ORIGINAL PAPER Cooperative modelling and design on the computing grid: data, flux and knowledge interoperability Antonio Laganà & Elda Rossi & Stefano Evangelisti Received: 31 October 2012 / Accepted: 1 April 2013 / Published online: 26 April 2013 Abstract The fast interconnections of the presently avail- able distributed platforms allow scientists to target highly complex problems by chaining software developed and maintained by experts of the relevant fields. A pillar of such cooperative endeavor in molecular and materials science and technologies is the so-called grid empowered molecular simulator that combines the expertise of molecular science theorists (electronic structure and nuclei dynamics) and experimentalists in order to build and validate ab initio models. This line has prompted an unprecedented level of data format standardization procedures, the bridging of high throughput and high performance platforms, the assemblage of ad hoc designed virtual experiments. In addition this approach has prompted the design and development of tools allowing the evaluation of the quality of the cooperative effort produced by the members of a given research com- munity as well as its rewards to such effort through a credit economy is reported. Keywords Cooperative . Data formats . D5cost . Grid computing . Grid economy . HPC . HTC . Molecular simulator . Q5cost Introduction The exploitation of advanced distributed power of the Eu- ropean Grid Infrastructure (EGI) [1] has allowed the design and implementation of a prototype Grid empowered molec- ular simulator (GEMS) [2] based on the combination of collaborative interoperable service oriented computing. GEMS, in fact, builds up, out of the first principles of physics, accurate evaluations of physical observables. To this end GEMS is articulated into the following steps: – Interaction: a step aimed either at performing ab initio calculations determining (at various levels of accuracy) of the electronic structure of the considered molecular systems within the Born-Oppenheimer approximation and at incorporating, if available, other existing ab initio and/or experimental information to the end of providing either a single point evaluation of the potential energy or an extended point-wise representation (including all the molecular geometries significant to the description of the evolution of the reactive process) of the related potential energy surface (PES); – Fitting: a step aimed at performing, if required, either a global or a local interpolation of the available ab initio data using a suitable functional form and enforcing (whenever is possible) the reproduction of critical ge- ometries and symmetries of the PES; – Dynamics: a step aimed at integrating nuclei dynamics equations either on a single potential energy point (pro- vided that is equipped with derivatives) or on the fitted PES; – Observables: a step aimed at averaging over unobserved parameters and extending to larger scales (non atomistic) to the end of providing a realistic estimate of a measurable physical property. In practice, however, the implementation of such scheme has issued unprecedented challenges in terms of software interoperability. These challenges and related solutions worked out by the authors for the Chemistry and Molecular & Materials Science and Technologies (CMMST) Virtual Organizations (VO) and Virtual Research Communities (VRC) A. Laganà (*) Dipartimento di Chimica, Università di Perugia, Perugia, Italy e-mail: lagana05@gmail.com E. Rossi CINECA, Casalecchio di Reno, Italy e-mail: e.rossi@cineca.it S. Evangelisti Universitè Paul Sabatier, Toulouse, France e-mail: stefano@irsamc.ups-tlse.fr J Mol Model (2013) 19:4215–4222 DOI 10.1007/s00894-013-1844-6 # Springer-Verlag Berlin Heidelberg 2013