Conceptual design

The conceptual design provides a high level of conceptual diagrams for the solution, using the data collected from the current state analysis of the existing environment (if existing), the application requirements, and the business needs and goals. All the data collected during the assessment is categorized into different categories:

• Requirements: This provides the business requirements that the designed solution must meet.
• Constraints: The conditions that provide boundaries to the design.
• Assumptions: Lists the conditions that are believed to be true but are not confirmed. All assumptions should be validated before the deployment.
• Risks: Factors that might have a negative effect on the design.

All business requirements, assumptions, and constraints must be used to support design and implementation decisions suited for mission-critical applications, considering also how risk can affect design decisions and how it can be mitigated.

From this analysis, a solution is usually defined where you should decide the number of vSphere objects, like data centers, clusters, vCenter Server, and so on.

Depending on the size of the solution, you may have some different deployment types:

Table 2.1: Different deployment types

A typical conceptual design is limited to a general high-level architecture, usually with the data centers, and the interconnection between them and maybe with the cluster at an abstraction level (without the details of the hosts, the vCenter Server, and other infrastructure elements). For example, for a single site, using the suggested deployment, the concept could look as follows:

Conceptual schema

The usage of a management cluster, usually with at least three nodes (but which can potentially also start with two), is to provide security and resource isolation of management workloads from production workloads, potentially with different roles and scopes for the administrators. A management cluster is also used to simplify management, to upgrade procedures, for troubleshooting, and for dependencies. The edge cluster is a concept more related to an NSX design, where lots of virtual appliances (called NSX edge) should be deployed to provide several network services.

The production cluster or payload cluster could be one or more clusters for the different workloads. They could be differentiated with a different type of cluster, related to the life cycle of the workloads (like test or dev or productions), or isolated clusters for Demilitarized Zone (DMZ) purposes, or island clusters with small groups of ESXi hosts that can run workloads with special license requirements.

The usage of DMZ dedicated clusters is considered old-school, surpassed by the new network security and micro-segmentation features provided by NSX. But without NSX, it is still possible to have dedicated separated network cards for segregate networks on different physical switches, or to use different VLANs on the same physical switches.