Accelerate Query Workloads to Streamline Business Operations

Gain successful outcomes that are driven by implementing the DB2 Analytics Accelerator

zAnalytics Technical Leader - MEA, IBM

Business environments today depend on rapid analysis of data in their extensive repositories to meet strategic objectives and remain ahead of the competition. As a result, the key to survival for many organizations is the capability to leverage information rapidly and be responsive to dynamic situations and fluid conditions.

The IBM® DB2® Analytics Accelerator software for the IBM z/OS® high-performance appliance in DB2 for z/OS database deployments enables the DB2 database to offload data-intensive, complex, static and dynamic DB2 queries for data warehousing, business intelligence, and analytics workloads. And it helps perform these workload offloads without requiring any application changes. The DB2 Analytics Accelerator deeply integrates IBM Netezza® technology with the IBM zEnterprise® platform at the kernel level. It enables queries to be executed significantly faster than was previously possible, while helping optimize processor utilization in DB2 for z/OS.

To understand the opportunities the DB2 Analytics Accelerator offers organizations, think of this technology as being conceptually equivalent to a hybrid automobile. A hybrid vehicle has a standard end-user interface that includes a steering wheel, brake pedal, accelerator pedal, dashboard instrumentation, and so on. In addition, it may at any given time be powered by gasoline or electrical power. The hybrid vehicle’s capability to switch between power sources automatically helps optimize fuel economy without requiring manual driver intervention or a change in the standard vehicle application programming interfaces (APIs).

Enhancing query workload response

The DB2 Analytics Accelerator can be highly useful in many organizational scenarios. It is designed to deliver dramatic enhancement in response times for unpredictable, complex, and long-running dynamic and static query workloads. It also helps meet service-level agreements (SLAs) and can shorten batching windows by offloading complex query workloads, which it accomplishes by helping improve response times for processor-intensive queries.

In addition, workloads that previously were not considered for mainframe processing or queries governed or shunted in DB2—such as ad hoc queries whose performance characteristics were typically unknown at runtime—can now be run using the DB2 Analytics Accelerator. DB2 can control where to run these queries or force these types of queries to the DB2 Analytics Accelerator to help prevent additional DB2 database consumption.

By offloading resource-intensive queries and the associated processing onto the DB2 Analytics Accelerator, organizations can reduce million service units (MSU) consumption. It also enables cost-effective storage, management, and processing of historical data using a near-line storage solution. The time necessary to perform general tuning and administration tasks supports cost-effective performance enhancement for resource-intensive DB2 for z/OS workloads.

Moreover, DB2 Analytics Accelerator deployments can dramatically reduce hardware and software procurements for data warehousing and analytics and provide cost-effective data movement, transformation, landing, storage, and systems maintenance. And in addition to the benefits of integrated, operational business intelligence (BI), the DB2 Analytics Accelerator can empower organizations to consolidate disparate data to existing zEnterprise platforms.

The DB2 Analytics Accelerator also helps increase organizational agility by providing the capability to more rapidly respond to queries with immediate, accurate information from which business users can derive strategic insights. The zEnterprise platform is designed to consolidate reporting where the majority of the data being analyzed exists, while maintaining its security and reliability.

Applying business intelligence and reporting

DB2 Analytics Accelerator technology can be a good fit in organizations that want to extend the use of operational data for business analysis, embed operational analytics in other applications, or view daily BI reporting. Reporting or operational BI initiatives can be run on zEnterprise using the DB2 Analytics Accelerator to enable line-of-business users and business processes to glean insight from the information.

Either the WHERE, GROUP BY, or ORDER BY aggregate function can be used with long-running DB2 for z/OS queries, such as those queries that are less than five seconds. For example, these queries may be run from a BI environment that provides important business information. Or, they may be scheduled in overnight batch processes so that they do not affect business users during daytime work hours. However, overnight schedules may mean that information is not available in a timely manner or that the full potential of having this information for other business processes is not realized.

Another good fit for the DB2 Analytics Accelerator is in organizations that may have existing but forgotten or abandoned queries, which may have been shelved because of performance problems. Some of these queries may have already been through exhaustive tuning efforts without success. Complex analytical and ad hoc queries that may not currently run in DB2 for z/OS may be able to provide significant, successful outcomes to line-of-business users who may not realize the impact of their ad hoc queries.

The DB2 Analytics Accelerator can be used to consolidate data and analytics workloads to a single, secure data environment to help reduce costs for maintenance, extract-transform-load (ETL) processes, administration, licensing, and so on. The high-performance storage saver feature enables organizations to move tables not used frequently or table partitions to the DB2 Analytics Accelerator and remove the data from DB2 for z/OS. All the data can still be maintained in the DB2 directory, and all the queries that target that data are directed only to the DB2 Analytics Accelerator. This feature helps store, manage, and process historical data cost-effectively while helping to achieve improved response times for queries against the historical data.

Some organizations may also benefit from using the DB2 Analytics Accelerator to analyze non-DB2 data such as IBM Information Management System (IBM IMS™) data, virtual storage access method (VSAM) data, or data in flat files. Organizations also have the option of loading non-DB2 data into the DB2 Analytics Accelerator when a business user has a need to analyze this kind of data.

Benefitting financial workloads

Application teams and organizations, particularly those organizations in the financial industry, can implement the DB2 Analytics Accelerator to achieve successful outcomes. In particular, they can offload processor-intensive query workloads from DB2 for z/OS to the DB2 Analytics Accelerator to help reduce monthly license charge (MLC) and independent software vendor (ISV) costs. In addition, by meeting SLAs and/or minimizing batch duration windows, they can enhance query response times for their line-of-business users. The following workload examples may be well suited for the DB2 Analytics Accelerator in financial industry scenarios:

  • Card issuance analysis
  • Credit underwriting efficiency and effectiveness
  • Customer churn reduction
  • Debit card reporting and card fraud reduction
  • Financial advisory services effectiveness
  • General month-end processing and ad hoc querying capabilities
  • Historical data analysis for understanding trends—analysis of historical data that may be archived
  • Increased cross-sell
  • Securities portfolio risk management
  • Web interaction analysis

Organizations of all stripes can deploy the DB2 Analytics Accelerator to help drive advanced applications and analysis cost-effectively to promote revenue growth, heighten customer satisfaction, enhance security, and bolster risk management. It also advances operational efficiency by helping organizations reduce the need to procure and maintain new data stores for analytics and to engage in complex tuning of workloads.

Please share any thoughts or questions in the comments.