Lesson 4: Virtual and Cloud Design Lab

Core path: 42 minutes

You have one physical lab workstation and several possible ways to deliver computing environments.

Your job is to choose the right model before configuring anything.

Host specification

CPU: 8 logical processors with virtualization support
RAM: 32 GB
Free SSD space: 300 GB
Network: Ethernet + Wi-Fi

Requirement A: Windows practice system

Needs:

• complete Windows guest OS;
• 8 GB RAM target;
• 80 GB virtual disk;
• approved external network access;
• snapshot before a major configuration exercise.

A full VM is a plausible fit.

Requirement B: Linux practice system

Needs:

• complete Linux guest OS;
• isolated initial network;
• 6 GB RAM target;
• 60 GB disk;
• cross-platform training from the Windows host.

Choose an appropriate VM network mode and explain why.

Requirement C: packaged web tool

Needs:

• isolated application/runtime dependencies;
• no separate full guest desktop required;
• repeatable deployment.

Decide whether a container concept is a better fit than a full VM and explain the architectural difference.

Requirement D: centralized desktop access

A fictional organization wants users on lightweight endpoints to reach centrally hosted desktop sessions.

Sketch the VDI dependency path:

endpoint
→ network
→ authentication/access
→ VDI infrastructure
→ desktop/session

Identify two failure points outside the endpoint itself.

Part A: resource budget

Create a host/guest resource table.

Do not allocate every byte of RAM or storage to guests. Reserve capacity for the host/hypervisor and operating headroom.

Part B: security, network, and storage requirements

For each virtualized workload record:

security boundary:
network requirement:
storage requirement:
backup/export requirement:
what a snapshot would and would not protect:

Part C: cloud architecture choice

Choose a deployment/service model for each fictional need:

1. Public SaaS collaboration app.
2. Private organization-only cloud workload.
3. Hybrid design using private resources plus public capacity.
4. Community cloud shared by organizations with a common requirement.
5. IaaS VM requiring customer control of the guest OS.
6. PaaS application deployment where the runtime is provider-managed.

For at least two choices add one characteristic that matters:

• shared vs dedicated resource;
• metered utilization/ingress-egress;
• elasticity;
• availability;
• file synchronization;
• multitenancy.

Fault injection

Your partner selects one:

• host RAM exhausted by over-allocation;
• VM virtual disk fills;
• virtual NIC disconnected;
• VM attached to host-only network when external access was expected;
• VDI endpoint has internet access but cannot reach the VDI service;
• containerized application starts but a required host/network dependency is unavailable;
• file synchronization propagates an unwanted deletion;
• cloud egress use exceeds the expected metered plan;
• snapshot exists but host storage fails.

Trace the failure to the narrowest supported boundary.

Deliverable

Submit:

• virtualization-model decisions;
• host/guest resource table;
• security/network/storage map;
• VDI dependency path;
• cloud deployment/service-model comparison;
• one troubleshooting ticket.

Success criteria

A passing design explains why the chosen virtualization/cloud model fits the requirement and names the dependencies that still have to work.