Section: New Results

A Chemical Approach for Autonomous Service Computing

Participants : Héctor Fernandez, Marko Obrovac, Thierry Priol, Cédric Tedeschi.

Chemistry-Inspired Workflow Management System for e-Science Applications

Participants : Héctor Fernandez, Cédric Tedeschi, Thierry Priol.

In the research track that aims at leveraging the properties of the chemical Programming models for autonomic computing, we have built a software based on the HOCL compiler (part of the HOCL-tools) that was actually deployed and experimented over the Grid'5000 platform. The experiments have shown, that envisioning the execution workflow as an autonomic chemical process is actually viable in practice. Experimented with different well known workflow-based e-Science applications, the software showed a performance level comparable to current top-rated scientific workflow management systems [25] .

Solving Workflow Patterns Through Molecular Composition

Participants : Héctor Fernandez, Cédric Tedeschi, Thierry Priol.

In the same area, but on a more conceptual point of view, we have shown how the expressive power of the chemical model can be leveraged to solve complex workflow patterns. This aspect was also integrated into the HOCL-tools and experimented over the Grid'5000 platform, following two architectures with a different level of decentralization, showing the advantages and drawbacks of decentralizing the workflow execution using a chemical workflow engine [26] .

Scalable Atomic Capture of Molecules

Participants : Marko Obrovac, Cédric Tedeschi.

Capturing the reactants involved in a reaction constitutes one of the main challenges in the execution of chemical programs. Doing it at large scale is one of the essential barriers hindering the actual execution of chemical programs at large scale. We proposed a protocol solving this issue on top of a distributed hash table (DHT). The DHT secures the scalability of the communications and provides a scalable discovery of reactants. Our protocol is triggered once reactants are found. It is made of two sub-protocols being used at different stages of the computation, according to the density of possible reactions. The protocol is validated through its proof of liveness and simulations showing that it is able to switch from one sub-protocol to the other efficiently, according to the execution's conditions [18] .