Sep 11 2019
Artificial Intelligence (AI) is meant to make life better and simplify the world but, without Computational Storage, it can actually be like adding new rooms and walls with an architecture that is never-ending. You might find yourself in a maze similar to the Winchester Mystery House that didn’t have any master building plan as evident by the doors that lead to nowhere and stairs that end at a solid wall. Had the blueprint considered options like Computational Storage then it would help simplify the floor plan and piece together the house with purpose and direction by making it simple, smarter and more efficient.
It’s no surprise to anyone when we say that AI has taken the technology industry by storm, but it has also meant, that the data AI churns on is an unstoppable tsunami of information. The crux of the matter is that enterprises deploying workloads, at the edge or even the core, to support data-intensive use cases like AI require a more effective and effecient way to process and analyze data to get valuable and needed results. As such, the computer storage industry is no longer just about storage and recovery but about intelligently analyzing data in real time. As AI and other data-intensive deployments increase, data today still needs to be moved over increasingly longer distances, which causes network bottlenecks and delays analytic results.
Next week NGD Systems will be in attendance at AI Hardware Summit, a conference dedicated solely to the ecosystem developing hardware accelerators for neural networks and computer vision. The 2019 summit will discuss emerging AI hardware technology that can solve the pain point of space limitations and processing powers the enormous amount of data AI workloads has created. Computational Storage which precludes the need to move the data around – can be especially valuable in workloads such as AI, big data analytics and machine learning because it enables data relevant to the query to be pinpointed on the storage media itself so that only specific data sets get transported back to the host, according to Storage Switzerland.
Additionally, SearchStorage explains that “The complexity of the workloads plus the volume of data required to feed deep-learning training creates a challenging performance environment. … “
Upshot, Computational Storage has emerged as a new trend which can deftly organize raw information into meaningful results – allowing an organization to ingest as many bits as possible and churns out just the right information on command and in real time at the storage level instead of forcing more work on an overburdened Central Processing Unit (CPU).
To learn more about Computational Storage, check out our recent blog that shares the definition and more! https://www.ngdsystems.com/page/Computational-Storage:-What-It-Is-and-Why-You-Need-It
Look for our logo at the Computer History Museum in Mountain View on September 17 – 18! We look forward to further discussing Computational Storage with you.