CESWP Usage Map
Stage
Deploy
Run
Acquire compilers
Acquire other tools
Acquire software and licenses
Acquire model (source code)
Build model
Deliver output to user (optional)
Instantiate Virtual Machine
Select Amazon Machine Image (AMI)
Bundle machine image
Archive output
Post-
Process
Archive input (optional)
Develop
Register the machine image in the cloud inventory
Register the model in the portal
How will scientists use CESWP*?
There are five types of activity involved:
Developing models
Staging models for deployment on CESWP
Deploying models on CESWP
Running simulations on CESWP
Post-processing results
The first activity (Developing models) and the last activity 
(Post-processing) are outside of the scope of CESWP.
The other activities (Staging, Deploying, and Running) 
form the core of CESWP.
Let's take a high level walk through the core activities:
This is the core of the scientists' activity. It may take many years before a model is mature enough to share with other scientists through the portal.
Once a model is stable and ready to share with other scientists, it can be staged on a CESWP Virtual Machine.
Here are some of the steps involved in staging a model:
Although model development occurs before CESWP, there are things that we can do now to  make it easier for the scientists to deploy their models to CESWP later.
The CESWP development team is helping by setting up a source code control system for use by the science team.
CESWP Activities
Create a manifest (i.e. list of everything needed to run the model on a Virtual Machine)
Now the VM is ready to be deployed to the Portal. We just have to do a couple of things:
Prepare
Use
We are also exchanging software engineering best-practices with the science team to help ensure that their work will be reusable and easier to share with others.
Need to know inputs and outputs
Before CESWP
After CESWP
Once a model has been deployed, scientists can run simulations based on the model. Here are some of the steps involved in running a simulation:
Acquire input:
Manually, and/or
From a file, and/or 
Automatically (from a known source)
Execute:
Manually (user) or
Automatically (time or event-based)
Produce Output:
to screen and/or
to file
Validate input:
Verify that data is complete
Validate that data is within acceptable ranges
Validate that simulation will execute within allowable resource constraints (time/cpu/output)
After the simulation has executed, the scientist may want to do some further processing on the output data.
For example, the scientist may want to produce a visualization based on the data.
Or the output of the simulation may be used as the input for another simulation, which can be run through the portal, and so on...
In some cases, if there are common sequences of model execution (e.g. first run model A, then model B, then model C), it may be possible to set them up and run them as a workflow in CESWP. (This is currently out of scope.)
Once a model has been staged, it is ready to be deployed to the portal. Here are some of the steps required to deploy a model:
First, the scientist navigates to the portal: 

www.ceswp.ca or 

www.cssdp.ca

and selects a model from the menu.
Then CESWP kicks in and:
gets the inputs
executes the simulation, and
produces the outputs

Here are some of the details...
*CESWP = Cloud-enabled Space Weather Modeling and Data Assimilation Platform