Introducing Plastic Fog
Plastic Fog targets ARM processors and the upcoming mbed IoT operating system. We think microcontrollers will evolve to become more than just stand-alone compute engines. We believe each microcontroller will ultimately be like a pixel in a holographic fabric that extends throughout the Internet of Things. Something similar to an all-knowing neuron, which learns from all other neurons.
What is Plastic Fog?
Plastic Fog is a patent-pending software framework for performing data analytics entirely within the fog-of-things, i.e. on and across a set of networked devices such as smart vehicles, factory floor machines, smart enterprise equipment, handheld wireless devices, connected home appliances, wearables, and so forth.
Why is Plastic Fog important for ARM processors?
For one thing, Plastic Fog is designed to perform powerful data analytics operations, even on and across highly resource-constrained microcontrollers. Part of the magic involves leveraging ARM mbed.
Our longer term vision is even more bold. Plastic Fog utilizes an underlying methodology that could easily be integrated into hardware and firmware, including System-on-a-Chip (SOC) products. This is because a singular technology base is used to efficiently combine data modeling results across nodes, even when a wide variety of software modeling plugins are implemented on top of that singular technology base. Thus, we hope to one day become part of the processor, itself. The part that enables the processor to be the all-knowing neuron.
- Plastic Fog is able to reduce, or eliminate altogether, the amount of Big Data that must be transported, ingested, and analyzed within centralized locations such as data warehouses, while still providing the benefits associated with deep data analysis across very large numbers of devices and data streams.
- Plastic Fog can perform data analytics functions that discover end user behavior patterns, while avoiding the privacy invasion issues surrounding the transport of user behavior data across networks, and the storage of this type of data within centralized locations. In fact, it is possible with Plastic Fog to provide the value of data analytics and machine learning—such as intelligently targeted web ads, vehicle-to-vehicle real time safety functions, and so forth—without sharing or persisting any user behavior data whatsoever in any centralized system.
- Plastic Fog can perform deep data analysis in real time at scale. The Internet of Things is predicted to grow to billions – or even trillions! – of nodes. Traditional Big Data approaches may have difficulty scaling to this level, given the requirement to provide real time data analysis at scale. Consider vehicle-to-vehicle (V2V) applications, where analyses must be performed in a fraction of a second, for example in order to avoid accidents. It is very difficult to assure this type of real time responsiveness using the centralized Big Data model. In addition, any long-distance network failures could disable the entire system at a potentially critical moment. Plastic Fog’s approach can function in real time, without centralized control (or with much more limited and non-critical centralized control) and thereby provide greater assurance of real time responsiveness.
- Plastic Fog can support massively scalable multi-company and multi-industry solutions, while allowing all participating organizations to keep their internal data private and secure. These kinds of large-scale solutions will provide extreme efficiency gains and other benefits that are not possible with single-company solutions. Plastic Fog may be the only possible way to build these types of applications, since companies will never share their proprietary data with each other within a single centralized repository. In fact, it is not feasible to create a centralized collection of data from thousands of companies. Yet, such a centralized data repository would be necessary, when using traditional Big Data approaches, in order to analyze the multi-company data wholistically. Plastic Fog gets around this with an elegant, flexible, and finely granular distributed solution, which amounts to a paradigm shift!
We encourage you to try out our early releases, and provide feedback. You can help determine the direction that this technology takes.
- Plastic Fog Beta - December 2015
- Plastic Fog GA - Q1 2016