Hardware and network architectures change faster than can be sufficiently addressed by programming models – accordingly, programming models are subject to a 3-5 years’ change. In order to achieve long-term sustainability, not only a more “consistent” approach is required, but also a significant simplification of the model itself to reach a broader user community: current models are restricted to a few thousand experts worldwide, but with the increasing uptake of multi-core and clustered processors in PCs and even mobile devices, the number of users able to exploit multi-core technologies need to rise to millions.
S(o)OS proposes a different approach towards this issue, by examining the underlying Operating System in more detail: current OS models take a monolithic, hardware coupled approach and stacking of middlewares and programming models only makes the system less efficient without actually addressing the problem. The proposal aims at (at least) one sustainable (in the sense of "still relevant in 2020") Operating System model that addresses in particular the current and future development of processor and network architectures. By drawing strongly from the Service Oriented Architecture approach, S(o)OS will enable dynamic integration of different resource types, via different communication models – ranging from multi-core processors to networked “resource fabrics”. This will effectively decouple the OS and thus code execution from the underlying hardware, thus addressing sustainability beyond near-future processor-, hardware- and network-architectures.