ABSTRACT

Company ABC was developing the next generation of their embedded system with resource constraints and significant real-time requirements. The architecture was an evolution of ABC’s previous (very successful) system. The system was going to be deployed in an environment with an available 64 Megs of RAM (much more than the previous system). The system also had strict real-time requirements for start-up. This time, however, they found they kept blowing their memory limit on run-up. The structure of the development organization had changed over the years and they were now implementing different modules of the system at five sites throughout Europe. This caused many back-and-forth interactions with the different development teams, requiring them to coordinate all the interactions among their components during run-up. The end result was that development took more than twice as long as anticipated, exceeded the original budget by 80 percent, and was still not able to meet the latency requirements or memory constraints for run-up.