X

Hybrid datacenter is Docker’s latest open source goal

Shutterstock

Docker, Docker, Docker — everyone’s talking about Docker. A surprising fact is that Docker wasn’t on anyone’s roadmap in 2014, but was on everyone’s roadmap just a year later. A lot of people are calling it the most phenomenal, amazing, and speedy adoption of technology that they’ve witnessed during their careers. But this sudden rise to fame hasn’t been a lonely one. Docker has shared this amazing growth spurt with an entire ecosystem of tools and services.

A part of the stack

It’s in our nature to think that the people at Docker can’t be too happy about this, but Docker CEO Steve Singh maintains that they “couldn’t be happier.” He also points out that Docker was open source from the start and always expected other companies to come and build on its success. The success Kubernetes has had with container orchestration, for example, is almost as big as the Docker story itself, and though they’re both different pieces to the puzzle that is a stack, it’s hard not to think of the two as competition for each other.

This is probably why during an interview with GeekWire, Steve Singh was asked two pretty blunt questions. The first one was “how do you feel about the success of Kubernetes,” and the other, “how do you make money on open source software?” That’s like asking Superman how he feels about Batman or asking Coke how they feel about Pepsi. It is without a doubt the burning question or the elephant in the room, and as expected, COO Scott Johnston was quick to jump in and point out that Kubernetes is just a part of the stack.

Countless opportunities for everyone

That, interestingly enough, is exactly what the folks at Kubernetes think about Docker, especially since container engines are now pluggable and you can use anything from rkt to clear containers. To be fair, Docker is also part of the Open Container Initiative, even if it’s only to enhance their perception according to some people. The CEO and COO duo also explained their genuine lack of concern over the steady rise in Kubernetes adoption by reiterating that the market in question here is so vast and deep that it’s literally just the surface being scratched at the moment.

With regards to the second elephant in the room and how do people actually make money off open source software. While open source is driving adoption and standardization and allowing companies like Docker to embrace the ingenuity of the world, everything isn’t open source. While open source is a great way to make use of the unlimited supply of intellect on the planet and continuously innovate, CEO Steve Singh reminds us that “all innovation is not done in open source” and there’s a lot of proprietary code at Docker, too.

Hybrid datacenter is the future

Another place Docker is paying close attention to is the hybrid datacenter. As enterprise companies continue to migrate to the cloud, many are looking to achieve hybrid cloud capability to deploy their applications. Survey data from IDC shows 73 percent of enterprise cloud adopters use a hybrid cloud strategy as it’s the best approach for cost, reliability, scalability, and security. Their infrastructure is typically made up of an on-premises datacenter combined with one or more public cloud providers in the mix. COO Johnston says 75 percent of their deployments are hybrid. The reason this setup is so popular is because there are a lot of practical applications and services that just have to be done on premises, some for security reasons, and others because they actually perform better on premises. At the same time, these organizations can’t afford to ignore the benefits of the cloud for scalability, cost, and agility. This is why Docker believes the hybrid datacenter is the future, and why they also believe they will be the ones to power it.

So this hybrid datacenter is where Docker is focusing their attention now, and people who buy IT software love it because it’s open source. IT buyers have long memories is something else Singh said in his interview, which is true because license renewal time is no fun when you’re the buyer. Everyone is “lock-in” wary while Docker, on the other hand, is something that companies are asking for right off the bat. Adoption is through the roof and with it comes a lot of customers willing to pay for some extra lines of code.

Swarming the opposition

In a recent survey of more than 200 IT buyers polled by CoreOS, 33 percent said they were interested in containers for efficiency, while 32 percent said they were interested in hybrid and cross-cloud integration capabilities. Surprisingly, Docker Swarm was cited as the most popular orchestration platform overall (36 percent), ahead of report sponsors CoreOS (27 percent), Kubernetes (22 percent) and Mesos (14 percent). Could this be why the CEO and COO of Docker both seem to be unimpressed with the success of Kubernetes and quite confident in Swarm’s capabilities moving forward?

They’re definitely going to keep pushing Swarm and Docker 17.06, which is the first version to be completely based on the Moby Project,  and also features a number of new features in swarm mode. One of those features is a configuration object for swarm mode that allows users to securely pass along configuration information in the same way they pass along secrets. Another interesting addition is Docker events that helps users get real-time event information from Docker in Swarm mode. Additionally, Docker service logs moved to stable release, which means that users can now get consolidated logs for an entire service running on Swarm.

Multistage builds

One of the highlights of version 17.06 announced at Dockercon 2017 is the launch of multistage builds, which is now going to be a standard feature on all versions. Multistage builds allow users to build cleaner, smaller Docker images using a single Dockerfile according to the blog post and work by building intermediate images that produce an output. This means you can now compile code in an intermediate image and use only the output in your the final image. This results in a substantially smaller image size, which could be a sign of things to come, especially since Steve Singh forecasts a data shift toward the edge.

Docker now also supports metrics through an API endpoint so users can expose Docker’s metrics to external monitoring tools that can be plugged in. However, we’re supposed to keep in mind that even though these metrics plugins are available on nonexperimental daemons, the labels are still considered experimental and could change in future versions.

The future is on the edge

Overall, things seem pretty optimistic for Docker right now, they have a new CEO and they’ve set themselves up in a great position to start making some serious money, perhaps even by powering the hybrid datacenter they envision many enterprises will soon have. It’s true they’re not making the kind of money they should be considering they’re valued at over $ 1 billion, but like their new CEO mentioned, it’s a really huge market. “Fifty percent of data will be created at the edge of the network over time” were his exact words, though he didn’t say exactly how much time. If we are to go by Docker’s track record, however, time isn’t really a concern, though only time will tell if this move to the edge and the hybrid datacenter is going to be powered by Docker containers or something new.

Photo credit: Shutterstock