University of Minnesota
Software Engineering Center

You are here

ML/AI and Observability Impacts in ‘Two Speed’ Implementations

Date of Presentation: 
Thursday, January 16, 2020
Presented By: 
As AI/ML moves towards a front line, operational need, legacy analytical architectures need to re-think acquisition, management and, above all, observability regarding the acquisition and management of these mission critical data elements and structures. Adaption of a two-speed architecture is a critical enabler for this, providing the framework and mechanisms to offer a traditional OLTP/RDBMS path and a separate, high speed and event driven path for solutions that persist data in different formats/platforms. These streams not only enable these types of solutions, but offer the ability to aggregate and observe data in nontraditional methods such as search, graph and document based solutions.