Scaling architecture through flexible deployment

Nancy Nunes demonstrates how to construct versatile software components that can be deployed on multiple platforms without performance hits normally associated with generalized behavior. Youll leave with understanding and an operational example of how to package software components to make them adaptable to running in a single processor or distributed processing system.

System architectures often influence the software architecture of a product. If the system architecture misses an important performance requirement, the software architecture might have to suffer through a major redesign effort which is likely to cause software delays. But what if you could protect your software against late-breaking requirements, like increases in scale, that change system architecture? What if you could allocate software components to system processing components during production rather than only during design, freeing your logical software capabilities from the realities of actual processing limitations until late in the production process? By employing some interface ground rules and leveraging a few design mechanisms you can deftly combine components as needed to ensure scalability for product line variation and product growth. Nancy Nunes demonstrates how to use a powerful production-time variation technique for creating software that scales by allowing you to deploy software components to various processing environments, from single process/single thread to distributed multiprocessor systems. Nancy walks you through an example showing how a set of software components can be combined to form separate optimized executables for different deployments. We will then build, run, and explore these components deployed for single process, multiprocess, and multiprocessor deployments.