Paper Title: Memory Requirements for Edge Computing in IoT applications

Paper Abstract: IoT devices often generate huge network traffic, while some require real-time control. If all traffic is sent to the cloud, then the sheer number of connections can create an overwhelming strain on network and compute resources. A simple solution is to use intelligent edge devices (so-called edge computing or "autonomous edge" to reduce network traffic and latency. The question then comes down to what can be handled efficiently at the local level and how much must still be sent to the cloud for processing. Obviously, there are tradeoffs here among local complexity, cost, and network capacity. Memory is a critical factor since it will be a large part of node cost and will have a major effect on node performance and power consumption.

Paper Author: Nabil Damouny, Chief Strategist, Autonomous Edge

Author Bio: Nabil Damouny is Founder and Chief Strategist at Autonomous Edge, a consulting firm focused on developing strategies for communications including IoT devices and cloud edge devices. He has over 30 years experience in the semiconductor and networking industries, including executive positions at Intel and network processor maker Netronome. He has been a leader in developing standards for software-defined networking (SDN) and network functions virtualization (NFV) and was Vice Chair, Market Education at ONF (Open Networking Foundation, focused on SDN). He was a co-founder of the network processor firm Basis Communications, which was acquired by Intel. He earned an MS in electrical and computer engineering from UC Santa Barbara. He holds 3 patents in networking and computer architecture, has several published articles, and has presented at many trade and technical events.