Microsoft Azure Arc Updates Go GA, Gain ML Capabilities

Microsoft continues to update its overarching Azure Arc platform that is becoming a central point for the cloud giant’s Kubernetes efforts. The latest updates were announced as part of the cloud giant’s Ignite event this week and also counter similar moves from its hyperscale rivals. Microsoft continues to update its overarching Azure Arc platform that is becoming a central point for the cloud giant’s Kubernetes efforts. The latest updates were announced as part of the cloud giant’s Ignite event this week and also counter similar moves from its hyperscale rivals. The biggest move was the push of its Azure Arc enabled Kubernetes platform to general availability. This basically makes what had been a preview available to all customers. Microsoft launched that preview last spring. It explained that Azure Arc enabled Kubernetes allows users to manage Kubernetes clusters across different infrastructure environments. The Kubernetes angle is central as it allows users to attach and configure Kubernetes clusters inside or outside of Azure. This includes data centers, edge locations, or across different clouds. The platform agent can be hosted by any Kubernetes distribution conformant with the Cloud Native Computing Foundation (CNCF) landscape. Microsoft had previously validated interoperability with Red Hat OpenShift, Canonical Charmed Kubernetes, and SUSE’s Rancher, and has now added VMware Tanzu Kubernetes Grid (TKG) and Nutanix Karbon. Microsoft is also adding its Azure Machine Leaning capabilities to Azure Arc. This will allow users to run their machine learning (ML) training on Kubernetes clusters running in Azure, Google Cloud Platform (GCP), Amazon Web Services (AWS), in edge devices, and on-premises through the Azure Arc-enabled Kubernetes platform. Once an ML agent is deployed, organizations can parse out access to data sets running in those different environments. Tangentially to the Azure Arc updates, Microsoft also moved its Azure Kubernetes Service (AKS) running on the Azure Stack HCI in a data center to GA. This allows users to deploy and manage containerized applications on Azure Stack HCI similar to running AKS in Azure, and it works on data center equipment from vendors such as Dell Technologies, Lenovo, Hewlett Packard Enterprise (HPE), Fujitsu, and DataON. Nick McQuire, chief of enterprise research at CCS Insight, noted in a statement that moves continue to highlight Microsoft’s cross-organization integration efforts. “Whether it’s a classic big data and machine learning path given the enhancements to Azure Machine Learning, or whether it’s a blueprint for using modern AI technologies of the future where areas like its Turing Model and massive AI models at scale come into play. Making these areas simultaneously relevant and accessible to customers in today’s climate is where Microsoft is placing huge bets,” McQuire wrote. Specifically to the Azure Arc updates, McQuire added that they show Microsoft moving toward where its customers want to go and also maintaining a feature set comparable to what is coming into the market. “The big trend we are seeing at the moment is customers wanting to leverage their existing hardware investments for their machine learning projects, and it is here where Arc’s service expansion is squarely positioned,” McQuire wrote. Customers need the flexibility to run ML across multiple environments such as in data centers, co-lo facilities, at the edge, or in the public cloud. While running Azure ML in multiple clouds will appear as another attractive benefit of this move, the reality is multi-cloud is not yet an operational reality for machine learning operations in most customers. The real advantage here is the hybrid capabilities, which are a real boost for Microsoft against the other clouds in AI, in my opinion.” Microsoft CEO Satya Nadella noted during the company’s most recent earnings calls that it had more than 1,000 customers using its Azure Arc platform. Microsoft unveiled the Azure Arc platform in late 2019. At a high level it extends the Azure control plane across rival cloud provider and on-premises locations. That cross-cloud management includes data and services running in virtual machines (VMs), containers, or Kubernetes, and also allows enterprises to enforce a single security posture across those data sources or applications. The platform is also Microsoft’s entry into the increasingly competitive hybrid-cloud management space. Hyperscale rivals Amazon Web Services (AWS) and Google both offer a similar platform in their Outposts and Anthos offerings, respectively. The updates come as Microsoft’s rivals have also been bolstering their similar platforms. IBM, for instance, earlier this week announced GA of its Cloud Satellite platform that allows users to run their IBM Cloud service on-premises or in edge locations managed through a single pane of glass in the public cloud. IBM had already struck a deal with AT&T to use the platform to bolster the reach and depth of its 5G network into enterprise markets.