In recent months, the rise of containers as a lighter-weight alternative to virtual machines has emerged as a major challenger to VMware, the dominant provider of virtual machines inside the enterprise. In response, VMware has already announced that it will support the management of containers from Docker alongside its virtual machines, a position the company just extended to include integration with the open source Kubernetes orchestration software project being led by Google, Mesosphere and Pivotal, a sister company of EMC.
But while VMware concedes that there has been a sharp increase in the usage of containers, Kit Colbert, VMware vice president and CTO for cloud native apps, says that VMware is betting that the vast majority of those containers will be running on top of virtual machines rather than physical servers.
Because of container security concerns, Colbert says most IT organizations are going to want to run various containers in isolation from one another. It’s already been proven that a Docker container has a vulnerability that allows that container to take over an entire machine, says Colbert.
As for alleged performance concerns involving the running of containers on top of virtual machines, Colbert says VMware tests show that at least within VMware environments, performance of a container running on a physical machine versus a VMware virtual machine is minimal.
For the most part, usage of Docker containers is currently limited to application development environments. But as Docker containers inevitably move into production environments, VMware is betting that IT organizations will want a unified approach to managing not only Docker containers and existing virtual machines, but also a new generation of Project Fargo virtual machines from VMware that make use of advanced cloning techniques to reduce the size of a virtual machine footprint.
It’s still unclear what impact all this innovation is about to have on the enterprise. But the one thing that is clear is that there will soon be a lot more diversity in the data center as IT organizations begin to figure out what types of application workloads run best and where using at least three distinct types of virtualization.