Tech2Tech
Ask the Experts
Take It to the Top
Powerful systems help drive organizations toward industry leadership.
Deploying a data warehouse often requires a fundamental shift in considering how information is used. Since 1979 Teradata has been building business intelligence (BI) and data warehouse solutions to enable companies to turn data into usable information. Three primary capabilities—workload management, ease of use and failover—are the cornerstones for a successful data warehouse and satisfied end users. This becomes evident as soon as organizations load their data and run a few queries.
Once these companies recognize that they can make timelier and more accurate business decisions with the Teradata solution, they often find themselves in the role of industry or market segment leaders. However, before this can transpire, an important transition must occur from focusing on the speed of transactions to understanding how business users can leverage and manipulate data.
To better understand this transition, Teradata Magazine spoke with Harry Huber and Jeff Gleason. Huber has been in IT for more than 33 years and with Teradata since 1988. He is a Teradata solutions marketing specialist and has significant experience with retail, telecommunications and the U.S. intelligence community.
Also a Teradata solutions marketing specialist, Gleason is a veteran of 24 years in IT, with more than 20 years focused directly in data warehousing and BI. He has been with Teradata since 1989.
Q: Can you start by contrasting the value of an integrated data warehouse with some of the alternative approaches commonly used to answer business questions?
A: Companies are familiar with operational systems that feature speed of data consumption and keyed record retrieval. Such systems are often purchased based on their ability to run the same queries or processes faster and cheaper.
Unfortunately, many companies use these traditional cost/benefit models as business intelligence solutions. After the solution is deployed, however, this often leads to less-than-desirable return on investment [ROI]. Organizations are looking for new business insight—not just the same questions answered faster.
To put it another way, the mathematical equation for the acquisition and deployment of a data warehouse is: returns - cost = value to the end users. A system that costs $2 million and allows the end users to drive $10 million of enhanced revenue to the corporation is far more valuable than a solution that costs $1 million and drives only $2 million of revenue to the corporation.
Q: What makes the data warehouse a better solution for organizations?
A: A data warehouse can do far more than deliver simple, standardized reports common with operational systems. The single most important element that enables deep business insight is workload management. With workload management, the system can simultaneously handle all types of queries and processes against a complex, integrated, single copy of the data.
Furthermore, Teradata systems support active data warehousing, in which queries that range from simple lookups to complex data mining are run concurrently with system utilities and data loads.
Q: How does Teradata compare to some other vendors in the way queries are handled?
A: Some platforms’ database management systems [DBMSs], hardware and operating systems [OSs] have distinct and separate processes or programs to optimize their operations. Often, however, these separate solutions directly conflict with elements from another process. Teradata is unique in the industry as it integrates the workload management processes in its hardware, OS and DBMS environments via Teradata Active System Management.
This integrated solution controls the different pieces of the workload that enter the Teradata system and processes them based on the time of day, level of importance and other pre-assigned factors. For example, ad hoc queries that are given a high priority would be processed during critical business hours, ahead of any low-priority transactions. Then these low-priority transactions would be processed as machine resources become available or during non-critical hours. At night, critical load jobs might be moved ahead in the queue and processed to prepare for the next business day.
Q: What sets the Teradata system apart from other vendors’ products?
A: While the Teradata system offers many of the same base performance enhancements—such as indexing and OLAP [online analytical processing]—as other vendors’ products, nearly all Teradata customers find that they need fewer of these items. The Teradata Database processes complex queries against the base normalized data, therefore eliminating a vast majority of pre-joined, indexed or replicated summary data.
Q: How does this affect DBAs?
A: The Teradata Database seamlessly manages the routine and time-consuming technical minutiae, allowing DBAs to focus on work that delivers higher value, such as the speedy deployment of applications. Rather than spend many hours creating simple reporting marts and moving data from one place to another, DBAs are given more freedom to create more intricate data models that will solve more complex business questions.
Additionally, the Teradata open solution integrates into all service-oriented architectures [SOAs] and third-party applications such as those from SAP and SAS, along with many others. From a DBA perspective, this integration allows for easier application development, deployment and management, and it enables the delivery of the right information to the right people when they need it.
Q: Data replication can be problematic for users. Which processes typically lead to these issues?
A: In a typical data warehouse system, standardized reports are the lowest-value element delivered. However, this simple workload is often what drives data redundancy because of the solution’s inability to process the data in a timely or easy fashion. Companies soon discover that continued data replication leads to ever-increasing costs and decision latency. Aggravating this replication problem are shrinking batch time windows, which occur because of the additional processing required to summarize and/or move the data.
Q: How does the Teradata system resolve data redundancy?
A: Simple reporting becomes a byproduct of the system’s capabilities as it simultaneously conducts higher-value work in the platform on the same data. This allows business users to drive more value out of the system faster and with more accuracy.
Q: Planned and unplanned system downtime can cause organizations to lose revenue and customers. What features does the Teradata system have in place to avoid this type of catastrophe?
A: Failover, a standard Teradata system feature, is the mission-critical capability of a data warehouse solution to recover from an error and prevent system downtime. A failover solution, known as “cliquing,” is implemented at the base level. This is in addition to several other enhanced mechanisms for handling errors. When a single node fails, a clique, or collection of nodes, will integrate the unfinished workload from the failed node. Because this is an automatic function, the intervention does not require any interaction from the support staff and is offered at no additional charge to the customer.
The Teradata system also supports a deeper, second level of disk-failure protection with a feature known as “fallback.” Because this function offers higher performance in some areas of query processing, fallback goes beyond traditional redundant array of independent disks [RAID].
To add to these integrated capabilities, Teradata supports an industry standard hot-standby node concept, as well as a unique solution known as “dual active,” which provides nearly 100 percent availability projection and disaster recovery in the same application.
Q: IT expresses that it is often difficult to quantify with upper management the benefits of building a data warehouse. When faced with tight capital budgets, how can a new system be justified?
A: If the acquisition cost is the primary consideration for choosing a data warehouse solution, some confusion may arise around the reason for its deployment. A data warehouse is designed to drive business value by analyzing the organization’s data and offering answers to complex business questions. For this reason, returning hard dollars to the bottom line of the corporation should be an ROI requirement.
In this regard, the key factor for creating and deploying a data warehouse is ensuring that the dollars expended on the solution are dwarfed by the return delivered by that solution. The data warehouse performance and capability—whether actual or proven potential—should exceed the price paid. Companies that deploy Teradata systems often find that the capabilities of the data warehouse exceed the original application demands outlined by the end users—by months instead of years, in many cases.
The three primary capabilities of the data warehouse—workload management, ease of use and failover—make a positive ROI easy to achieve. And when an organization’s bottom line is reinforced through the delivery of BI, reaching the level of industry leader becomes an attainable goal.
Photography by Veer
