Homeostatic Architectures for Robust Spatial Computing

For open-ended computational growth, we argue that: (1) Instead of hardwiring and hiding the spatial relationships of computing components, computer architecture should expose and soften them, and (2) Instead of minimizing reliability as just a hardware problem, robustness should climb the computational stack toward the end users. We suggest that eventually all truly large-scale computers will be robust spatial computers - even if intended neither for overtly spatialized tasks, nor harsh environment deployments. This paper is an introduction for the spatial computing community to the Movable Feast Machine (MFM), a computing model in the spirit of an object-oriented asynchronous cellular automata, with which we are exploring robust, indefinitely scalable computations. We briefly motivate the approach and present the model, then illustrate some robustness mechanisms such as redundancy, sloppiness, and homeostasis, showing how a basic task like sorting can be reconceived for robustness within a homeostatic architecture.