As Microsoft® approaches the formal release of Azure™ Stack during the coming months, they recently released that which is said to be the third and final technical preview. Azure Stack is a way to deploy Azure in an organization’s own datacenter. This is appealing to many, as it will bring the same Azure tools and functionality on-premises, either exclusively, or in a hybrid environment. This is particularly important to developers, as the APIs, Portal, PowerShell™ and DevOps tools will be identical whether using Azure or Azure Stack.
Unfortunately, deploying on-prem doesn’t mean an organization can use their existing hardware. The original offering was expected to give users a choice of hardware, but Microsoft later decided it will be sold as an “integrated system”, bundled with hardware from Dell™, HP™ or Lenovo™ (Cisco® will be added later this year). Services and support will also be included. Microsoft defends this decision as a way to ensure consistency and reliability, since they won’t have to support a wider range of systems. When we consider the fast pace of enhancements and updates to Azure, this appears to be a valid approach, but it is being met with considerable resistance.
Pricing for Azure Stack is also similar to that of Azure, in that there are no upfront licensing fees and they will employ a pay-as-you-use model. In most cases, services will be metered on the same units as Azure, but the pricing will be lower since the Stack customer is using their own facilities and hardware. Customers who are unable to have their metering information sent to Azure will be offered a fixed-fee “capacity model”, based on the number of cores in the system.
Microsoft has promised availability “mid-2017”, although it’s unclear whether all hardware integrations will be available simultaneously.