Azure Arc — Hybrid and Multi-Cloud
With Microsoft becoming more focused on the adoption of hybrid technologies in the cloud, Azure Arc may be something you have already heard about. If not, this will be a quick rundown on how Microsoft are raising the bar for multi cloud offerings.
What is Arc?
Throughout the Ignite sessions, Microsoft highlighted that whilst their customers had been enjoying the services that Azure had to offer, there was still some gaps in their automation practices due to some services being hosted on separate cloud providers or on on-premise solutions. This is where Azure Arc comes in, Arc allows for applications to be deployed and managed using the Azure suite anywhere. Microsoft have announced they’re aiming for on-premise and multi-cloud applications to be capable of being managed via the Azure control pane in a single pane of glass fashion. The three main offerings of Arc are the ability to manage VMs running outside of Azure, deploying and maintaining Kubernetes clusters outside of Azure as well as data services such as SQL and PostgreSQL.
At MS Build and Ignite this year, Azure announced that the VM and Kubernetes side of this offering had made it into GA and is available for Production workloads, in addition to Arc enabled data services and Arc enabled machine learning being announced for preview.
Hybrid Virtual Machines
Arc for VMs allows for instances that are housed outside of Azure to be connected to the Azure control pane and represented as a resource within the portal. The instances are connected to the pane via the installation of the Azure Connected Machine Agent which will have versions available for both Windows and Linux. Once the agent has been connected, the instance is now considered hybrid and will send regular heartbeat alerts to Azure. As long as the instance remains healthy, it is available for configuration updates.
As Arc enabled Virtual Machines can be housed in resource groups with Azure native Virtual Machines, they can all be managed under the same policy groups, regardless of where they are hosted. Meaning the same level of governance can be applied to all machines which helps to reduce the management complexity of multi-cloud environments.
Azure provided examples from their customers demonstrating the scale Arc is capable of, a particular example boasted Arc enabled services running in 3 global data centres as well as infrastructure spanning over 124 countries. As with native instances, Arc enabled virtual machines are able to monitor for changes within the OS. Any change in state configuration will be reported to the Azure Control Centre and each VMs level of compliance with its policies will be recorded.
In addition to policy and configuration management, a whole host of tools within the Azure management pane such as Azure Cost Management, Azure Monitor and Azure Resource Graph will be available for use with Arc enabled VMs.
Hybrid Kubernetes
In a similar fashion to VMs, Kubernetes clusters that are housed outside of Azure can now be represented as resources within the Azure portal, regardless of where they are hosted. This includes AWS EKS, Google Kubernetes Engine and on-prem solutions. Azure is aiming to allow engineers to leverage the Azure control pane across a number of Kubernetes distributions with flavours from a number of vendors.
If an all Azure solution is the preference, Arc enabled Kubernetes is now natively integrated with AKS on Azure Stack HCI. Clusters are connected in the same way as Virtual Machines with the use of an agent. Once the agent has been configured and the cluster is considered hybrid, it is then capable of receiving GitOps based configuration updates and can send telemetry data to Azure Monitor.
With Azure Arc also enabling Azure policy across all of your clusters, policies can be authored to ensure that the infrastructure is not only configured but continuously watched to ensure they remain compliant. The Arc enabled cluster agent will watch the repository and recognise changes made to either configuration or policy and immediately begin synchronising the cluster with the latest changes.
What’s Next?
Arc Enabled Data Services
Announced in Autumn, arc enabled data services remains in preview, but Azure noted it’s been generating a lot of interest with customers with hopes of making it to GA later this year. Arc enabled data services allows for the management of data assets running on premise alongside those running in familiar tools such as Azure Portal and Azure Data Studio.
Arc Enabled Machine Learning
With this MS Ignite, Azure announced that Arc enabled machine learning would now be available in private preview. The aim for this again is allowing the use of the Azure machine learning tool sets to build, deploy and train models no matter where the data lives. Over the coming months we’ll hope to see more info on the experiences with Arc enabled ML.
Interesting Use Case
Throughout the talks Azure noted some interesting use cases for the Hybrid Virtual Machine and Kubernetes offerings. As opposed to just be being useful with regard to the management of on-prem and multi-cloud environments. Some teams have been using Arc enabled services to scope and gain telemetry on environments being prepped for migration. Allowing for an accurate representation of what the environment will look like on Azure services and also allows for tools such as Azure Cost Management to be utilised to give an accurate representation of the running costs.
