The landscape of today’s business is shaped by the mountains of data being produced, with rapid growth in the volume, variety, and velocity of data due to the explosion of smart devices, mobile applications, cloud computing, and social media. Much of this growth has been in unstructured data; however, by 2020, internet business transactions—business-to-business and business-to-consumer—are predicted to reach 450 billion per day. Smart organizations are seeking innovative ways to turn this explosion of data, called big data, into actionable insights. This book explores the attributes that make IBM’s DB2 11 for z/OS the ideal database for big data and business-critical analytics in the new era of computing. It is packed with rich information about features and business benefits of the relational database management system’s newest software release, including even more out-of-the-box CPU savings, enhanced resiliency, capabilities to excel at business-critical analytics, and simpler, faster upgrades for quicker return on investment (ROI). Find out why DB2 is the database of choice for big data and analytics by top businesses around the world.
|Sold by:||Barnes & Noble|
|File size:||3 MB|
About the Author
Cristian Molaro is an independent DB2 specialist and an IBM gold consultant. He was recognized as an IBM Champion in 2009, 2010, and 2011. He is the coauthor of DB2 9 for z/OS: Distributed Functions and DB2 10 for z/OS: The Smarter, Faster Way to Upgrade. Surekha Parekh is IBM’s worldwide marketing program director for DB2 for z/OS and leads the social media strategy for information management on System z. She is the coauthor of DB2 10 for z/OS: Cost Savings . . . Right Out of the Box, DB2 10 for z/OS: The Smarter, Faster Way to Upgrade, and The Business Value of DB2 for z/OS: IBM DB2 Analytics Accelerator and Optimizer. Terry Purcell is a senior technical staff member with the IBM Silicon Valley Lab, where he is lead designer for the DB2 for z/OS Optimizer. He is the coauthor of The Business Value of DB2 for z/OS: IBM DB2 Analytics Accelerator and Optimizer. He lives in Springfield, Illinois. Julian Stuhler is a principal consultant with Triton Consulting, a UK-based company specializing in the provision of DB2 consultancy, education, software, and managed services to clients throughout Europe. He is the coauthor of DB2 10 for z/OS: Cost Savings . . . Right Out of the Box.
Read an Excerpt
DB2 11: The Database for Big Data and Analytics
By Cristian Molaro, Surekha Parekh, Terry Purcell, Julian Stuhler, First Edition
MC Press Online, LLCCopyright © 2013 IBM
All rights reserved.
DB2 11 for z/OS: Unmatched Efficiency for Big Data and Analytics
by Julian Stuhler
Few IT professionals can have missed the big data phenomenon that has manifested itself in recent years. Industry publications and IT analysts have devoted a huge percentage of their output to the subject (creating a big data challenge all their own in the process). There can be little doubt that the advent of new technologies and methods of customer and business interaction have created unique challenges for organizations wishing to create actionable insight from very large amounts of unstructured data. Innovative tools and techniques have been developed to cope with these "big data" challenges (and indeed some of them are discussed in this paper, in the "Hadoop and Big Data Support" section).
However, beyond this somewhat narrow definition of big data, many organizations have been dealing with the challenges of processing, maintaining, and analyzing ever -increasing amounts of more traditionally structured data for many years. The inherent scalability and resilience of IBM® DB2® for z/OS® and the underlying System z® platform have proven to be a compelling combination for such applications, and IBM continues to invest in extending DB2's capabilities with each new release.
From transparent archiving to greater in-memory scalability through the use of 2 GB page frames, DB2 11 for z/OS, the latest release of IBM's flagship database, contains many new features specifically designed to help customers to address the challenges of managing traditional big data. A wealth of material exists on the technical changes within DB2 11, but finding descriptions of how those new features will improve your business results can be a challenge. The main body of this paper provides a high-level overview of the major new features from an IT executive's perspective, with emphasis on the underlying business value that DB2 11 can deliver.
This is the fourth paper in this series, with previous editions highlighting the business value offered by DB2 for z/OS V8.1, DB2 9 for z/OS, and DB2 10 for z/OS:
DB2 for z/OS 8.1: Driving Business Value (J. Stuhler, Triton Consulting, 2004)
DB2 9 for z/OS: Data on Demand (J. Stuhler, Triton Consulting, 2007)
DB2 10 for z/OS: A Smarter Database for a Smarter Planet (J. Stuhler, Triton Consulting, 2010)
NOTE:Throughout the remainder of this document, all references to "DB2 9," "DB2 10," and "DB2 11" refer to the relevant release of IBM DB2 for z/OS.
DB2 11 for z/OS: The Database for Big Data and Analytics
In this section, we take a detailed look at the major features of DB2 11 for z/OS and see how many of IBM's most innovative enterprise customers plan to use them to deliver an enhanced IT service to the business. Many of these enhancements can deliver benefits "out of the box," with little or no effort required to begin exploiting them, reducing the time-to-value for a DB2 11 upgrade. See "DB2 11 New Features by Implementation Effort" (opposite) for a breakdown of the effort required to exploit each new feature.
This section is organized around the key DB2 11 themes:
Efficiency. Reducing cost and improving productivity
Resilience. Improving availability and data security
Business analytics. Enhanced query and reporting
Even in the most favorable economic climate, businesses need to control costs and increase efficiency to improve their bottom line. In today's increasingly challenging business environment, this continues to be a key factor for the survival and success of enterprises of all sizes.
This section examines the major DB2 11 enhancements that are aimed at delivering the highest efficiency for core IT systems that rely on DB2, a key design objective for the new release. These features can help reduce ongoing operational costs, improve developer and DBA productivity, and enhance customer experience by increasing performance and delivering a more responsive application.
Most DB2 for z/OS customers operate on a CPU usage-based charging model, so any increases or decreases in the amount of CPU required to run DB2 applications can have a direct and very significant impact on overall operational costs.
Traditionally, IBM has tried to limit the additional CPU cost of adding new functionality into each release, keeping the net CPU impact below 5 percent. The move to a 64-bit computing platform in DB2 for z/OS Version 8 was an exception to this rule and introduced some significant processing overheads that resulted in many customers experiencing net CPU increases of 5 to 10 percent following the upgrade.
DB2 9 for z/OS helped to redress the balance somewhat by delivering modest CPU improvements for many large customers, but the advent of DB2 10 completely changed the picture. IBM delivered the most aggressive performance improvements of any DB2 release in the past 20 years, with many customers seeing net CPU savings of 5 to 10 percent or more in their traditional DB2 online transaction processing (OLTP) workload without any application changes being required. Unsurprisingly, these savings proved to be very popular and are consistently quoted as being one of the major reasons for customers to upgrade to DB2 10.
IBM has further developed the CPU reduction theme within DB2 11, with initial savings of up to 5 percent expected for customers running simple OLTP workloads. Significantly higher savings are possible for complex OLTP and query workloads, as discussed below. Because these improvements are due to internal DB2 code optimization, they are available in DB2 11 Conversion Mode, without the need for any application changes. Additional CPU savings are possible once customers begin to use some of the other DB2 11 enhancements that require application change, as described elsewhere in this section.
Some workloads will benefit more than others from the performance enhancements offered by DB2 11. Figure 1 breaks down the anticipated CPU savings by workload type.
The most significant benefits are expected to be seen within query workloads. Complex reporting queries can see up to 25 percent savings for uncompressed tables and up to 40 percent for queries on compressed tables. Reporting queries with heavy sort processing may also see additional DB2 CPU savings.
Traditional OLTP workloads are also likely to benefit from the efficiency enhancements in DB2 11. Savings of up to 5 percent are expected for simple OLTP, with reductions of up to 10 percent for more complex transactions. Finally, update-intensive batch workloads may enjoy CPU reductions of 5 to 15 percent.
Figures 2 and 3 depict some actual observed CPU reductions for sample workloads, run as part of IBM's internal performance testing for the new release. These figures are broadly in line with the high-level expectations detailed above.
The overall out-of-the-box CPU savings within DB2 11 are expected to be one of the major factors supporting the business case for upgrading to the new release.
In August 2012, IBM announced the latest-generation IBM zEnterprise® EC12 (zEC12) enterprise servers, with up to 101 configurable processors per server, each running at an industry-leading 5.5 GHz. In addition to an impressive list of general performance and capacity improvements over the previous-generation z196 enterprise servers, the zEC12 models include a number of features that DB2 11 will specifically exploit.
2 GB page frames. DB2 10 for z/OS introduced support for 1 MB "large page frames," an enhancement designed to reduce processing overheads for very large DB2 buffer pools by letting z/OS manage the underlying storage in fewer 1 MB pieces rather than many more 4 KB pieces (Figure 4).
Many customers with larger DB2 buffer pools were able to achieve CPU savings of up to 4 percent by exploiting this capability. However, as memory prices fall and workloads increase, DB2 buffer pools continue to increase in size, and the overheads of managing even the larger 1 MB page frames start to become significant.
In recognition of these trends, when running on an zEC12 server DB2 11 will support even larger 2 GB page frames, each of which will map onto more than half a million 4 KB pages (Figure 5).
Those customers using very large DB2 buffer pools will see further CPU reductions by moving to 2 GB page frames. Other sites may not have sufficiently large pools for 1 MB page frames to be a significant limitation today, but that situation will undoubtedly change in the future as buffer pool sizes continue to grow. By moving early to support 2 GB page frames, IBM has recognized and eliminated an important future scalability issue.
DB2 code using large page frames. As discussed in the previous section, DB2 10 and DB2 11 have exploited 1 MB and 2 GB large page frames to allow more efficient handling of large buffer pools. However, despite the extensive use of large memory objects in the past few releases of DB2, the storage used for DB2 code (as opposed to the data held in buffer pools) remained backed by standard 4 KB page frames.
DB2 11 is able to utilize large page frames for DB2 code objects and log output buffers, in addition to buffer pools. This enhancement reduces the z/OS overheads associated with DB2 code objects, lowering CPU consumption and operational costs. (Support for running DB2 code in large page frames requires z/OS 2.1.)
Many new releases of DB2 introduce enhancements or new features that require application and/or SQL code to be changed. These include additional SQL reserved words, changes to DB2 behavior or processing and even changes to SQL return codes. Although IBM tries to minimize these "incompatible changes," they cannot always be avoided. They may be required in order to ensure that DB2 adheres to evolving SQL standards, to support new functionality, or perhaps to address an earlier defect in the DB2 code.
A major part of planning for a new release is to analyze the impact of these incompatible changes and arrange for the necessary amendments to be made to DB2 application code so it will continue to work as designed under the new release. This situation poses some challenges for DB2 customers:
Analysis of the impact of incompatible changes can be difficult, time consuming, and error-prone. Missing one or more of the required changes may result in application outages when DB2 is upgraded (or worse, the application may continue to work but return unexpected results).
Finding the necessary resources to undertake any required remedial work (and scheduling the associated change slots) can be expensive and require significant elapsed time. All of the changes within a given subsystem or data sharing group must be completed before the upgrade can commence, so a lack of resources within a single application team could impact the upgrade schedule for the entire environment.
Figure 6 depicts these challenges.
To address these issues and allow customers to upgrade their DB2 systems with less effort and risk, IBM has introduced some new capabilities in DB2 11 for z/OS that remove the hard dependency on all remedial work being conducted before a version upgrade and allow the impact of incompatible changes to be more easily assessed. Figure 7 summarizes these enhancements.
When upgrading to DB2 11, customers will be able to defer some or all of the remedial work for incompatible SQL DML and XML changes and allow the DBA or developer to request that DB2 behaves the same as it did for DB2 10 on an application-by-application basis. Although the remedial work will still need to be done at some point, DBAs and developers are now free to schedule it a later date and in a more manageable, staged fashion that conforms to the requirements of the business (e.g., as part of a regular application release). In the meantime, other applications can benefit from the enhancements in the new release.
Furthermore, IBM has provided additional trace data in DB2 11 that can identify applications using incompatible SQL and XML statements after the version upgrade has been implemented. This will enable DBAs and developers to identify applications requiring remedial work much more efficiently, and with less risk of some being accidentally missed. Because the intention of this feature is to allow more manageable implementation of remedial work, not to defer it work indefinitely, this capability is limited in the number of previously supported releases. In DB2 11, this feature provides backwards compatibility only for DB2 10. Beyond DB2 11, compatibility for up two previous releases will be provided. This means that the release following DB2 11 will support both DB2 10 and DB2 11 compatibility, thereby allowing plenty of time for any remedial work to be undertaken.
By breaking the hard dependency on performing all remedial work prior to an upgrade and providing valuable tools to assist with the identification of that work, the DB2 Application Compatibility feature addresses many of the issues associated with handling incompatible changes in each new DB2 release. This capability should be a huge benefit to customers struggling to line up the necessary application development/DBA resources to address incompatible changes prior to DB2 11 implementation.
A common requirement for any IT application is to be able to archive old or less frequently accessed data. Regulatory restrictions may require data to be retained for many years, but access frequency tends to drop off dramatically as the data ages). Moving older data to a separate archive can reduce the cost of retrieving and maintaining more frequently used data and allow slower but much less expensive storage devices to be used.
Unfortunately, archiving is usually one of the last areas to be considered and developed for a new application, and it is therefore common to see it deferred until later code releases (or bypassed completely) if time and/or funding is scarce. Even when it is properly implemented, many hundreds of person-hours can be spent in implementing and testing the necessary logic to allow older data to be placed automatically in the archive store while ensuring that the application retains access to it when required.
DB2 11 introduces some new features to simplify application development for archiving data, as well as improve consistency between applications and reduce the amount of time required for testing. When defining the operational DB2 table, the DBA also defines an identical archive table and connects the two via an ALTER TABLE ... ENABLE ARCHIVE statement. Any subsequent changes to the operational table (e.g., adding a column) will automatically be made to the archive table so they remain in step. If required, the archive table can be placed on older, cheaper disk devices, with the more frequently accessed operational data residing on faster storage. Once an archiving relationship has been defined, DB2 can automatically and transparently handle archiving and retrieval of data from the operational and archive tables. The new DB2 11 global variable support (discussed further in the "Global Variables" section) is used to provide simple application-level switches to enable or disable arching functionality at run time, as shown in Figure 8. For static SQL, DB2 automatically prepares two access path strategies: one for use when archiving is enabled and another for use when it is not.
Once an archiving relationship has been defined, DB2 can automatically and transparently handle archiving and retrieval of data from the operational and archive tables. The new DB2 11 global variable support (discussed further in the "Global Variables" section) is used to provide simple application-level switches to enable or disable arching functionality at run time, as shown in Figure 8. For static SQL, DB2 automatically prepares two access path strategies: one for use when archiving is enabled and another for use when it is not.
This approach is very flexible, providing automatic archiving and transparent access to archived data while also retaining the ability to disable that functionality via a simple SQL statement (or BIND option) when performance is critical and/or archiving functionality is not required.
Note that Version 3 of the IBM DB2 Analytics Accelerator product also offers some interesting options for handling archive data. For further details, see the "IBM DB2 Analytics Accelerator Enhancements" section.
Overall, the new transparent archive feature promises to significantly reduce the cost of designing, developing, and testing data archive processes for DB2 applications. While it will be of limited value for those applications that have already implemented such functionality manually, it could reduce developer/DBA effort by hundreds of person -hours for newly developed applications (or existing applications that did not originally implement an archiving strategy).
Solutions developed using this feature will also benefit from the ability to dynamically enable and disable access to archive tables at run time, ensuring that no performance overhead exists for processes that require access to only the current operational data.
Excerpted from DB2 11: The Database for Big Data and Analytics by Cristian Molaro, Surekha Parekh, Terry Purcell, Julian Stuhler, First Edition. Copyright © 2013 IBM. Excerpted by permission of MC Press Online, LLC.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.
Table of Contents
ContentsAbout the Authors,
Introduction by Surekha Parekh,
DB2 11 for z/OS: Unmatched Efficiency for Big Data and Analytics by Julian Stuhler, Triton Consulting,
Improved Query Performance in DB2 11 for z/OS by Terry Purcell,
IBM DB2 Utilities and Tools with DB2 11 for z/OS by Haakon Roberts,
How DB2 for z/OS Can Help Reduce Total Cost of Ownership by Cristian Molaro,