The Unified Distributed Architecture

What is UDA?

AllBlock’s long-term goal is to initiate the open-source development of an architecture capable of supporting fully decentralized and distributed cloud and fog computation. No central servers. No trusted parties. Computational resources are offered on the network by peers, other peers contract to use those computational resources, work is done, and money changes hands safely and seamlessly. Data can be stored securely with 100% availability and searched like an advanced database. Web services can be made fully autonomous. And in the coming world of IoT devices in every building and on every corner, a pedestrian walking by can augment her smartphone with the power of the whole network, effectively making it a mobile supercomputer.

Mobile Supercomputer

Imagine a personalized virtual PC that persists in the cloud, safely holding your documents and other data and allowing you to access them from anywhere. But unlike the traditional cloud, you won’t have to place your trust in any company or other institution; code never lies. Harness the power of the cloud and edge devices around you to play graphically intensive games or enter augmented reality with only a low-powered phone or tablet interface. You’ll have more cheap computational resources at your fingertips than you could ever use.

Autonomous Virtual Agents

Recent advances in cryptography allow for computation to be done on encrypted data without decrypting it first. The implications are astounding: ‘virtual agents’ (essentially distributed programs) will be able to run on the UDA layer completely autonomously as long as there’s enough money in the accounts associated with them. Autonomy means a virtual agent could run a consortium blockchain (always orders of magnitude simpler and faster than public blockchains) without the consortium. Autonomy means a virtual agent could be placed in charge of mission critical systems, like traffic control, free of the single largest intrusion vector: user error. Sure, current implementations of homomorphic encryption are a bit slow, but that’s never stopped you.

Swarm Intelligence Agents

A suitably designed virtual agent could facilitate the self-coordination of driverless cars, allowing them to share local data and collectively compute near-optimal solutions for large-scale pathing and traffic flow. Similarly, a virtual agent might one day powr the medical nanobot swarm acting as your artificial immune system, allowing them to collectively adapt to changing conditions and sustain moderate levels of attrition without loss of functionality.

What is Biome-Computing?

Computing is moving to ‘cloud’ (remote network) and ‘fog’ (edge device) architectures. In the process, it’s becoming more and more distributed, less reliant on central points of failure, and increasingly flexible, powerful, and convenient for end users. Many of the new protocols and algorithms making this possible are inspired by natural processes and behaviors —electric flow, immune systems, swarm and flock behaviors, neural activity. This suggests an evocative conceptual analogy: many physical agents, from servers to PCs to IoT edge devices, who don’t necessarily trust one another interacting through distributed, modular, often organic-like means to form an ecosystem of virtual agents that populate the fog and clouds above. Biome computing —the future.