Effective Computation Resilience in High Performance and Distributed Environments

The work described in this paper aims at effective computation resilience for complex simulations in high performance and distributed environments. Computation resilience is a complicated and delicate area; it deals with many types of simulation cores, many types of data on various input levels and also with many types of end-users, which have different requirements and expectations. Predictions about system and computation behaviors must be done based on deep knowledge about underlying infrastructures, and simulations' mathematical and realization backgrounds. Our conceptual framework is intended to allow independent collaborations between domain experts as end-users and providers of the computational power by taking on all of the deployment troubles arising within a given computing environment. The goal of our work is to provide a generalized approach for effective scalable usage of the computing power and to help domain-experts, so that they could concentrate more intensive on their domain solutions without the need of investing efforts in learning and adapting to the new IT backbone technologies