Platform Approach for SDN Predictive Management Using AI and ML
AI and ML are the latest hot buzz words not only from the top technology companies but also from businesses large or small around the world. AT&T has promoted the cause of AI/ML over the last 20 years …
| Talk Title | Platform Approach for SDN Predictive Management Using AI and ML |
| Speakers | David Lu (Vice President, SDN Platform & Systems, AT&T), George Sultana (Assistant Vice President, AT&T) |
| Conference | Open Networking Summit North America |
| Conf Tag | |
| Location | Los Angeles, CA, USA |
| Date | Mar 26-30, 2018 |
| URL | Talk Page |
| Slides | Talk Slides |
| Video | |
AI and ML are the latest hot buzz words not only from the top technology companies but also from businesses large or small around the world. AT&T has promoted the cause of AI/ML over the last 20 years in predictive network management. AT&T is upping its game in AI/ML analytics by partnering with the industry and open source communities with new AI/ML platforms like ACUMOS, as well as by using ONAP as a platform to provide a framework for easy integration. We recognized early on that one of the challenges in applying new ML is the conflict of aggressive prediction from AI/ML vs. network reliability. AT&T has implemented a “sandbox” to enable experts in operation to trial AI/ML rules with real time production data in an experimental environment first, then deploy them into production without code changes. This talk will describe the strategy and implementation with a quick video demo.
