Hybrid Containers: The Middle Way between VM Technology and Disruptive Upheaval
By Jeff Farinich, Glen Tindal, Mark Thiele and Mark Campbell
The Containerization Revolution is a hot topic these days that’s garnering a substantial amount of ink in the IT press. Conversely, the stalwarts of traditional virtualization (wow, when did something as young as virtualization become traditional?) view this trend as just a modest innovation to be subsumed by VM technology and not as a disruptive upheaval. But is there a third option?
Containers provide several benefits beyond those offered by virtualization. Containers improve the speed and ease of application deployments through the use of an image format that allows seamless portability between both on-premise and cloud-based environments. Containers isolate processes and encapsulate dependencies, allowing containers with conflicting dependencies to run on the same host. Containers also enable the next generation of scalable application architecture referred to as “micro-services,” in which small, discrete, modular services simplify development and allow computing resources to be precisely allocated to suit performance requirements.
The use cases for containers emanate primarily from the development process as they improve development team collaboration by allowing the entire team to work more efficiently on the same code base. Containers can also improve software quality by eliminating inconsistency between environments, which lessens the burden on IT operations staff and container build jobs, reduces configuration complexity, and allows for continuous integration. Nonetheless, even with these valid use cases, the adoption of containers can stall due to organizational awareness, infrastructure maturity and orchestration, monitoring, and automation limitations.
Despite these challenges, containerization is picking up support faster than any technology in history. In fact, one can compare the maturity of the Docker era container after three years very favorably to the spread of VMware virtualization after eight years.
The sea change in enterprises, big box vendors and cloud providers to demonstrate a leadership position in containers means that the market has the necessary support to continue building a strong ecosystem and enable adoption. Strong adoption interest is easily demonstrated by more than 500 million downloads.
The potential for disruption in the infrastructure market crosses legacy virtualization players and cloud providers alike. Yesterday’s fresh flower (VMware) is now legacy, and cloud providers must scramble to demonstrate why their infrastructure won’t become a transient stop on the way to cheaper or faster infrastructure.
Today the question is, “Where will this stop?” Frankly, it won’t. Virtualization disintermediated hardware suppliers in ten years and containers will do the same to virtualization in five. The end state of full market penetration will occur in less than ten years, when 80% of workloads will run in containers. Now the real question is, “What next?” It’s possibly micro kernels.
“Hybrid” is a term the IT Industry is batting around now more than ever: Hybrid Cloud, Hybrid Storage, Hybrid Computing and Hybrid WAN to name but a few examples. There is yet another example of hybrid evolving at the very edge of the computing universe: Hybrid Containerization.
As most are aware, the merits of virtualization vs. containerization are hotly debated. The virtualization camp is forwarding a mature and slowly evolving migration of VMs into the cloud, while the containerization pundits are focusing on the added benefit of decomposing applications into smaller entities, abstracted from the underlying operating system.
Hybrid Containerization seeks the middle ground of an “and” vs. an “or” world. Virtual machines and containers cannot just coexist – they can play together elegantly. Virtualization and its underlying virtual machines are a staple of today’s IT infrastructure, business processes and workflows. As a result, IT organizations have invested software, processes and budget into the automation of virtualization. Would it make sense to cast this investment aside, even for all the benefits that containerization offers? Hybrid proponents say, ‘Probably not.” Instead, a more productive option is to implement containers inside virtual machines. That leverages the current VM ecosystem to move, add, change and delete virtual environments and thusly create a hybrid container environment.
For now, Hybrid Containerization offers enterprises a middle road to leverage current investments and virtualization’s mature ecosystem, while opening the door for containerization. Running containers in a virtualized environment is a good first step down a road that will, undoubtedly, lead to an all-container enterprise. But even this too is just a stop on the ever-evolving journey of IT. Enjoy the trip.
About the Co-Authors: Jeff Farinich, Glen Tindal, Mark Thiele and Mark Campbell
The co-authors are all members of the Trace3 Innovation Council. Jeff Farinich is IT Manager of Infrastructure Operations at Schools First Federal Credit Union. Glen Tindal is Cloud Computing Principal for Capstone Consulting. Mark Thiele is EVP of Data Center Technologies at Supernap. Mark Campbell is Director of Research for Trace3.