Section: New Results

PSYNC: A partially synchronous language for fault-tolerant distributed algorithms

Participants : Cezara Drăgoi [correspondant] , Thomas Henzinger [IST Austria, Austria] , Damien Zufferey [MIT, CSAIL, USA] .

Fault-tolerant distributed systems, Programming languages, Verification Fault-tolerant distributed algorithms play an important role in many critical/high-availability applications. These algorithms are notori- ously difficult to implement correctly, due to asynchronous com- munication and the occurrence of faults, such as the network drop- ping messages or computers crashing. We introduce PSYNC in [4], a domain specific language based on the Heard-Of model, which views asynchronous faulty systems as syn- chronous ones with an adversarial environment that simulates asyn- chrony and faults by dropping messages. We define a runtime sys- tem for PSYNC that efficiently executes on asynchronous networks. We formalize the relation between the runtime system and PSYNC in terms of observational refinement. The high-level lockstep ab- straction introduced by PSYNC simplifies the design and imple- mentation of fault-tolerant distributed algorithms and enables auto- mated formal verification. We have implemented an embedding of PSYNC in the SCALA programming language with a runtime system for asynchronous networks. We show the applicability of PSYNC by implementing several important fault-tolerant distributed algorithms and we com- pare the implementation of consensus algorithms in PSYNC against implementations in other languages in terms of code size, runtime efficiency, and verification.