Crowd computing and the cloud: climate modelling

Crowd computing and the cloud: climate modelling

Volunteer or crowd computing is becoming increasingly popular to solve complex research problems, from a diverse range of areas. The world's largest climate experiment, climateprediction.net, for example, is a volunteer computing, climate modelling project created in 1999 and based at the University of Oxford.

Traditionally, climate models have been run using supercomputers because of the vast computational complexity and high cost. However, climateprediction.net runs hundreds of thousands of state-of-the-art climate models, each very slightly different from the others, representing the real world. This technique, known as ensemble modelling, requires an enormous amount of computing power but is not an efficient use of large HPC facilities. Hence using instead the power of thousands of ordinary computers, each tackling one small part of the larger modelling task.

The majority of crowd computing projects use the Berkeley Open Infrastructure for Network Computing (BOINC) platform, which provides a range of different services to manage all computation aspects of a project. The BOINC system is ideal in cases where not only does the research community involved need low cost access to massive computing resource, but there is also a significant public interest in the research.

Climateprediction.net has now been running for more than 10 years and faces a number of evolving challenges, including an increasing and variable need for new computational and storage resources and restrictions imposed on using more complex models by the processing power and memory of volunteers' computers.

A new paper published in Geoscientific Model Development, authored by the Centre's Professor David Wallom, Peter Uhe and other researchers[1], discusses the way in which Cloud services can help BOINC based projects deliver results in a fast, on demand manner. This is difficult to achieve using volunteers, and at the same time, using scalable cloud resources for short on demand projects can optimize the use of the available resources.

The study shows how this design can be used as an efficient distributed computing platform within the Cloud, and outlines new approaches that could open up new possibilities in this field, using climateprediction.net as a case study. 

It successfully demonstrates that it is possible to run simulations of a climatic model using infrastructure in the Cloud, which has never previously been tested. In addition, cloud services enable a given number of tasks to be completed in some cases five times faster than using the regular volunteer computing infrastructure - although at a high cost.

Geoscientific Model Development is an interactive open-access journal of the European Geosciences Union.

Read the full paper here.

 

 [1] Enabling BOINC in Infrastructure as a Service Cloud Systems, Diego Montes (Universidade de Vigo), Juan A. Añel (Smith School of Enterprise and the Environment, University of Oxford), Tomás F. Pena (University of Santiago de Compostela), Peter Uhe and David Wallom (University of Oxford's e-Research Centre).