I was looking at:
I quite like the ability to create a logical deployment model but the model in the youtube movie is not better than the current Savara roles/relationships. So I started thinking about how one might use the current roles/relationships in a similar fashion (really we need to pin the roles to participants because these represent locations). And I came up with the following idea.
First of all we could take the same technique of using scenarios to simulate (test models) and the other way around (using the scenario to animate the model). The scenarios could include time-delays to understand what would happen from a time perspective if delays in sending or receiving or interacting occur. Equally the model could have semantic annotations for SLA's added to the roles and relationships.
Then you could drop a scenario onto a model and see what happens. If the SLA's are breached this could show up on the model and on the scenario and be highlighted in some way.
The aim is to use the Savara model and the Savara scenario collectively to simulate deployment scenarios and so better choose what platforms t run on.
Yes these are good ideas. The inclusion of a deployment model has been in the plan from day one, but we haven't yet fleshed out how it should be used.
In terms of the simulation functionality, it sounds like it would be useful to have different 'implementation profiles' associated with a deployment model, representing (1) the alternative technologies that may be used, or (2) different performance information related to the various hardware/software/services involved.
It would also be good to capture realtime information from the production use of some of those components (whether servers, network, services, etc) and therefore enable the simulation to be based on realistic information.
Can you create some jiras for the different features - one per discrete requirement, with relevant links between them to show dependencies.
|Retrieving data ...|