AUSTIN, Texas—Last May, the open-source OpenStack Foundation announced that it was moving from a restricted integrated project release model to the Big Tent model. Under the Big Tent, more projects are included under the definition of OpenStack, providing a wide array of capabilities to users. Now, a year after the Big Tent was announced, leaders of the OpenStack project discussed what’s wrong and what’s right with the Big Tent at the OpenStack Summit here.
In the integrated release model, OpenStack followed an approach where projects first had to be accepted into incubation and once they were considered to be mature, they had to be voted on to become part of an integrated release. The integrated release approach, however, was not working well and couldn’t scale very fast.
“We were completely stuck at the end of the Juno release,” Thierry Carrez, release manager for OpenStack, said. “Now we are no longer stuck.”
The OpenStack Juno release debuted in October 2015, adding the Sahara data processing project as part of the integrated release.
Monty Taylor, distinguished engineer at IBM and a member of the OpenStack board of directors, said that with the Big Tent approach, a person or a group of people are now free to work on a project without first petitioning the OpenStack project for incubation or integrated status.
Carrez added that in the integrated release approach, by the time of the OpenStack Juno release, the requirements for new projects were so high that it limited the ability of any new project to become part of the OpenStack release.
“Now we also have more collaboration in spaces where there was none before,” he said.
With the Big Tent model, OpenStack is also able to be more reactive to changing market dynamics. For example, Carrez noted that the OpenStack Magnum container project likely wouldn’t have easily been integrated in the old model for OpenStack projects.
While the Big Tent enables more collaboration, the inclusiveness of the approach also can create more competition. In the integrated release model, a single project would own a specific problem space, according to Carrez. With the Big Tent, there can now be multiple projects tackling similar capabilities.
While projects in the OpenStack Big Tent are not subject to the same restrictions that applied in the integrated release era, they are still required to follow what is known as the “OpenStack Way.”
“The Big Tent forced us to document the OpenStack Way,” Carrez said. “So it’s not just passed down campfire folklore.”
The OpenStack Way defines the processes and open-source community approaches that all OpenStack projects are expected to follow.
While there have been positive benefits to the Big Tent model, there are also negative aspects. Currently, Carrez said, there is no requirement for organizational diversity for projects in the Big Tent. As such, a specific project might only have developers working for a single company. The problem with that, according to Taylor, is that if that company has a change of heart, the project could potentially cease development.
Another challenge of using the Big Tent model is that it can get confusing. Because there now are so many projects, Taylor admitted that he no longer knows all of them. There is also a risk of projects that just don’t grow and progress.
“There are some corner cases of dead projects that we accepted into the Big Tent, hoping they would grow, but they are not going anywhere,” Carrez said. “Keeping them in the tent makes the tent smell like death.”
Sean Michael Kerner is a senior editor at eWEEK and InternetNews.com. Follow him on Twitter @TechJournalist.