A New and Improved Computing System for the FireMAP Project
Faculty Mentor Information
Dr. Barry Myers Dale Hamilton
Presentation Date
7-2017
Abstract
Research projects often have complex algorithms, machine learning, or powerful programs that need to execute in a reasonable amount of time. The goal of installing and maintaining new servers is to create a powerful, flexible, and reliable ecosystem for these large processes to run.
This project is in support of the NNU Computer Science FireMAP research project, and as such has many different requirements to satisfy. The goal of the new servers is to host virtual machines that are capable of quickly creating orthomosaics and 3D models using Pix4D, storing terabytes of georeferenced orthomosaics and other imagery data, and hosting a web application to facilitate the selection of machine learning training data from the imagery. An advantage to having three separate new servers with four physical cores each is that algorithms written for them can take advantage of those cores through parallelism, speeding up calculations tremendously. Another key feature of this project is a VPN that allows researchers inside and outside the campus access to the file share and virtual machines. These capabilities provide a computing environment to enable researchers the tools needed for the success of the FireMAP project.
A New and Improved Computing System for the FireMAP Project
Research projects often have complex algorithms, machine learning, or powerful programs that need to execute in a reasonable amount of time. The goal of installing and maintaining new servers is to create a powerful, flexible, and reliable ecosystem for these large processes to run.
This project is in support of the NNU Computer Science FireMAP research project, and as such has many different requirements to satisfy. The goal of the new servers is to host virtual machines that are capable of quickly creating orthomosaics and 3D models using Pix4D, storing terabytes of georeferenced orthomosaics and other imagery data, and hosting a web application to facilitate the selection of machine learning training data from the imagery. An advantage to having three separate new servers with four physical cores each is that algorithms written for them can take advantage of those cores through parallelism, speeding up calculations tremendously. Another key feature of this project is a VPN that allows researchers inside and outside the campus access to the file share and virtual machines. These capabilities provide a computing environment to enable researchers the tools needed for the success of the FireMAP project.