The English writer Samuel Johnson once remarked, "The trouble with life isn't that there is no answer, it's that there are so many answers." This statement is entirely appropriate for the term "database" which represents a wide spectrum of use cases, with On-Line Transaction Processing (OLTP), Reporting, On-Line Analytical Processing (OLAP) and Analytics being the most common ones. With such diverse and varied sets of requirements, what is the best method to optimize "database performance"? This paper explores the characteristics of the predominant database implementations and offers suggestions on building a highly efficient and performant infrastructure to support this application category.
Phil Muncaster reports on China and beyond