In-memory databases used to be the territory for niche technology vendors and equally niche applications. But things have changed. Vendors of all database stripes -- SQL, NoSQL, NewSQL -- now offer in-memory technology, some as standalone products and others as add-ons to disk-based database management systems. That includes relational database market leaders Oracle, IBM and Microsoft as well as business applications bigwig SAP with its HANA system.
In an interview with SearchDataManagement's Jack Vaughan, business intelligence and data management consultant William McKnight said that as the price of RAM declines, "memory in a lot of ways is becoming the new disk." There are good reasons for that. Keeping analytical data in memory can sharply reduce query response times and enable end users to run deeper analyses, McKnight said. "When you can do that," he added, "you're hopefully producing a better business."
IDC analyst Mike Rosen wrote in a January 2014 report that in-memory databases could provide "transformational performance improvements" in both operational and analytical applications. But in a video posted on YouTube the following month, Rosen said the heavily hyped technology also has "the potential to be the next failed silver bullet from IT." Challenges abound, he cautioned, include a possible need for costly new hardware, data migration issues and the proliferation of data silos that make it hard to do real-time analytics effectively. "Before you make an investment in this technology," Rosen said, "it's important to understand what problems in-memory can solve and whether your organization can make the changes necessary to benefit from such a solution."
SearchDataManagement and sister sites SearchOracle and SearchSQLServer have recently published a variety of content that explores in-memory database trends and offers advice to help you get started on deciding whether the technology is right for your company. In a full version of our Q&A with McKnight, he more deeply dissects the mainstreaming of in-memory technology. In another article, consultant David Loshin assesses applications -- and types of organizations -- that are a good fit for in-memory databases. Another story looks at recent in-memory software releases by the top relational database vendors. We examine those moves more closely in separate stories that focus on Microsoft's In-Memory OLTP option for SQL Server 2014, IBM's BLU Acceleration technology for DB2 and Oracle's in-memory add-on for its 12c database. If you do move forward with an implementation of in-memory database software, hopefully the deployment will be a memorable one -- for all the right reasons.