Containers are being touted by many as the next big thing.  What are containers? Containers are a form of lighter weight virtualization that allows the sharing of the operating system among the host and all guest operating systems.  In short, containers run in the same operating system as the host.

Google is the most famous company currently using containers.  Google stunned the virtualization world in 2014 when they noted that they had completely converted to containers and that they stood up 2 billion containers each week.

This is important because containers enable higher performance, particularly with respect to heavy I/O workloads, than traditional virtual machines (by traditional, we mean virtual machines operating in VMware vSphere, Microsoft Hyper-V, and other hypervisor-based virtualization environments.)  This higher performance means that tier 1 applications can run faster with less overhead in a container – and it means that the consolidation ratio (the number of virtual machines a host server can support) can be much higher than in traditional virtualization.

Another reason that containers have an advantage over traditional virtualization is visibility.  IT staff are able to have deeper visibility into container-based architectures because there is less abstraction, i.e., there is one operating system to monitor and manage rather than having to monitor and manage the host operating system and then each operating system for each virtual machine.

Is there a downside to containers?  Of course.  The biggest downside is that Windows is the most common virtual machine operating system, and Windows doesn’t support containers.  Thus containers are excellent for Linux-based applications but are not supported for Windows (at least yet.)

As you’re designing next generation data centers, you want to ensure you remain agile and adaptable enough to support containers architecturally

This is part 12 of an on-going series. Part 1: virtualization isn’t the next big thing (NBT) because it was the last big thing (LBT); part 2: data center IP traffic growth; part 3: data center IP traffic sources; part 4: cloud workloads; part 5: large data centers and administrator to server ratios; part 6: strategy overview – adapt or be crushed; part 7: automation, agility, adaptability;  part 8, automation vs agility and adaptability; part 9: virtualizing everything; part 10: bare metal cloud; part 11: bare metal cloud performance.