In-memory database systems worth thinking about, McKnight saysDate: Aug 20, 2014
One of the instructors at the 2014 TDWI World Conference in Boston was William McKnight, president of McKnight Consulting Group, which offers strategic and technical consulting services for master data management, big data, data warehousing and business intelligence (BI) initiatives. At the educational conference, which was colocated with the TDWI Executive Summit, McKnight spoke about such topics as BI project management and calculating ROI projections for BI and information management projects. He also discussed the potential uses of in-memory database systems and analytics tools in a video interview with SearchDataManagement.
In the past, in-memory databases were standalone technologies that targeted niche applications. But the leading database vendors are now incorporating in-memory functionality in their flagship relational software -- for example, Oracle released an Oracle Database In-Memory add-on in July, and Microsoft included an in-memory engine for transaction processing in the SQL Server 2014 update it launched in April. In the video, McKnight suggests that organizations should consider investing in in-memory technology for applications that require the highest levels of performance. In general, he said, performance is the top requirement for the success of processing workloads. Using an in-memory database can enable an organization to do deeper and more iterative analytics than is often feasible with conventional disk-based processing, he said.
And nowadays, McKnight added, "almost every organization that I've come across" has workloads with performance requirements that make in-memory processing a good fit. For data science applications in particular, it's important to give data analysts a platform that provides as much processing power as possible, he noted. In-memory database systems do still cost more compared to disk-based ones -- but he said that decreases in memory prices have made the in-memory approach more viable from a price-performance perspective.
Watch this two-minute video, hosted by SearchDataManagement Executive Editor Craig Stedman, to hear more of what McKnight had to say about deploying in-memory databases in business organizations.