digineer™ OI5 DataEngine™ delivers:
- Speed—Performs fast and efficient processing of massive volumes of transactional data, resulting in a consolidated data store that accurately represents the organization’s operational dynamics.
- Scale—Integrates data from multiple, disparate silos without requiring any modifications or enhancements to the existing data sources.
- Depth—Includes powerful data query capabilities that efficiently and economically enable data mining of massive amounts of transaction-level event data.
- Breadth—Provides an outstanding price and performance ratio, making the solution accessible to medium- and large-sized organizations.
PwC Consulting fully integrates the OI5 DataEngine™ into the client’s business processes and technologies. PwC Consulting’s background and experience in business intelligence and analytics produces a solution that enables organizations to rapidly collect, aggregate, manage, analyze, filter, customize, and distribute information throughout their value chains. By providing consolidated views of the dispersed data––rapidly and cost-effectively–– organizations can identify trends and quickly respond to changing needs and situations, thus improving overall business performance.
The Intelligence Engine solution runs efficiently on a 4-way Intel® Itanium® processorbased server. Supported operating systems include (Win64) Microsoft* Windows* 2000 server (Q3’02), Linux* (Red Hat*, version 7.2 for the Itanium® processor), and HP-UX* version 11 (Q4’02).
Components of the Digineer™ OI5 DataEngine™ rely on the following technologies:
- Core Data Engine™—C++ with multi-threading
- Result Set Viewer—Java* using Swing*
- MetaData Manager—Visual Basic*
- MetaData Database—Any ODBC-compliant database
The OI5 DataEngine™ runs optimally using Intel® EPIC architecture, taking maximum advantage of the Itanium® processor family’s multiple memory registers, and massive on-chip resources. The 128-bit general and floating-point registers excel at supporting the complex operations involved in analytics. Tuning the implementation to achieve maximum performance through parallelism, and tapping into the power of the Itanium® processor family’s multiple execution units makes it possible to rapidly process and analyze the large volumes of data necessary to accomplish this solution.
PwC Consulting solutions employ a variety of technologies using Digineer™ OI5 DataEngine™. By being able to create data stores through the importing and exporting of data files, this solution makes it unnecessary to devise complex, custom integration schemes. Support for each of the following sources and targets is provided:
- Source databases
- BM* DB2*
- Microsoft* SQL Server*
- Target analysis engines
- EMC* Symmetrix Manager*
- SAP* portals
- IBM* Intelligent Miner*
- Cognos* Powerplay* and Improptu*
- Custom applications built in Visual Basic, Java, C++, and HTML using Microsoft Internet Information Server* (IIS) and Transaction servers
- Digineer™ OI5 DataEngine™ has been deployed for use with SAP* and PeopleSoft* ERP packages as sources for data and targets for additional analysis. Future plans include development of interfaces for MQSeries*, and clickstream data sources, scheduled for release by mid-2003.
WHO THE SOLUTION WILL BENEFIT
The Intelligence Engine solution provides rapid analysis of data from multiple sources to gain a better picture of emerging trends and patterns. For a strong competitive advantage, companies will need to develop a quantitative, as well as qualitative, understanding of their business. The Intelligence Engine solution complements existing data warehouse and business intelligence solutions. It can also maximize the performance and extend the functionality of systems already in place.
The key vertical integration areas in which the Intelligence Engine solution can provide significant value include:
- Retail—Domestic and international (Western European) retailers with annual sales of more than $1 billion, and large-scale customer, supplier, point-of-sale, and operational databases.
- Manufacturers—Mid-market manufacturers with annual sales of more than $500 million and one or more large silo databases.
- Financial Services—Financial services organizations with multiple geographic locations and stringent operational and regulatory reporting requirements.
- Healthcare—Pharmaceutical and biotechnology companies that depend on data from multiple systems to manage and optimize their operations.
- Government—Agencies under pressure to provide more extensive reporting of data, over which the organization has little governance. In addition, data sources continue to propagate, grow in size, and are increasingly complex and divergent from each other. 5
The PwC Consulting Intelligence Engine solution using Digineer™ OI5 DataEngine™ adapts well to environments where operational data is distributed among several different systems––sometimes in different data formats––and where the report requirements change frequently. Implementing the Digineer™ OI5 DataEngine™ on servers powered by the Intel® Itanium® processor family offers these benefits:
- Provides faster high-volume transaction processing. Supplies the massive computing power needed to run and consolidate memory-intensive databases across channels. This enables key decision makers and business owners to obtain accurate business intelligence and analytics, and quickly gain deeper insights into the dynamics of their operations.
- Supports multiple data formats. Supports multiple applications and loads required to consolidate and analyze data throughout the company. This enables the solution to easily access existing data stores and business intelligence solutions without modifications or retrofitting.
- Adapts easily to changing data needs. Accommodates an organization’s evolving data needs and models so that new informational requirements can be met with minimal investment of time and money. Integrates different processes and organizes these in an easier and more cost-effective manner with a modular, open environment based on industry standards and Intel® Architecture. Solutions based on the OI5 DataEngine™ as deployed on servers powered by the Intel® Itanium® processor family provide a highly competitive price and performance ratio. 6
FUNCTIONAL BUSINESS CONCEPT
Maintaining a current and accurate view of an organization’s operational status requires the capability of locating and analyzing information from a very large data store in a short period of time. This problem can be compounded when data resides in disconnected data silos, or has been extracted from a variety of non-correlated reports. Reducing the scope of the data analyzed in order to save time or expedite processing can create an inaccurate view of current operations. Processing bottlenecks and inefficiencies often result from solutions implemented on platforms without a tuned architecture capable of handling data quantities that reach terabyte levels. While increasing the quantity and complexity of a solution’s hardware and software may improve performance, it also raises costs. A better approach is to deploy an implementation on a platform designed to accommodate massive volumes of data with processing power that is suitable to the task. The Intelligence Engine solution using the OI5 DataEngine™ on a platform powered by the Intel® Itanium® processor family performs very high-speed, analytical processing of massive amounts of event data. This event data can be dispersed across any number of data sources, including enterprise resource planning (ERP), SCM, CRM, and legacy systems–– without reducing the effectiveness of the solution. Processed data can then feed downstream data repositories, analytical and business intelligence tools, and executive reporting systems, as shown in the following diagram.
The OI5 DataEngine™ uses a patented software approach that leverages the latest advances in microprocessor technologies featured in the Intel® Itanium® architecture, including 64-bit addressing and parallel processing capabilities.
The open architecture of the OI5 solution allows data to be extracted and transformed directly from legacy systems. OI5 provides a massively scaleable, high-performance solution for the most demanding intelligence and analytics applications
The first step to using OI5 is to define the metadata, including record layouts, files or tables, and relationships between tables. The OI5 design removes the complexities of understanding table relationships. Common keys and other relationships between tables only need to be defined once. Instead of having to understand the “where” statement in SQL to add department names to a sales report, OI5 lets users select the department name and then add it to their report. OI5 automatically locates the correct data.
Next, define the views (queries):
- Select the transaction files to read
- Specify the appropriate output format
- Indicate the appropriate filters to apply
- Indicate the data to sort or summarize
- Design columns, and indicate any calculations to perform
- OI5 does the rest
During data file output, OI5 produces multiple files in one execution of the process. This feature reduces the time needed to supply data to other systems. OI5 views can reformat data,perform logic operations, and make calculations. With the open API, custom functions can be created for other types of transformation. These features make OI5 a powerful data extraction and transformation tool, as well as a flexible application engine.
The OI5 DataEngine™ helps large organizations meet the scaleability and performance challenge of assessing key organizational operations when the number of events being measured and analyzed becomes too great. The solution supports queries directed to extremely large volumes of event-level data in a single pass.
Data exposed by the OI5 DataEngine™ can reside in different physical locations on multiple systems, including ERP, CRM, SCM, and POS systems. The front end of the OI5 DataEngine™ maps these “in-place” data sources into the data engine’s MetaData definition modules. The engine then uses the MetaData definition to intelligently pre-process the queries before generating the data views and outputting data stores for analytical processing. The data output can either be viewed with the OI5 Viewer or directed to other business intelligence and reporting systems.
The core, patented technology of OI5 is configured for refreshing and managing a largescale operational data store (ODS) derived from multiple data sources. As shown in the following diagram, it includes three integrated modules that deliver the full power of the OI5 technology and its massive data throughput: OI5 MetaData Manager, OI5 View Manager, and OI5 Core Processor. The OI5 DataEngine™, as implemented on servers powered by the Intel® Itanium® processor, runs on Microsoft* Windows* 2000 Server (64-bit) and HP-UX systems.
The parallel query-processing capabilities of the system follow a defined software algorithm that is optimized for speed. The algorithm first pre-processes all the data queries based on the data engine’s knowledge of where and how the event data is structured and formatted (determined by the MetaData definition in combination with Read Exits), as well as the required output data stores (known as Views). Optimized code generated by this query pre-processing step gets sequenced into a logic table. When the logic table is processed, the OI5 DataEngine™ spawns parallel query instructions that minimize instruction path lengths to the available microprocessors. This explicitly parallel query-processing directly takes advantage of the EPIC architecture of the Itanium® processor family. The OI5 DataEngine™, using instruction path optimization and single-pass data queries, offers unmatched query-processing and exceptional handling of the data view output.
The OI5 DataEngine™––the basis of this data warehouse analytics solution––is optimized for the Intel® Itanium® processor family. The following diagram shows how information from existing data warehouses is delivered to the OI5 DataEngine™ for advanced analytics processing. The OI5 MetaData and View Manager applications, running on a workstation powered by the Intel® Pentium® 4 processor, define the relations and views for the associated processing and analytics. Processed data can be directed to Web portal solutions, as well as custom applications and reports, or supplied in a form for re-entry into an existing data warehouse.
The Web portal is powered by the Intel® XeonTM processor family. The OI5 DataEngine™ is powered by a 4-way Intel® Itanium® processor-family based server.
PwC Consulting has used the OI5 DataEngine™ in numerous scenarios that have been designed and deployed for clients. This solution achieves favorable results in knowledge discovery applications that involve massive numbers of transactions. Running on servers powered by the Intel® Itanium® processor family, the underlying OI5 software data engine performs very high-speed, batch-analytical processing of large volumes of event data–– while simultaneously writing multiple data views to downstream data warehouses, online and offline reporting tools, and real-time analytical processing applications. When used in combination with Intel® Itanium® processor family-based platforms, the OI5 DataEngine™ offers an exceptional price and performance value.
LEARN MORE ABOUT THIS INNOVATIVE SOLUTION
For general information about the products described in this data sheet, visit: www.digineer.com
If you have a specific question about implementing this solution within your organization, contact your Intel representative or e-mail us at:
Digineer, the Digineer logo and OI5 DataEngine are trademarks or registered trademarks of Digineer, Inc.
Copyright ©2002 Digineer. All Rights Reserved.
Intel, the Intel and Intel Inside logos, Pentium and Xeon are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries.
Copyright ©2002 Intel Corporation. All Rights Reserved.
*Other names and brands may be claimed as the property of others. Information regarding third party products is provided solely for educational purposes. Intel is not responsible for the performance or support of third party products and does not make any representations or warranties whatsoever regarding quality, reliability, functionality, or compatibility of these devices or products.
250790—001 and 250789—001