The National Resource Conservation Service was faced with some staggering figures: more than 100 scientific modeling tools deployed to each and every one of its 12,500 planners, who were spread out in more than 2,000 offices.
Each NRCS employee needs access to those applications as they travel into the field to help farmers and ranchers determine how best to handle the critical issues that affect their businesses and livelihoods.
But the NRCS, part of the U.S. Department of Agriculture, realized it had an increasingly costly and ineffective system on its hands.
So in 2009, the NRCS started looking at cloud computing as a way to centralize resources, cut costs, and provide a more scalable and flexible IT backbone for the organization and its thousands of field workers.
The agency, though, needed a tool to help move its legacy applications to the cloud. It developed the Cloud Services Innovation Platform (CSIP) to meet that goal.
The CSIP is designed to facilitate the migration of existing scientific modeling applications to any Amazon EC2-compatible infrastructure-as-a-service cloud environment. One of its key components is its ability to take an existing environmental model and turn it into a Web service that can run in the cloud.
The agency also sees compatibility of this open-source infrastructure with commercial cloud-hosting offerings as critical, since that compatibility will enable both successful production-level deployment and reliable availability on a stable operational platform.
The NRCS tested this new framework using Eucalyptus open-source private cloud software. Using the Eucalyptus-based CSIP, the NRCS demonstrated that these environmental modeling tools could be easily accessible as a Web service from the cloud. This capability helps the agency deliver better support to farmers and ranchers by enabling, for example, an NRCS planner to simply use a mobile phone from the field to get answers to questions about soil erosion.
The agency is using two criteria to evaluate the success of this project: whether computational services can be quickly established on an open-source infrastructure and then moved to a commercial cloud provider, and whether the procedure generates significant cost savings that make it a compelling option for future development activities.
So far, the NRCS is reporting successes on both fronts. It has been able to set up legacy computational scientific models within a week for each one before moving them to commercial providers. Meanwhile, the agency is expecting to save money over its current server-farm approach.
The NRCS sees this project as an ongoing effort and expects to constantly update the underlying open-source infrastructure to fit changing development needs.
Read more about the 2012 Computerworld Honors Laureates.
Read more about Cloud Computing in Computerworld's Cloud Computing Topic Center.