While one release is in the field, another release is being deployed, a third and often fourth release are in testing, the next releases are being defined. With so many nearly simultaneous releases, it is imperative that the processes for defining, producing, and deploying AWIPS run efficiently, quickly, and cleanly. To this end, there are four phases involved in getting AWIPS releases to the forecast offices.

The first phase consists of determining the release contents. During this phase, the NWSHQ works with NWS forecast offices Development Organizations to prioritize requirements, determine feasibility, and potential schedule. Working groups like the SwEG and SET provide forums for discussion. The first phase is both the longest and least structured of the phases.

The second phase belongs to the Development Organizations. Throughout this phase, the actual design and implementation occurs. Occasionally, forecast offices participate in very early prototype or pre-Alpha tests to help focus development.

The third phase is the shortest and most structured phase. Within this phase, the entire release is built, integrated, and tested in a configuration-managed (CM) environment. The software is build under strict CM guidelines for repeatability and auditability. The software is installed and tested on four different baselined hardware configurations (WFO, OCONUS, collocated WFO/RFC) located in the PRC facility in McLean, Virginia. During the latter portion of the formal testing, Alpha site installs are scheduled for site verification of the release. After a formal verification review, the release is approved for the fourth phase -- deployment and support.

During the final phase, the release deployment is scheduled, installed, monitored, and accepted. After the deployment, support of the release begins in earnest with the NCF and SST playing major roles.