Fast memory and storage architectures for the big data era

The rate of data growth is expected to propel. People and things continue to generate data, some of which are then transmitted, stored and analyzed. Stored data pose a great value to those who have access to them - but only if the data can be accessed and analyzed in a timely manner. For many interactive applications, traditional storage systems built with rotating magnetic media (hard drives) fall short because of their slow data access rate and high energy consumption. Clearly, solid-state memory and storage systems will play the central role in demanding big data analytics applications. Relevant questions are many, such as: How quickly can you detect fraud in on-line transactions? How much memory is needed to hold critical data for fast decision making? How much stored data can one process in a given time to offer relevant in-situ business intelligence? Emerging memory- and storage-centric system architecture innovations will enable cost-effective high-performance data analytics. In this presentation, recent advances in solid-state memory and storage system technologies and R&D challenges in the context of big data processing will be presented. Certain broader driving forces and trends in the memory industry will be outlined as well.