Abstract
In-situ processing, in-storage processing, smart SSD, computational storage… Many names to define the same concept: a closer integration of computing capabilities and data storage in order to reduce data movement leading to better performance and lower power consumption. This is a new trend in storage and computing architectures. How can we define this new type of products: storage with embedded processing, or processing with embedded storage? This talk presents an analysis of the in-storage processing trend, including a comparison of different architectures, the presentation of a computing and storage technologies roadmap, and a list of applications use cases. Learning Objectives: 1. In-storage processing architectures 2. Technology roadmap for integration 3. Application use cases