Oracle is a popular relational database management system (RDBMS). With several online forums and tools available to the database administrator (DBA), deploying and managing Oracle has become easier. However, it is not uncommon to have a working solution hit roadblocks as it scales up. This book presents historical approaches and the shortcomings of guess-based approaches. In addition, it explains how to approach and how to solve specific problems.
The book starts with a very interesting chapter by Alex Gorbachev: “Battle Against Any Guess.” The author uses real-life examples to stress the importance of using analysis (rather than guesswork), and he highlights the difference between being efficient and being effective. The second chapter, “A Partly Cloudy Future,” by Jeremiah Wilton, starts by pointing out the hype surrounding cloud computing. After listing a few providers, he quickly moves on to explain Amazon Web services and running Oracle, including the cost savings associated with it. Chapter 3, “Developing a Performance Methodology,” by Connie Green, Graham Wood, and Uri Shaft, emphasizes the need to design the system for required performance levels. It provides a list of design pitfalls to guard against and it covers the use of diagnostic tools in performance analysis. The next chapter, “The DBA As Designer,” by Melanie Caffrey, covers the DBA’s role in application design. The importance of the DBA as a member of the design and development team is vividly explained, with examples. Niall Litchfield’s chapter, “Running Oracle on Windows,” presents some differences between Windows and Unix/Linux platforms from a DBA perspective. In “Managing SQL Performance,” Karen Morton attempts to highlight the importance of hidden issues--for example, it duly brings out misleading methods for measuring expected performance instead of measuring actual performance. Often, people use custom procedural language/structured query language (PL/SQL) functions in their SQL statements, but then leave the optimization to the default Oracle optimizer, which can result in performance issues. Jože Senegačnik, in chapter 7, “PL/SQL and the CBO,” highlights the importance of using an extensible optimizer to incorporate custom values when using custom PL/SQL functions.
To achieve the desired performance, one needs to understand the methods available for optimization and then choose the appropriate ones. Chapter 8, “Understanding Performance Optimization Methods,” by Charles Hooper, and chapter 9, “Choosing a Performance Optimization Method,” by Randolf Geist, provide a detailed account of this. They cover several optimization methods and discuss how to make the right choice. They also cover as many fronts as possible, from file input/output (I/O) to central processing unit (CPU) performance. Together, these two chapters make up about a quarter of the book--this alone suggests the importance of this topic.
Chapter 10, “Managing the Very Large Database,” by Tim Gorman, covers issues related to designing, partitioning, and backing up as the database scales up. Getting the statistics for performance evaluation and monitoring can be tricky; Jonathan Lewis addresses this topic in “Statistics.” The next chapter, “Troubleshooting Latch Contention,” by Riyaj Shamsudeen, explores some of the internal mechanisms of latches. Since this is a lesser-known topic, the author tries to provide step-by-step instructions for latch contention-related issues. Robyn Sands’ chapter, “Measuring for Robust Performance,” covers tuning and the effect of tuning on performance.The book concludes with two chapters by Pete Finnigan: “User Security” and “Securing Data.” Both are crucial reads given the current cyber-threat-prone environment.
Overall, I recommend this book to readers who handle serious Oracle deployments that need to perform and scale, and still remain secure. The authors do a good job of covering things that are either inconspicuous or often ignored.
