Docker buys Conductant to catalyse coding development

CodingContainer technology pioneer Docker has bought start up Conductant, best known for creating the Aurora strand of the Apache Mesos clustering system. Conductant’s software is used to catalyse faster development of large scale code.

Announcing the acquisition on its website Docker spokesman Solomon Hykes placed more emphasis on the talent, rather than the technology, that is being brought in with the take over of an early stage start up. Welcoming the Conductant ‘team’ to the Docker ‘family’ Hykes outlined the contributions that founders Bill Farner, David Chung and John Sirois made to operating and scaling production systems at Google, Twitter and Zynga.

Farna, who created the Aurora Project, will lead the process of integrating the clustering technology into the fabric of its container software. Docker’s expansion policy is to buy emerging software tool makers and integrate them into its container software core, according to Hykes. In January BCN reported how Docker has acquired Unikernel Systems in order to channel its hypervisor and unikernel experience into the development of Docker’s container systems. “We believe our job is integrating these technologies in tools that are easy to use and help people create new things. We did this for Linux containers, to help make applications more portable,” wrote Hykes.

Aurora, an extension of the Apache Mesos clustering system, is specifically designed for hyper scale production environments. Hykes claimed it is recognized as the most scalable and operationally-robust component of the Mesos stack, which in turn helps to create the conditions for operations-driven development (ODD). The experiences of the Conductant team, operating global scale clouds for Google, Twitter and Zyng, forced them to develop new techniques for rapid development. Bill Farner’s team at Twitter built Aurora to automate massive server farms that could be managed by handful of engineers.

Docker now plans to incorporate the best ideas from Aurora into Docker Swarm, which allows for any app to go on any infrastructure on any scale, and integrate Aurora as an optional component of the official Docker stack. One option is to integrate Aurora with Docker Swarm to form a powerful large-scale web operations stack.

While Swarm is designed to be the standard base layer to scale all kinds of applications, Aurora is optimized for large-scale consumer apps reaching hundreds of millions of users. “By making two of the most popular open-source infrastructure projects interoperate better, we believe both communities will benefit,” said Hykes.