Networking for Supercomputing
About the Speaker
Senior Network Engineer at Oak Ridge National Laboratory
Daniel Pelfrey is a Senior Network Engineer at Oak Ridge National Laboratory in the Supercomputing Division. Daniel's Information Technology career began at an Internet startup company in 1995. He was there from customer number 0, and all the way up to 75,000, and from location #1 to over 35+ locations in three states. This eventually led to a nationwide ISP, and ultimately to ORNL where he is on the network team that takes care of the supercomputing network. Daniel's first round of formal educational experience was in electrical and computing engineering, and he has a Bachelor of Science in Information Systems at Strayer University. His hobbies include PC gaming-especially the game Factorio, retro gaming, electronics, science, anything space related and programming.
Presentation Resources
Not to be confused with the Supercomputing conference, this presentation is an overview of a high performance computing center network for a leadership class computing facility. This presentation will cover network requirements and challenges of designing, implementing, operating, and monitoring a high performance computing network and will compare and contrast different network environments. How well do you know the limit of every piece of network gear in your network? Where are the bottlenecks in your topology with respect to these limits? How do you monitor for these limits? What are your interoperability challenges in a heterogeneous network environment? What are some of the design challenges to ensure the network meets performance and scalability? This presentation will cover these and many more questions. Supercomputing networks is very much building airplanes in the sky.