Data intensive scientific computing

Applications are continually growing in both size and scale. As a result of this growth the need for distributed processing has become increasingly apparent. Cloud computing offers an effective method of hiring machines to facilitate the execution of specific applications. Users of cloud services have the flexibility to assemble configurations of instances that satisfy their specific computational requirements. Cloud computing also adopts a pay as you go nature ensuring that users only pay for the services that they use. Due to its high cost effectiveness and elasticity, cloud computing has become a desirable platform for many different applications and services. In this thesis we attempt to develop scheduling solutions for a scientific application called Montage. There is an existing cloud native workflow execution framework that has been tested in Amazon EC2.