Biomolecular committor probability calculation enabled by processing in network storage

Abstract

Computationally complex and data intensive atomic scale biomolecular simulation is enabled via processing in network storage (PINS): a novel distributed system framework to overcome bandwidth, compute, storage, organizational, and security challenges inherent to the wide-area computation and storage grid. PINS is presented as an effective and scalable scientific simulation framework to meet the unbounded requirements of a ‘user of infinite need’. The novel hybrid database–filesystem architecture enables the high throughput computation and data generation required by our scientific target. Biomolecular simulation methods are correlated with the primary PINS components, including: client tools, hybrid database/file management service (GEMS), computation engine (Condor), virtual file system adapter (Parrot), and local file servers (Chirp). Performance for the PINS prototype is reported for the committor probability calculation of a solvated protein domain requiring 500 independent simulations and the generation of over 1,000,000 output files.