Intel and Micron’s announcement of a new memory technology, 3D XPoint, will add further momentum to the adoption of in-memory processing of large data sets and extend the scope of real-time analytics to a wider set of business processes.
It is a timely reminder of the compounding effect of hardware and software innovation that continues to push the boundaries of information technology.
Over the last five years, the move to the next generation of in-memory databases, together with falling DRAM prices and the mainstream adoption of solid-state storage (SSD), has enabled a new class of analytics-equipped business applications.
These applications are being used to dramatically speed up business operations and decision making in areas as diverse as marketing optimisation, supply chain forecasting, risk management, and financial close.
It is claimed the new technology is 1,000 times faster than the NAND storage used in SSD, and does not suffer the same degradation problems, offering 1,000 times greater endurance.
Initial costs are likely to be high, estimated to be somewhere between those of SSD and DRAM, but are likely to fall as demand and production increase.
Storage vendors will incorporate the new technology as an additional tier in their infrastructure, using it to cache the “hottest” data in the same way SSD does currently.
An order of magnitude increase in density over NAND will also stimulate new applications in consumer devices.
Faster processing of even larger data sets will accelerate the move to in-memory database technology, and enable real-time analytics to be applied to more complex business processes, with initial examples in areas such as financial services, scientific computing (including medical research), and fraud detection.
The new technology will also increase the rate of progress in machine learning and artificial intelligence.
By Tim Jennings - Research Analyst, Ovum