Container Toolbox Grows With Azure Support
Microsoft moved this week to expand its Azure cloud container service with the release of extended native support for the Kubernetes cloud orchestrator along with upgraded support for Mesosphere datacenter operating system (DC/OS).
The software giant (NASDAQ: MSFT) rolled out the cloud-native container initiatives during a conference in Seattle this week sponsored by the Cloud Native Computing Foundation. The "preview" release means the Azure cloud now supports Kubernetes, DC/OS and Docker Swarm for container orchestration.
Meanwhile, other tools unveiled at the Kubernetes conference aim to improve management of container and legacy workloads.
Microsoft also released an upgraded version of DC/OS that includes virtual networking capabilities along with job-scheduling and container orchestration based on Marathon, the container orchestration platform for Apache Mesos and Mesosphere's DC/OS. The upgrade also includes faster installation of platforms such as the enterprise version of the DataStax database built on Apache Cassandra.
Since announcing support for Docker Engine last year on Windows Server 2016, Microsoft has been moving to expand cloud native support for container orchestration. Corey Sanders, director of compute for Microsoft Azure, asserted in a blog post that Azure is the first public cloud platform to support all three leading container orchestrators.
"We are seeing organizations of every size move their container-based solutions from dev/test environments to production in the cloud," Sanders added.
Microsoft also said Monday (Nov. 7) it would release to the open source community the code for its Azure Container Service Engine. Sanders said the GitHub project would share details on deploying DC/OS, Docker Swarm, and Kubernetes and orchestrating containers on both the public cloud and on the Azure Stack.
In July, two years after it announced support for Kubernetes on Azure, Microsoft hired Brendan Burns, a lead developer for Kubernetes at Google (NASDAQ: GOOGL), to oversee its Azure Resource Manager service. Burns helped launch the Kubernetes container orchestration initiative along with two other Google engineers.
In a blog post, Burns noted that the introduction of Kubernetes as a supported container orchestrator on Azure means a functional cluster can be spun up quickly via the Azure portal or using a Python-based command line tool. The Kubernetes cluster is then integrated with other Azure cloud resources.
As public cloud vendors such as Microsoft expand support for a range of container orchestration options, management tool vendors are attempting to plug gaps in the container ecosystem as the breadth of supported infrastructure and application components expands. With that in mind, workload manager Univa released a tool on Monday to handle workload scheduling and policy management for containers running on Kubernetes.
Chicago-based Univa also unveiled "mixed workload support" for running legacy and container workloads with a Kubenetes cluster, meaning applications such as Spark and Hadoop could run alongside micro-service-based applications sharing the same server infrastructure.
That capability, industry analysts note, would allow enterprises to achieve efficiencies by pooling container and traditional computing cluster, and then managing those resources as a single entity.