Licensing disputes have long been a hallmark of the commercial open-source landscape. Some major vendors, like Grafana and Element, have switched to more restrictive “copyleft” licenses, while others, such as HashiCorp, transitioned to a fully proprietary model last year with Terraform.

However, one $8 billion company is taking a different approach.

Elastic, the developer of the enterprise search and data retrieval engine Elasticsearch and the Kibana visualization dashboard, surprised many last month by announcing its return to open source nearly four years after adopting proprietary “source available” licenses. This decision contrasts with the trend of many companies abandoning open source altogether and even creating new licensing models like “fair source,” which several startups have embraced.

In 2021, Elastic moved to closed source licenses following years of conflict with Amazon’s AWS, which was offering its own managed version of Elasticsearch. While AWS was legally permitted to do so under the permissive Apache 2.0 license, Elastic was unhappy with how AWS marketed its version, particularly using the branding “Amazon Elasticsearch.” The company believed this caused significant confusion among customers, who may not fully understand the nuances of open-source projects and their associated commercial services.

“People sometimes think that we changed the license because we were upset with Amazon for taking our open-source project and providing it ‘as a service,’” Elastic co-founder and CTO Shay Banon told TechCrunch in a recent interview. “To be honest, I was always okay with it because it’s in the license that they’re allowed to do that. The issue we consistently faced was the trademark violation.”

Elastic pursued legal action to compel Amazon to cease using the Elasticsearch brand, a situation reminiscent of the ongoing WordPress controversy from the past week. While Elastic eventually settled its trademark dispute with AWS, such legal battles consume significant resources, diverting focus from the company's primary goal of protecting its brand.

“When we considered the legal route, we felt we had a strong case, which we ultimately won. However, that became less relevant after the change we made to the Elasticsearch license,” Banon explained. “But it was just taking too long—you can spend four years winning a legal case, and by that time, you’ve lost the market due to confusion.”

### Back to the Future

The decision to revert to open source was a sensitive topic internally, as the company had to use phrases like “free and open” instead of simply “open source.” However, the strategy proved effective, compelling AWS to fork Elasticsearch and create a variant called OpenSearch, which the cloud giant transitioned to the Linux Foundation just this month.

With enough time elapsed and OpenSearch now firmly established, Banon and his team decided to reverse their earlier decision and make Elasticsearch open source again.

“We knew that Amazon would fork Elasticsearch, but there wasn't a grand master plan here—I hoped that after some time passed with the fork, we might be able to return to open source,” Banon said. “And to be honest, it’s for a very selfish reason—I love open source.”

Elastic hasn’t completely returned to its previous approach, though. Instead of re-adopting the permissive Apache 2.0 license it once used, the company has chosen the AGPL, which imposes greater restrictions by requiring that any derivative software also be released under the same AGPL license.

For the past four years, Elastic offered customers a choice between its proprietary Elastic license or the SSPL (Server Side Public License), created by MongoDB, which the Open Source Initiative (OSI) did not recognize as “open source.” While SSPL provided some benefits of open source—such as code visibility and modification—by adopting AGPL, Elastic can now officially label itself as open source, as this license is recognized by the OSI.

“The Elastic [and SSPL] licenses were already very permissive and allowed you to use Elasticsearch for free; they just didn’t carry the 'open source' stamp,” Banon explained. “We understand this space intimately, but most users don’t—they simply Google ‘open source vector database,’ see a list, and choose based on that because they care about open source. That’s why I care about being on that list.”

Looking ahead, Elastic hopes to collaborate with the OSI to create a new license or at least engage in discussions about which licenses qualify as open source. According to Banon, the ideal license would sit “somewhere between AGPL and SSPL,” though he admits that AGPL might suffice for most purposes.

For now, Banon feels that simply being able to refer to itself as “open source” again is a significant achievement.

“It’s still magical to say ‘open source’—‘open source search,’ ‘open source infrastructure monitoring,’ ‘open source security,’” Banon stated. “Those two words encapsulate a lot—it signifies that the code is open, along with all the community aspects. It embodies a set of freedoms that we developers cherish.”