Decentralized execution of constraint handling rules for ensembles

Abstract

CHR is a declarative, concurrent and committed choice rule-based constraint programming language. In this paper, we adapt CHR to provide a decentralized execution model for parallel and distributed programs. Specifically, we consider an execution model consisting of an ensemble of computing entities, each with its own constraint store and each capable of communicating with its neighbors. We extend CHR into CHR e , in which rewrite rules are executed at one location and are allowed to access the constraint store of its immediate neighbors. We give an operational semantics for CHR e , denoted ω e 0 , that defines incremental and asynchronous decentralized rewriting for the class of CHR e rules characterized by purely local matching (0-neighbor restricted rules). We show the soundness of the ω e 0 semantics with respect to the abstract CHR semantics. We then give a safe encoding of the more general 1-neighbor restricted rules as 0-neighbor restricted rules, and discuss how this encoding can be generalized to all CHR e programs. * Funded by the Qatar National Research Fund as project NPRP 09-667-1-100 (Effective Programming for Large Distributed Ensembles)

Comingle: Distributed Logic Programming for Decentralized Mobile Ensembles

“…Our prototype, discussed in Section 4, is instead based on a concurrent, node-centric model of computation, where each node manipulates its local facts and exchanges message with other nodes. We achieve this by compiling Comingle rules into the code that runs at each participating node [9].…”

“…An example is the selection of the indexing structures used by the Comingle runtime to carry out multiset matching with the best possible asymptotic time complexity [10]. Once a program has been statically checked, the compiler first applies a high-level source-to-source transformation [9] that converts a class of system-centric Comingle programs into node-centric rules. In addition to preserving soundness, the resulting node-centric program explicitly implements the communications and synchronizations that are required to correctly orchestrate the distributed execution of multi-party Comingle rules among a group of participating devices.…”

“…This system-centric view underlies popular frameworks such as Google Web Toolkit [7] (for client-server web development) and Map Reduce [4] (for parallel distributed algorithms on large-scale computing clusters). In earlier work [9,10], we generalized this approach to a much broader class of distributed computations by relying on a form of logic programming to orchestrate interactive distributed computations [9,10]. Comingle specializes this work to distributed applications running on mobile devices.…”

“…Comingle specializes this work to distributed applications running on mobile devices. Comingle is based on committed-choice multiset rewriting extended with explicit locality [9] and multiset comprehension patterns [10]. This provides declarative and concise means of implementing distributed computations, thus allowing the programmer to focus on what computations to synchronize rather than how to synchronize them.…”

“…This provides declarative and concise means of implementing distributed computations, thus allowing the programmer to focus on what computations to synchronize rather than how to synchronize them. The present work extends [9] by introducing triggers and actuators to integrate the Comingle multiset rewriting runtime with traditional code from mainstream mobile development frameworks (specifically Java and the Android SDK). This allows a developer to marry the best of both programming paradigms, using Comingle to orchestrate distributed computations among devices and traditional native code for computations within a device (e.g., user interface functionalities, local computations).…”