Docker Inc.’s acquisition of a Spanish startup called Tutum yesterday is in keeping with Docker’s growth plan thus far, acquiring small firms such as SocketPlane and making full use of their talent, while partnering with larger firms such as Microsoft.
This is important, if you’re considering a possible experiment with containerized applications in your organization, and you’re concerned about the long-term health of the movement’s leading vendor.
The Universe Expands
Until this spring, Docker was the center of gravity for containerization — the radical re-compartmentalization of workloads into small, highly standardized units that can scale up or down simply by adjusting the coefficient that multiplies them.
That began to change, perhaps with necessity, after Docker made the bold move of launching the Open Container Initiative — an industry consortium that enabled key players to exploit containerization without relying so much around Docker Inc.
Because of OCI, no single company can acquire Docker Inc. and own the container ideal as intellectual property. On the other hand, Docker must now demonstrate its own long-term viability by offering competitive tools and services, including some that can be stacked against orchestration platforms like Google’s Kubernetes and Mesosphere’s Marathon in a fair comparison.
Until yesterday, Docker offered its own Docker Swarm as its orchestration platform, with the idea that organizations that need more complex or muscular systems could choose Marathon or Kubernetes instead without incurring some type of license penalty. (Industry veterans who make fun of such things typically prefix them with “VM-”.)
Swarm’s key virtue is that it can deploy containers in a test environment on the developer’s laptop, and in a full-scale production environment, with but minor modifications to the script.
A Ford Escort can also be made to run on the Indianapolis Motor Speedway (assuming you get a permit), but you can’t expect it to compete.
Tutum builds the Docker ecosystem into something reasonably more competitive. While Swarm will likely still exist in some form, Tutum has virtues that place it toe-to-toe against other deployment and orchestration options.
The reason most CIOs and executives give for why their organizations have not extended the containerization movement from the development phase into production, is because the tools haven’t matured to production-quality quite yet.
Security still ranks up there as a problem, but even the security issues could perhaps be tackled if organizations understood how the tools worked. As CMSWire has shown you, very sophisticated orchestration platforms do exist.
But, rather importantly, those platforms have not found a way to either subsume or coexist with virtualization platforms, led by VMware, that are not going away soon, or maybe at all. In fact, VMware is taking full advantage of this little detour in Docker’s otherwise smooth sailing course.
Swarm, and the other members of Docker’s home-grown platform such as Compose and Docker Machine — released just last February — are effective at getting containerized workloads onto a deployment platform, and running there. These are largely script-oriented tools, so when the platform changes — for instance, from whatever the development team uses to the organization as a whole — the scripts often have to change with them.
Tutum takes a very different tack, adopting a style that should be familiar to both software developers and DevOps professionals who provision services on Amazon AWS. Its aim is to treat each application an organization produces, including those that face its customers, as a single service with X number of containers.
That X is a scalable value, and Tutum handles the scalability using its own load balancing.
“A load balancer dispatches requests to the same application server, which is an application container in the Docker world,” according to a Tutum blog post last August. “For instance, if we link service A (containing 3 containers) and service B (containing 2 containers) to a load balancer, the load balancer will balance the traffic on 3 containers when accessing service A, and on 2 containers when accessing service B respectively.”
Tutum uses a browser-based portal that leads the developer or DevOps professional step-by-step through the process of making services active, including on multiple platforms simultaneously (such as hybrid clouds, or part-Amazon and part-Azure).
Anybody who has used Amazon AWS to get a VM up and running should feel comfortable using Tutum. This is what distinguishes Tutum’s approach from the more dazzling, sometimes sci-fi, approach of Mesosphere — it’s more down-to-earth, more fundamental, and less dazzling for the sake of being dazzling.
But most importantly of all, it could change CIO’s first impressions of Docker, both as a company and as an ecosystem.
The process of commanding Docker Swarm to set up pools of resources for immediate deployment of containers has drawn spontaneous applause... from a room full of software developers.
From a business manager’s perspective, the appearance of a developer deploying containers using Swarm on the Linux command line, and that of having witnessed a 10-year-old demonstrating a reconstruction of the first Pythagorean Theorem on the first Apple II, do not look all that different from one another.
I once overheard a business manager saying, “They really expect us to trust a multi-billion-dollar enterprise to one guy with a keyboard?”
The radical simplification of what has historically been treated as an enormously complex process is hard to believe. Swarm truly does what Swarm truly does.
Yet Tutum is more believable, and to that end, more sellable. A PowerPoint presentation to the Board of Directors showing Tutum at work will be received far more positively than a YouTube video of someone showing off Swarm at DockerCon.
When an organization considers making a risky investment in a new technology, that technology can’t enter the boardroom looking like it needs a shave and a shower.
So for CIOs and DevOps managers looking to move their organizations towards a much more manageable, less expensive, more effective approach to deploying applications in the cloud, just the very existence of Tutum in the Docker ecosystem — whether or not these organizations even end up choosing Tutum as their orchestrator — makes those difficult arguments much, much easier to articulate.