Data Consistency in Distributed Virtual Reality Simulations Applied to Biology

Abstract

We perform biological simulations in the virtual reality context. In order to run large simulations, we choose to put together a set of standard computers and create a grid in charge of distributing the biological simulations. We propose to make the distribution on to the ReISCOP generic model, developed in our laboratory, which allows us to easily design biological simulations. This method is also based on the replication of passive elements. It is a spatial distribution in which local simulations are periodically synchronized and the consistency of replicated data is checked. This synchronization is not a strong synchronization but a weak one. The consistency method is built on the transmission between nodes of the grid of the state variations of the data. The software used for distributing simulations is named DIVA, and is an individual based software located on each node of the grid. DIVA confers a peer to peer architecture upon the grid.