I would like to get an advice about the following situation:
Say I have a three tier application (TheApplication) with two deployments in production.
One of the deployment uses a "canonical three tier architecture" where the three elements are deployed on differents nodes (one node is the desktop, the second hosts the application server and the third hosts the D.B).
The other deployment (also in production) uses a "simpler" three tier architecture with only two nodes (one node is the desktop and the other node hosts both the application and the DB).
I want to describe the "application deployment technical architecture" of these two solutions.
Does it mean
1) I have to define (in the application provider) two different application deployment for production ?
or
2) I have to define one application deployment with two "deployed application instance" ?
Remark : I think I will not be able to describe the technical architecture if I use the later
Multiple technical architectures for application deployment
-
jonathan.carter
- Posts: 1087
- Joined: 04 Feb 2009, 15:44
First, can I start with a question?
Do the two application deployments (both production) serve different user communities or are they something like a failover pair? That is, users are not explicitly aware of connecting to one or the other?
What I would like to understand is whether these are two deployments of the same application (e.g. TheApplication in North America and TheApplication in Europe).
It sounds like this is the case and not just that the application is itself distributed across multiple physical architectures such as in clustering (in such cases we use the second option that you describe)
We support both approaches, by the way!
There is an element of making sure we capture in the model the important/interesting aspects and in this case, the true 3-tier vs simplified 3-tier is what we want to understand.
In this case, what we can do it define two separate Technology Product Builds that include the physical server hosts as part of the model. One for the 'canonical 3-tier' and one build for the 'simpler 3-tier'.
We can create two separate Application Deployments (which we should anyway, based on the above TheApplication in North America / TheApplication in Europe example). We then associate the relevant Technology Product Build with each Application Deployment.
Of course, in terms of modelling workflow, we can define the relevant Technology Product Builds as we go, starting from the Application Deployment.
So far, I've come at this from the Technology point of view but I'd like to quickly touch on the Application perspective in terms of how it relates to its users.
If each of these deployments is supporting different user communities - (which we capture by relating a Physical Process to an Application Provider) then we need to define two separate Application Providers for each deployment. Each of these will have its own Application Deployment which can be related to a Technology Product Build that accurately reflects the 3-tier architecture.
Let me know some about these deployments
Jonathan
Do the two application deployments (both production) serve different user communities or are they something like a failover pair? That is, users are not explicitly aware of connecting to one or the other?
What I would like to understand is whether these are two deployments of the same application (e.g. TheApplication in North America and TheApplication in Europe).
It sounds like this is the case and not just that the application is itself distributed across multiple physical architectures such as in clustering (in such cases we use the second option that you describe)
We support both approaches, by the way!
There is an element of making sure we capture in the model the important/interesting aspects and in this case, the true 3-tier vs simplified 3-tier is what we want to understand.
In this case, what we can do it define two separate Technology Product Builds that include the physical server hosts as part of the model. One for the 'canonical 3-tier' and one build for the 'simpler 3-tier'.
We can create two separate Application Deployments (which we should anyway, based on the above TheApplication in North America / TheApplication in Europe example). We then associate the relevant Technology Product Build with each Application Deployment.
Of course, in terms of modelling workflow, we can define the relevant Technology Product Builds as we go, starting from the Application Deployment.
So far, I've come at this from the Technology point of view but I'd like to quickly touch on the Application perspective in terms of how it relates to its users.
If each of these deployments is supporting different user communities - (which we capture by relating a Physical Process to an Application Provider) then we need to define two separate Application Providers for each deployment. Each of these will have its own Application Deployment which can be related to a Technology Product Build that accurately reflects the 3-tier architecture.
Let me know some about these deployments
Jonathan
Essential Project Team
