Plan and Deploy Azure Stack Hub
What the exam expects
Azure Stack Hub is a connected or disconnected private cloud appliance delivered as an integrated system. Planning starts with the capacity and scale unit model: a scale unit is the unit of expansion, and its node count and hardware SKU are fixed by the OEM partner. Understand the difference between connected (telemetry, marketplace syndication, Azure AD identity, billing) and disconnected (no internet, ADFS identity, offline marketplace, capacity-based billing) deployments.
Know the deployment prerequisites and decision points: identity provider (Azure AD vs ADFS), Azure registration and billing model, certificate requirements (the PKI certificate set: deployment, portal, admin portal, ARM, graph, blob, table, queue, Key Vault, and the wildcard/CA chain), DNS zones and delegation, time sync, and network design including the BMC network, public VIP pool, and infrastructure/private networks.
Traps to watch
- Confusing the privileged endpoint (PEP, used for OEM/first-party operations) with the administrator portal.
- Assuming Azure AD is required; disconnected systems use ADFS and cannot use Azure AD.
- Mixing up region (the Azure Stack Hub stamp) with Azure regions.
- Forgetting that the public VIP pool must be sized for both tenant and infrastructure services.
- Certificate questions that ignore the internal vs external namespace split.
Rule of thumb
If the scenario mentions "no internet access" or "classified network," the answer almost always shifts to ADFS identity, capacity billing, and offline marketplace. If it mentions "hybrid" or "single sign-on with Azure," it points to Azure AD and pay-as-you-use billing.
Worked example
A customer needs a disconnected stamp for 500 VMs with no internet. Plan: ADFS identity, capacity-based billing, offline marketplace, and size the scale unit so usable capacity after infrastructure reserve covers 500 VMs plus growth headroom.