Statically verified refinements for multiparty protocols

Abstract

With distributed computing becoming ubiquitous in the modern era, safe distributed programming is an open challenge. To address this, multiparty session types (MPST) provide a typing discipline for message-passing concurrency, guaranteeing communication safety properties such as deadlock freedom.While originally MPST focus on the communication aspects, and employ a simple typing system for communication payloads, communication protocols in the real world usually contain constraints on the payload. We introduce refined multiparty session types (RMPST), an extension of MPST, that express data dependent protocols via refinement types on the data types.We provide an implementation of RMPST, in a toolchain called Session ⋆ , using Scribble, a toolchain for multiparty protocols, and targeting F ⋆ , a verification-oriented functional programming language. Users can describe a protocol in Scribble and implement the endpoints in F ⋆ using refinement-typed APIs generated from the protocol. The F ⋆ compiler can then statically verify the refinements. Moreover, we use a novel approach of callback-styled API generation, providing static linearity guarantees with the inversion of control. We evaluate our approach with real world examples and show that it has little overhead compared to a naïve implementation, while guaranteeing safety properties from the underlying theory. CCS Concepts: • Theory of computation → Automated reasoning; Distributed computing models; • Software and its engineering → Source code generation.

CAMP: Cost-Aware Multiparty Session Protocols

“…This section presents a number of benchmarks used to evaluate the predictive power of CAMP. Our benchmarks are taken from multiple different sources, mostly MPST-based tools [Castro et al 2019;Castro-Perez and Yoshida 2020;Imai et al 2020;Ng et al 2015;Zhou et al 2020], but also a subset of the Savina actor benchmarking suite [Imam and Sarkar 2014]. We categorise our benchmarks following the structure of the Savina benchmarking suite: (i) microbenchmarks, (ii) concurrency benchmarks, and (iii) parallel algorithms.…”

“…Two-Buyer Protocol (twobuy). We use an F★ implementation taken from [Zhou et al 2020], and extracted into OCaml. Sleeping Barberx, Dining Philosophers and Cigarette Smoker (barb, dphil, csmok).…”

“…We prove that the cost analysis by CAMP is sound with respect to the operational semantics of a given global type instrumented with sizes and execution costs; and extensible to analyse communication optimisations. Our main contributions are: a) we define the semantics of CAMP, integrating global and local type semantics with local computation costs, that can be used to explore the costs of particular traces ( §3); b) we instrument global types with size and local cost information, and use it to statically estimate an upper bound of the execution cost of a protocol, that we prove sound with respect to the operational semantics ( §4); c) we define multiple metrics on the cost recurrences associated with recursive protocols, that can be used to effectively analyse the performance behaviour of potentially infinite executions ( §5); d) we extend CAMP to handle asynchronous message optimisations, enabling us to statically quantify the potential performance gains when performing such optimisations/reordering ( §6); e) we implement a DSL for specifying global types, from which we can extract cost equations ( §7), and we compare our cost model predictions with real benchmarks used in MPST implementations in different languages: C-MPI [Ng et al 2015], C+pthreads [Castro-Perez and Yoshida 2020], Go [Castro et al 2019], OCaml [Imai et al 2020] and F★ [Zhou et al 2020]. Additionally, we apply CAMP to a subset of the Savina benchmarking suite (Scala) [Imam and Sarkar 2014].…”

Communication-safe web programming in TypeScript with routed multiparty session types

“…3). The local type of an endpoint can be then used in the code generation process, to generate APIs that are correct by construction [17,20,35].…”

Zooid: a DSL for certified multiparty computation: from mechanised metatheory to certified multiparty processes

“…The behaviour of global and local types is defined by (asynchronous) labelled transition systems (LTS) whose sound and complete correspondence is key to provide: progress of processes [28], synthesis of global protocols [16,33], and to establish bisimulation for processes [32]. Practically, typelevel transition systems are particularly useful for, e.g., dynamic monitoring of components in distributed systems [14] and generating deadlock-free APIs of various programming languages, e.g., [10,29,34,39,53].…”