Advertisment

2016: Some Un-Future Tense: Containers

It was quite a year with the third gear towards segments that were alien so far: Big-sized Incumbents, Enterprises, and even the so-called dyed-in-the-wool rival - the VMs

author-image
Pratima Harigunani
New Update
ID

INDIA: In many aspects the year 2016 was a cold-turkey moment for indifference towards technologies and paradigms that we had lazily tucked way ahead in the future.

Advertisment

If we thought that AI will take time to rub its eyes and wake up from the cradle, the Go boardgame challenge happened where Google’s AlphaGo battered not just the game but many cynics and assumptions on strong-level AI. If we thought that Bots are going to remain ensconced in sci-fi movies only for some more years, we saw major banks like ICICI, Yes Bank and companies like Meru Cabs giving them a not-so-hesitant hug this year. If we were not ready for the VR and AR twins yet, Pokemon Go left us no choice but to chase pixies of fantasy in the real world.

Surely, in our ever-amorphous universe of technology, there’s no telling what shifts, flips and ticks and when. An year back we were trying to get our heads around a new word called ‘Containers’ and 2016 completely hammered the notion that this technology is about a future that is on its way.

This year made it unequivocally clear that containers are already shipping with ferocity into the turnstile of enterprise playgrounds.

Advertisment

Then

Ironically, they breathed silently like a wallflower in this very industry since the days when they were a core feature of Linux and even when 2013 onwards, open source company Docker put them on the centre table in an attempt to make them easy-to-use, portable and appealing for developers; they were a fuzzy novelty. Unlike VMs, and other alternatives, Docker containers ensured that there were no guest OS environment variables or library dependencies to manage. There was enough company for Docker, in the form of Google’s Kubernetes that was trying to smoothen the edges on this concept on the management or orchestration side.

Then there was Rocket from CoreOS; Canonical Ubuntu Linux-based OS, with the LXD containerization engine for Ubuntu; Mesos; and other aboriginals of the open source and developer-leaning world.

Advertisment

It was yet to be seen how and where containers make their way into production environments and become recognisable for enterprises and the big boys too.

Microsoft, Google, Amazon were the front-runners in this regard. They sensed the charm and punch of containers well and tried to hitch their wheels on to this wagon with their own hooks. Blox, an open source project from AWS, orchestration versions from Kontena and Mesosphere and integration of Swarm container in its platform from Docker were notable moves that signaled the shift this technology is taking.

But what is more hair-raising this year is watching the blue-blood of the industry jump in these new waters in a clear and un-poker-faced way.

Advertisment

Now

This year, we saw, among other passengers jumping on this ship, Microsoft Docker container features coming with the first release of Windows Server 2016 for pushing containers toward production in enterprise environments. When a company like Microsoft revises the Windows kernel to make room for open source technology, it leaves little room to doubt the impact that containers have created in the real world.

EMC Corporation also announced a hardware-agnostic solution in the form of RackHD Machine for open deployment of Docker solutions directly on bare metal, reflecting its integration with Docker Machine to manage the lifecycle of physical servers for the Docker Engine and future Docker solutions. IBM on the other side, rolled out the availability of IBM Containers within cloud platforms like IBM Bluemix Local and IBM Bluemix Dedicated instances. Intel was also seen sharing plans for making Kubernetes run better on its hardware.

Advertisment

But perhaps the most popcorn-worthy of these spins was the one from VMware, because obviously, as it has been generally perceived, containers are construed as substitutes for VMs (Virtual Machines) which, in their own might, happen to be gravy boats for virtualization majors’ dinner tables.

Yet, VMware has come up with vSphere Integrated Containers (VIC) and this re-affirms VMware's stance that containers are not rivals but complementary forces that can run inside virtual machines. They have maintained their confidence about the relative maturity of VM-centric ecosystem and the familiarity as well as robustness that VM environments offer for enterprises.

Interventions like Project Photon from VMware, Oracle’s orchestration tool and promised container service for its public cloud and Cisco’s drooling over the acquisition of ContainerX, leave little room for doubt that containers 'were' a developer-only novelty.

Advertisment

Meanwhile pioneers and other players like Flatpak, Subuser continue their evolution. We witnessed desktop management versions and acquisitions like Socketplane and Infinit that Dockers grabbed this year.

But an important transition was the hard fork that is now shaping up with the core technology. Recently when Docker’s core Containerd component's spinning off was announced, it signaled that now the main marrow would be governed under a separate community so that Docker can focus on the product side, while hackers and developers can continue fiddling with their projects and infrastructure innovation the way they have so far, without either side creating avoidable chaos.

With this slice of code i.e. the Containerd that was, so far, distributed in different parts of the Docker Engine, now a loose form is available as a stand-alone component for companies and even competition that want to absorb containers in their own manner.

Advertisment

Later

Despite all these developments and adoption leaps, the container world also saw the struggle that ClusterHQ, an early start-up in this realm gave in to. The notion and significance of state-ful-ness for the success of containers took a new proportion altogether when CEO Mark Davis revealed the shutdown in a blog. This also heralds Container 2.0 as the next era where stateful containers would be critical. Even Kubernetes 1.3 has added beefed-support for stateful containers, Docker has acquired Infinit for this stateful slant and Kontena is walking on that stateful applications track as well.

As always, what will matter from here is how much stability, statefulness, freedom of choice, flexibility and agility can be offered to enterprise users.

Containers have a lot of radical advantages, they can run all kinds of applications, can translate immediate gains in efficiency for memory, CPU and storage; besides showering developers with immense granularity and control.

Containers share the same OS kernel as the host, but VMs need separate OS instances so the efficiency card goes in favor of Containers. That software will run, no matter where it is deployed – is such a relief for developers now. They have all the dependencies and components necessary to run the desired software- be it files, environment variables, libraries, anything; and since the host OS lords over the container's access to CPU, Memory or other physical; a single container cannot consume all of a host's fuel.

Their easy and organic enablement of microservices makes the idea of monolithic application sound too obsolete and unnecessary. Developing applications, testing them and then accessing them or working on them becomes way easier with containers.

They ensure that, unlike as in case of VMs, here the only operating system on the server is just one host operating system and containers can talk directly to it, thus cutting down overheads to a large extent.

But then, VMs have an edge because the native old software being run at big companies is better moved to actual scenarios or cloud services like AWS or Microsoft Azure with VMs.

What Containers also face is the security fear because of the presence of one OS kernel – the advantage point here translating into a concern ironically. Even if people argue that because of the play of microservices, containers are harder to exploit; the surface area of one OS here instead of a hypervisor-based virtualization, looms large from a safety point of view.

There is also the need to make Containers more consumable, and equipping the industry with enough skills, tools and ways to operationalise the software and environment. That is what seems to be happening and if 2016 is a glimpse of what’s about to come, 2017 may be just the year when Containers stomp right into the trenches and become a strong ‘present tense’.

containers docker