AmbientsPeer-to-peer Programs and Data.
Description
The decentralized web has shown great promise for highly available, massively distributed, open, censorship-free, privacy-preserving and non-exclusive application ecosystems. The technology landscape is, however, becoming fragmented due to a lack of collaboration and safe interoperability between platforms. This fragmentation exacerbates the uncertainty, risk and cost of decentralized application development and deployment. Furthermore, competing decentralized technology projects promote application programming models that are logically centralized, removing incentives for pursuing a fully decentralized web.
To counter this problem, we introduce the Ambients protocol to create and verifiably execute distributed programs in decentralized, peer-to-peer networks. The main contribution of the Ambients protocol is a process-algebraic specification of its purely functional, safe and composable programming model. Combined with a content-addressed and location-agnostic execution model, what we present is a novel protocol for distributed computation.
A deterministic evaluation of Ambients programs is guaranteed by a novel, strongly confluent, distributed rewrite system based on ambient calculus. Deterministic verification of the rewrite system, and its integrity and authenticity, are guaranteed by recording the execution and control flow of programs as Merkle-DAG-based, partially ordered, immutable event logs. Because of these guarantees, the Ambients protocol enables building and composing truly decentralized applications, databases and services, making safe interoperability and collaboration feasible within existing networks and platforms.
Protocols