Modeling the Cloud to Enhance Capabilities for Crises and Catastrophe Management
Abstract
In order for cloud computing infrastructures to be successfully deployed in real world scenarios as tools for crisis and catastrophe management, where large amounts of dynamic information even real time information have to be processed, novel algorithm designs, that can address the challenges of resource dynamism, scalability and virtualization in cloud environments, are needed. The overarching goal of this project is the design and development of a flexible mathematical modeling framework for the cloud infrastructure that can also leverage existing mathematical representations (e.g. graph theory), performance models (e.g. network models) and analysis tools (e.g. statistical analysis). In pursuit of this goal, we conducted an initial study to understand the impact of various cloud hardware and job parameters on performance. As part of this study, a cloud simulation environment on 32 compute nodes was used to run test programs under varying load conditions. The results and analysis of the initial performance study was used to explore adaptive algorithms designs for social network analysis for large and dynamic networks. We also identified a scenario, based on the challenging and computationally intensive problem of modeling resilience of social groups that we will use to validate our cloud modeling framework. It is worth noting that while the original performance period was 3 years, the project had a truncated performance period of less than 16 months.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 16, 2016
- Accession Number
- AD1037352
Entities
People
- Eunice E. Santos
Organizations
- University of Texas at El Paso