Subtasks of a single job can benefit from different cloud providers.
Scalytics considers each subtask as an element for optimization. Scalytics' optimizer always takes into account the data movement costs that will be incurred by applying multi-cloud processing.
Scalytics operators can run on different cloud providers and big data platforms.
Users no longer need to be aware of or to re-implement their applications to use different platforms. Whenever an alternative platform improves performance or reduces monetary costs for a particular task, Scalytics will notice this opportunity and proceed to replace the platform without any effort from users.
By breaking down a job into subtasks, Scalitycs can measure the benefits or running each subtask on different cloud providers.
Then, according to the user's objectives, Scalytics chooses the optimal cloud provider for each subtask.
Scalytics schedules subtasks to the cloud providers that achieves the best overall performance or lowest monetary cost.
Scalytics can reoptimize and re-schedule a subtask at runtime if this benefits the job execution time or cost. Users do not have to worry about this and can still leverage the advantage that each cloud provider has to offer.