The Impact of Scale
Many systems of the future will be of ultra-large size on one or many dimensions – number of lines of code; number of people employing the system for different purposes; amount of data stored, accessed, manipulated, and refined; number of connections and interdependencies among software components; number of hardware elements to which they interface. They will be ultra-large-scale (ULS) systems. They will be socio-technical ecosystems. Is the software community ready to tackle ULS systems? Will incremental changes in our current software development and management practices be sufficient?
In fact, the characteristics of ULS systems, already evident in some of today’s largest systems, imply changes in the fundamental assumptions that underlie today’s software engineering approaches. The gaps are strategic, not tactical. Issues that are not significant at smaller scales become significant at ultra-large scales. Our current practices are undermined by the characteristics of ULS systems. A new multi-disciplinary perspective and research are needed.
This talk shares the results of a year-long study on ULS systems, documented in Ultra-Large-Scale Systems: The Software Challenge of the Future (ISBN 0-9786956-0-7), as well as recent work motivated by the study.