Print

Send to colleague

E-mail us

Lessons Learned: Teradata's Dual Active Solution

Obtaining flexibility and benefits for any technology environment.

by Imad Birouty

As data warehouses have become more mission-critical for businesses, full-time availability has become a priority.

A few years ago, the only people within a company who needed access to a data warehouse were typically technicians or executives. In today's business environment, however, front-line staff and executives throughout the organization need to access to the information stored within the data warehouse. So 100% uptime has become a necessity for data warehouse environments and true business continuity. Failures, even for a few minutes, can mean lost opportunities for sales, revenue and profit.

Historically, business continuity was focused on protection against unlikely but potentially large events such as fire, flood and other natural disasters. Yet with real-time enterprises, even the smallest interruptions can have serious business consequences. In order to offer true business continuity, a data warehouse system must offer:

	Availability—the ability to prevent outages from occurring or significantly reduce the impact of outages if they do occur
	Recoverability—the ability to rapidly restore systems or data to a normal operational state
	Data protection—the ability to preserve data in the event of a user's error, system failure or local disaster

The Teradata Dual Active Solution—part of Teradata's Business Continuity Solutions—offers the highest levels of availability to applications and users while maintaining a single view of the business to ensure consistency as well as availability of the data.

(click to enlarge)

A dual-active environment has three technical requirements: data synchronization, query routing, and monitoring and control.

Synchronization of data on the two systems is accomplished primarily through dual loading of data or replication. Other synchronization techniques may also be employed, depending on the needs of the customer.

Dual loading of data is generally used when large amounts of data are introduced from an external source.

Replication produces similar results and is primarily used for data that is changed within the warehouse. Regardless of the technique, data changes are monitored and automatically replicated on each system to ensure that anyone accessing either system will receive the same data.

Teradata Query Director determines which system a user or application should be routed to. It employs sophisticated techniques to maximize load balancing between the systems while quickly and transparently routing users and applications to the appropriate system during normal operations and during system outages.

Monitoring and control capabilities give data warehouse administrators a single view of the dual-system environment. From this single console, administrators can control data loads, change Teradata Query Director settings, monitor the operational state of each system, and control planned and unplanned system outages, among a whole host of other capabilities.

Different user environment—similar requirements and benefits
Enterprises of different sizes, different industries and different technology infrastructures are seeing the benefits of the dual-active systems deployment in their operations.

Customer A

A large retailer with more than 2,000 stores needed a system that provided disaster protection for all applications and high availability for critical applications. The retailer elected to use its two existing data centers, which were 800 miles apart, to supply the necessary disaster protection. The company had a complex data warehouse environment, with nightly, weekly and monthly batch extracts from source systems, as well as application changes to data from within the data warehouse. The requirement was to develop a solution that gave the company the high availability and disaster protection it needed while minimizing changes to existing load jobs and control scripts. The company also requested automated monitoring of the dual-system environment for the administrators. For application access, the company requested transparent routing, which employed load balancing, preferred routing and required routing, depending on which user was accessing the system. Only the portion of the data that was deemed mission-critical was to be duplicated on the second system. This resulted in two systems of unequal size and different hardware generations: one with 100% of the data and a second with a subset of the data.

Customer B

A large multinational bank with thousands of users and hundreds of applications across diversified businesses in retail, commercial, mortgage, consumer lending and insurance operations needed a system that offered high availability for critical applications, disaster protection and optimization of resources for data enhancement. The bank requested a solution that used their two existing data centers, which were located 15 miles apart, to give them enough physical separation for disaster protection. Their requirement for building this solution was to have automated monitoring of the dual system environment with no major rewrites of load or data enhancement jobs. Further, the routing—including load balancing—had to be transparent while maintaining high throughput. The customer's choice was to duplicate 100% of their data on both systems, resulting in two systems of different generation hardware that had the same data model, the same data and nearly equal performance.

Customer C

A mid-sized retail bank with more than 500 different locations, each location running its own system and all systems feeding back into a centralized database for access by multiple applications. The bank required high availability for all its applications and a solution that resulted in no data loss at any time. They also required that all ETL (extract, transform and load) operations be done through Informatica, which was their standard. Like the previous two customers, they requested transparent routing for application access, which employed load balancing, preferred routing and required routing. This customer chose to employ two new systems of identical size and capacity to provide 100% duplication of their warehouse environment.

Each of these customers is quite different from the others in terms of industry, amount of data, size of system, data center environment, location, etc., and in each of these three cases, the Teradata Dual Active Solution was architected to meet 100% of the requirements set forth by the customer. All environments have built-in capabilities to handle planned and unplanned shutdown of either system, re-routing of users and applications following a shutdown, and recovery of the system back into production following an outage. These three customers have their dual-active environments operational and in production today.

Even though the exact needs and technology environments of each were quite different, their overall interest was the same: high availability and disaster protection delivered transparently to their users.

The Teradata Dual Active Solution architecture can benefit many companies, wherever true business continuity is needed. T

Imad Birouty is the marketing program manager for Teradata's Business Continuity Solutions.

© Teradata Magazine-March 2006

RELATED LINKS:
Implementing the Teradata Dual Active Solution Teradata Dual Active Solution Teradata Services: Business Continuity

back to top