;

Modelling Different Technology Types

7 min

Physical Servers

Modelling Physical servers are about creating Technology Nodes in the architecture – you should NOT create a Hardware Instance for the hardware of the node in question.

Hardware instances should be used to capture 'interesting' hardware items that exist on a Technology Node, e.g. additional network interface cards, and attached hard disk drives that you require to capture.

The process for modelling physical servers is:

  • Create a Technology Node for the physical server and define what type of technology this node is using the Technology Product field.

  • Create Technology Instances, e.g. Infrastructure Software Instances for things like Operating System, as required and deploy them to the Node.

Virtual Servers

While virtual servers appear to all the systems that run on them as real servers, virtual servers require slightly different treatment than physical servers.

To model a virtual server:

  • Create a Technology Node for the virtual server and deploy Technology Instances to it as required

    • Do NOT deploy any Hardware Instances – any physical attributes, CPU, Memory, or disk space should be defined against the Node

  • At some point create a Technology Node for the physical server that is hosting the Virtual servers

    • Deploy Technology Instances as required to describe that node.

    • Include the virtual server Nodes created in the set of contained Technology Nodes for the physical hosting server.

Appliances

Appliance technology is becoming more and more popular. This article describes how we recommend you capture such things in your architecture model.

To model the use of an Appliance, e.g. Integration Appliance or Database Appliance:

  • Create a Technology Component for the specific type of Appliance (e.g. Integration Appliance) if one doesn't already exist

  • Map this to Technology Capabilities in terms of what it is that the appliance does. For example, do NOT include things such as software platform services (e.g. Application Runtime Services), which are packaged into the appliance platform – for the Integration Appliance, it will be things like Transformation Services, Messaging Services etc. a Database Appliance will realise Structured Data Management Services

  • If a product can do more than just the Capabilities that the Technology Component that you've defined provides, then still map the product to the relevant Technology Component (e.g. Cast Iron as Integration Appliance)

  • However, at the Product-As-Role level, you can show additional things that a particular Appliance product can do, e.g. a product that is an Integration Appliance but can also act as a J2EE Application Server to run user-defined Java applications

Clustered Environments

Clustering is a physical architecture decision that we capture in the Physical Technology Layer. How you model it depends on how the clustering technology works.

Ideally, the application and information deployments are unaware of the physical nodes that make up the cluster.

To model a cluster where the application and information are deployed just to the cluster node:

  • Define a Technology Node for the cluster as a unit

  • Deploy Application Software Instances, Information Software Instances etc. to this Node as required

  • Define a sub-Technology Node for the actual physical hardware devices that are used to make up the cluster, e.g. the servers

    • Deploy the Hardware, Infrastructure Software instances as required to these sub-nodes

In scenarios where the application and information instances must be manually deployed to each node in the cluster:

  • Define a Technology Node for the cluster as a unit

  • Deploy Application Software Instances, Information Software Instances etc. to this Node as required

  • Define a sub-Technology Node for the actual physical hardware devices that are used to make up the cluster, e.g. the servers

    • Deploy the Hardware, Infrastructure Software instances as required to these sub-nodes

    • Deploy Application Software and Information Store instances to the sub-nodes.

Updated 31 October 2023

Contact Us