Describe Azure Arc-enabled data services

  • 7 minutes

Contoso, a global entity, hosts multiple SQL Server instances with various versions in their on-premises environment. The IT staff wants to explore multicloud strategy, in order to use cloud service providers around the world. Azure Arc-enabled data services provides a single pane of glass and unified tools for management, deployment, and data sovereignty compliance in certain markets.

What is Azure Arc-enabled data services?

Using Azure Arc-enabled SQL Managed Instance, you can enroll all of your database instances located anywhere to Azure and then manage these databases in a consistent way. Azure Arc-enabled data services:

  • Enables you to be on an evergreen SQL instance.
  • Provides flexible scaling options. You can easily and quickly add or remove CPU and Memory from instances.
  • Takes advantage of cloud billing models.
  • Provides a platform as a service (PaaS)-like experience on-premises and across multiple clouds.
  • Comes with high availability, monitoring services, automated upgrades, patching, backup and restore services out of the box.

Management capability comparison between various SQL services

The following chart compares management capabilities by deployment model:

A chart that compares the management capabilities by deployment model.

Connected and disconnected modes

Contoso has operational capabilities in countries/regions where data is required to be on-premises due to regulations and low latency needs. This situation makes Azure Arc-enabled data services a perfect solution for them. Since the data is stored in a country/region, you can meet regulation compliance or data sovereignty requirements. The underlying latency issues are also resolved by enabling the databases locally or in close proximity to operations.

The disconnected mode of deployments doesn't automatically load any data sets back to Azure. For example, billing files, logs, and metrics aren't loaded automatically. Inventory information is only used for billing purposes. You would manually export and load the billing file to Azure portal at least monthly. Although disconnected mode doesn't automatically load the file, you still get the scaling, evergreen SQL, and other benefits from the solution.

In the direct connected mode, you get the full benefits of Azure Monitor including alerts, inventory, and management capabilities via the Azure portal. In the connected mode, deployment data is still processed on-premises and you can optionally upload logs to Azure. The inventory information is only a metadata representation of assets that you have.

SKU comparison

Azure Arc-enabled SQL Managed instance comes in two different types of SKU offerings.

  • Business Critical = Enterprise Edition
  • General Purpose = Standard Edition
Area Business Critical General Purpose
SQL feature set Same as Enterprise Edition Same as Standard Edition
CPU limit per instance Unlimited 24 cores
Memory limit per instance Unlimited 127 GB
High availability Contained availability groups over Kubernetes redeployment. Single instance, with Kubernetes redeploy and shared storage.
Read scale out Availability group None
Disaster recovery Available via failover groups Available via failover groups
Scale up/down Available Available
Monitoring Built-in available locally. Optionally export to Azure Monitor Built-in available locally. Optionally export to Azure Monitor.
Logging Built-in available locally. Optionally export to Azure Log Analytics Built-in available locally. Optionally export to Azure Log Analytics.
Point in time restore Built-in Built-in
AHB exchange rates for IP component of price 1:1 Enterprise Edition
4:1 Standard Edition		 1:4 Enterprise Edition
1:1 Standard Edition
Dev/test pricing No cost No cost

How does Azure Arc-enabled data service operate?

The Azure Arc-enabled SQL Managed Instance operates on any supported Kubernetes cluster and hardware.

Azure Arc-enabled data service is deployed and managed via an agent-based solution. The agent that's deployed to the environment is called the data controller. This agent acts as the Kubernetes orchestrator for the solution. To run the services in direct connected mode, you must:

  • Have access to a Kubernetes cluster that is sized appropriately for your workload.
  • Use a supported Azure CLI version (>= 2.16.0 and <= 2.29.0).
  • Add the following arcdata and associated k8s CLI extensions:

    • arcdata

    • connectedk8s

    • customlocation

    • k8s-configuration

    • k8s-extension

      To add any of the above extensions, run:

      az extension add --name <extension>

You must also:

  • Meet the appropriate network requirements for connecting a Kubernetes cluster to Azure Arc.
  • Provide a ReadWriteMany (RWX) capable storage class for backups. If no storage class is specified for backups, the default storage class in Kubernetes is used. If this storage class isn't RWX capable, the Arc SQL Managed Instance installation may not succeed.
  • Register providers for Azure Arc-enabled Kubernetes:
    • Microsoft.Kubernetes.
    • Microsoft.KubernetesConfiguration.
    • Microsoft.ExtendedLocation.
  • Create a destination Resource Group in your Azure subscription.