As one of the first startups dedicated to managing Docker containers, ClusterHQ is trying to help IT organizations get in front of a Docker container sprawl issue before it happens. To that end, ClusterHQ this week announced support for the Fig file format that Docker created to simplify the management of container images across multiple servers and data centers.
ClusterHQ CTO Luke Marsden says that as developers move to quickly embrace Docker containers as an alternative to traditional virtual machines, it’s already apparent that not only will the number of Docker containers soon dwarf the number of virtual machines, but those containers will be distributed across multiple data centers in and out of the cloud. In the case of the Flocker management platform developed by ClusterHQ, those public clouds include Amazon Web Services, Rackspace and Digital Ocean. That’s critical, says Marsden, because it also enables IT organizations to deploy stateful applications using Docker containers in a distributed computing environment.
Where those Docker containers will actually run inside those data centers is an ongoing debate, though. One group of Docker proponents argues that containers eliminate the need for virtual machines altogether. Others argue that running Docker containers on virtual machines provides greater isolation, which in turn enhances security. Yet others contend that it’s more efficient to deploy Docker containers in a platform-as-a-service environment.
Marsden says ClusterHQ is betting that many IT organizations will want a management framework specifically optimized for Docker containers versus relying on existing management platforms that were designed for traditional virtual machine environments.
Regardless of the deployment approach, millions of Docker containers have already been created. Marsden says that by using Fig, IT organizations are not only going to find it simpler to manage Docker containers, they will also see higher server utilization rates. As a result, it may be only a matter of time before both developers and IT operations folks actually find a way to come to an agreement as to what the next best approach to virtualization in the data center actually is.