How it works

A multi-cloud application’s (re-)deployment is realised by executing a deployment process. This process includes the following sequence of activities presented below:

Step 1

The user describes the mixed application via the CAMEL language, including requirements for both normal and serverless components.

Step 2

The initial deployment plan for the application is calculated by executing components of the platform’s Upperware module. – The CP Generator conducts application profiling and finally produces a CP (constraint optimisation problem) model, which imprints the deployment alternatives for all application components. – The Metasolver (especially the constraint solvers that it incorporates and manages) performs deployment reasoning over the CP model.

Step 3

The deployment is executed across multiple clouds based on the initial deployment plan calculated.

Step 4

After its deployment, the application is monitored by the Event Processing subsystem.

Step 5

Based on gathered and aggregated metric measurements, the application adaptation at runtime will be performed.

Step 6

The runtime application reconfiguration will be run, based on the new deployment plan generated, again via the Adapter and Cloudiator components.

The way MELODIC works

More about MELODIC

The MELODIC platform

learn more about MELODIC platform and features.

Download MELODIC

It's open source, with full source code available. Download here!