Attitudes to Cloud Based Simulation

The concept of running Engineering Simulations on “somebody else’s computer” is not new. Many experienced analysts, of a certain age, will tell you stories of taking punched cards to a bureau computer back in the day and the frightening costs of computer time. Those stories of course date back to the 1970s and early 80s when very few people had a computer on their desk and even if they did, it wouldn’t have been capable of running simulations.

Things changed in the late 80s with the introduction of workstations and high-end PCs, and simulation increasingly came in house. The state of the art has progressed steadily to where we are today, with most engineering organisations having multiple machines capable of running highly advanced simulations with users’ desktop machines handling routine work.

However, there is still room for improvement. In most cases of high-fidelity simulation, analysts still need to carefully consider hardware capacity and run times when building models, making assumptions and segmenting the problem to be solved to obtain an acceptable level of accuracy in an acceptable time. IT departments typically have to procure and maintain non-standard hardware for simulation often leading to an uneasy relationship with analysts. Hardware and software licensing schemes traditionally have not had flexibility to cope with peaks and troughs in demand. Engineers are increasingly working in geographically dispersed teams, reducing the benefit of on-site hardware, and presenting barriers to collaboration. As models move closer to the real world they also become larger and more complex, the quantity of data increases as does the value of the IP associated with the model, however management of simulation data is still not widely deployed or adopted.