Abstracts

Educational Session Tutorials

Teaching Big Data and Open Source Software on Chameleon Cloud

Gregor von Laszewski, Indiana University

CyberLearning Education Platform

Paul Rad, UTSA

This session presents Cloud-eLab, an open and interactive cloud-based learning platform for Cyber and Machine Learning, aiming to inspire i) Deep and Wide learning, ii) Cognitive and Adaptation learning concepts for education. It has been successfully used in various cyber and machine learning courses in practice, and has the expandability to support more ML modules. In this session, we describe the block diagram of the proposed education platform, and provide two education application scenarios for unfolding Deep and Wide learning as well as Cognitive and Adaptation learning concepts. Cloud-eLab education platform will deliver personalized content for each student with flexibility to repeat the experiments at their own pace which allow the learner to be in control of the whole learning process.


Chameleon Tutorials

Introduction to Chameleon

Joon-Yee Chuah, TACC

This tutorial for new users describes the resources and capabilities of Chameleon Cloud and how to discover, allocate, launch and access bare-metal resources. Participants will perform a simple performance measurement experiment using Spark.

Advanced Topics: GPU Power Measurement and Complex Appliances

Joon-Yee Chuah, TACC

A two part tutorial.

Part 1: Machine Learning GPU Power Measurement on Chameleon

This part highlights the bare-metal capabilities of Chameleon Cloud, as well as the diverse hardware that is available for experimentation. Participants will measure GPU power consumption on a machine learning task using Jupyter notebook to control and measure the experiment. No prior knowledge of machine learning is necessary.

Part 2: Securing an NFS Complex Appliance

This part demonstrates the complex appliance capabilities of Chameleon Cloud. Participants will modify a HEAT orchestration template to secure an NFS service between a server and client and verify data-plane security.