Cognitive Cloud Networking - Architectures and Applications. 3 credits, 3 contact hours

Catalog description

Prerequisite: CS 656 . If anything can characterize today's networking environment is its newly found significance in delivering cloud applications at scale. Everything has become a cloud - even the formerly-rigid wireless proprietary nodes, known as base stations, are broken up to independently developed network micro-services running on cloud-native runtimes supported by commodity compute and storage. Networking has become a software business and complete network disaggregation is the end game. In this new world, network operators are increasingly becoming content providers and they abandon R&D-heavy network disaggregation initiatives run by the Linux Foundation and increasingly trust public cloud providers to run their networks and take full advantage of the public cloud completely abstracted set of APIs to build applications faster than they could ever do before. Drawing from these trends, the course aims on one hand to provide a foundation of the so called multi-cloud network environment and on the other, to provide the student with the tools and knowledge required to understand how these multi-cloud networks are operated / managed via a set of data-driven distributed applications, the Cognitive Network Apps, that bring Machine Learning / AI and big network data analytics together. We treat the most important networking architectures and technologies that empower today’s cloud provider networks at the core / edge such as CDN and 5G and learn how self-healing, self-optimization and self-configuration applications provide the reliability, resiliency and performance metrics needed by cloud applications such as Netflix, Zoom and many others.