Analytics on the Edge

Abstract

One of the driving forces for edge computing is the ability to run analytics on edge devices. Companies who have spent years creating and improving their algorithms or machine learning models that are used in their backed batch processing workflows are trying to understand how they can leverage that existing intellectual property in new ways and in new areas using the edge. In many cases, the batch processing paradigm imposes intolerable delays throughout the rest of the dependent systems, which motivates companies to look for ways to speed up the path to get the answers they want. The faster the results (results, not just data) are generated, the faster the business can react. So, companies have started introducing edge solutions that place this advanced logic (algorithms and models) as close to the source of the data as possible, which enables decisions to be made in real time, rather than having to wait on decisions until the data has been processed by backed batch processes.