How serious is Docker about supporting massively scalable workloads? More serious than ever, it seems, following the company’s acquisition of Conductant, bringing large-scale cloud management expertise to the Docker team.
Docker announced the acquisition on March 3. Conductant developed Aurora, an open source cluster management framework based on Mesos. Aurora is best known as the orchestration platform that helps run Twitter, but it has been an Apache open source project since 2013, allowing developers of all stripes to take advantage of the code.
The Conductant team also has experience in running massive cloud workloads for other big-name vendors, including Google and Zynga.
That expertise in managing huge cloud infrastructures via automated tools is what attracted Docker to Conductant. CEO Solomon Hykes played up the benefits of the deal for bringing Docker developers the skillsets they need to make containers work better at massive school, as well as to extend Docker’s existing orchestration toolset.
“Very few tools available today have been battle-tested at such operational scale,” he wrote. “There are many commercial distributions of Mesos, but none of them incorporate Aurora. We believe that is a wasted opportunity. We plan on incorporating the best ideas from Aurora into Docker Swarm, and are exploring integrating Aurora as an optional component of the official Docker stack.”
The acquisition has an added a bonus for Docker, too. One member of the Conductant team, Bill Farmer, not only founded Aurora but also worked on Google Borg, another cloud orchestration tool that helped manage massive workloads.
From the perspective of the container ecosystem as a whole, Docker’s decision to acquire Conductant is significant for two main reasons. First, the move reflects Docker’s interest in expanding its focus beyond containers themselves. Like Docker’s Unikernels acquisition in January, the addition of Conductant gives Docker resources that will help it to become more than a company that develops container software based on LXC. Now, Docker is assuming a broader purview by offering solutions for deploying and orchestrating large server clusters more generally.
Second, the Conductant acquisition underlines Docker’s interest in catering to enterprises with very large workloads. Docker containers may be useful for smaller datacenters, too, since they can simplify application deployment and increase portability. But Docker clearly believes that the commercial potential of containers lies with companies that have massive infrastructures to manage. Conductant and the Aurora expertise that come with it position Docker to pitch container-based solutions to the likes of Google and Twitter.
That’s a very different sort of strategy from focusing on containers for start-ups and SMBs. Docker may not be ruling out the applicability of containers for smaller environments, but those are not where it sees its business and revenue streams developing.