Hyperscale data is the next evolution of big data. With the massive increase in data generated in fields such as healthcare; securely storing the data & performing meaningful analytics have become a serious issue. Computational storage aims to solve this problem with a ridiculously simple approach: treat storage as a smart, computationally independent unit.
Computational storage is the notion of a traditional storage unit, paired with a general purpose CPU core. The goal of these devices is to perform most of the storage related tasks onto these brilliant smart storage units. Computational storage is an important piece of technology to enable hyperscale data analytics.
Computational storage brings incredible performance acceleration for storage bound tasks such as search, indexing & analyzing feature vectors such as those associated with genomics data.
I will be discussing the basics of computational storage, how to write applications to leverage computational storage & use cases that perform the best when optimized for infrastructure that leverages computational storage.