The following three demo videos demonstrate key features of our platform.
Performance Modelling with Resource Utilisation Feedback
This demo shows how to upload and profile a new application on the HARNESS platform. It compares two different approaches to application profiling: extrapolated profiling that uses resource utilisation feedback to speed up the process of creating an application performance model; and a slower, blackbox profiling approach without any feedback from the platform.
Resource Scheduling with Network Constraints
This is a demo of an application deployment on Grid’5000. It demonstrates how to schedule cloud resources with and without network constraints. The former resource allocation satisfies better the application objectives during execution.
Dataflow Engines in the Cloud
This demo shows a deployment of the Reverse Time Migration (RTM) application. RTM uses one of more Dataflow Engines (DFEs) to compute the seismic waves propagation. The demo highlights two different aspects of our platform: an in-memory execution that uses our novel DFE virtualisation technology to scale the application across multiple DFEs automatically; and, a disk-based execution that uses the XtreemFS performance isolation mechanisms of our platform for multi-tenancy.