DataSys: Data-Intensive Distributed Systems LaboratoryData-Intensive Distributed Systems Laboratory

Illinois Institute of Technology
Department of Computer Science

Networking and Distributed Systems in High-Energy Physics at Fermi National Laboratory     Gabriele Garzoglio

Dr. Gabriele Garzoglio
Distributed Offline Computing Services Group
Grid and Cloud Computing Department

Fermi National Accelerator Laboratory

Stuart Building 201
Monday, April 30th, 2012


This event will present speakers (Gabriele Garzoglio) from Fermi National Laboratory, discussing two summer internship opportunities at Fermi Labs.

Title: Test utilization of the 100 Gbps network ANI test bed for the use cases of High Energy Physics (HEP)

Abstract: Fermilab is an active participant in the Advanced Networking Initiative (ANI) 100 Gbps network test bed. We are planning to use the infrastructure at our disposal to "fill the pipe" with data patterns typical of HEP and evaluate this future network technology. Clients and server machines in the test bed are available to us to install appropriate infrastructure (Lustre, Squid, GridFTP, xrootd, etc.) to run our tests. This would be a great project for a student interested in network on this not-yet-generally-available technology. The student would work with Fermilab staff expert in the HEP use cases.

Title: A visualization system for Cloud infrastructure metrics

Abstract: In the past few years, Cloud infrastructure to dynamically instantiate virtual machines has emerged as the preferred means to achieve flexibility, adaptability, and scalability in computing resource deployment. FermiCloud is an Infrastructure as a Service that targets the computational use cases of the Fermilab scientific stakeholders. The FermiCloud project is enhancing its infrastructure to better analyze the resource usage trends and users behavior. To this end, FermiCloud is being integrated with accounting and monitoring systems to gather detailed information about the utilization of the system. As the type and amount of relevant information grows, the need for visualizing the information gathered becomes more and more important. This project aims at defining and implementing novel visualization techniques to help with the analysis of the FermiCloud resource utilization. The visualization will display and correlate Cloud metrics using graphs, plots, histograms, charts, etc. The consumers of this visual material include the FermiCloud operators, the users, the Computing Sector management, and the funding agencies. Goal of the project is to simplify with visual representations the correlation and analysis of the FermiCloud utilization metrics. We seek a collaborator with expertise in web development on UNIX platforms. The language involved in the development of the visualization system will include Java, Javascript, HTML, python, and mysql.