JaaS: the Team that Builds Jitsi Can Now Also Run it for You! Start now


The Atlassian Acquisition – 3 years later

Published on: June 4, 2018 by Emil IvovCategories: Atlassian | Jitsi Community

It has been 3 years Atlassian became Jitsi’s new home! To help give us an objective view on our accomplishments and where we are going next, we asked Chad Hart of webrtcHacks to do a guest post for us.

See his post below.

Your personal Jitsi team

It has been three years since Atlassian acquired BlueJimp, the original company behind Jitsi. I wrote a blog post speculating if this was a good or bad thing for the burgeoning WebRTC community that was just getting comfortable with Selective Forwarding Unit (SFU) concept. The track record of large companies acquiring small open source projects is often not good – the acquirer generally needs the expertise of the staff for its own projects and siphons away resources from the open source work. Without the same dedication and focus of the core team, the open source project starts to wither. In a new, fast moving technology like WebRTC this can be an effective death for the project.

I thought this would happen to Jitsi for a few reasons: Jitsi was Atlassian’s first open source acquisition. Atlassian is huge, Jitsi was tiny. The team was moved to Austin to be more closely embedded inside Atlassian and was not kept physically separate. Atlassian also had big needs for video conferencing in its portfolio – there was plenty of work to do inside of Atlassian instead of for the community.

I assumed Jitsi as an open source project was doomed.

I was very wrong.

Accelerated Development

In the past 3 years, the Jitsi team has accelerated its development. Atlassian has put money into the core team, growing the original Jitsi team, with many more involved on Atlassian related Jitsi projects (more on that below). The number of repos maintained under github.com/jitsi has grown from around 16 at the time of acquisition to 70 today with 40 of these repos receiving some update within the past year.

With so many project repos and a continuous deployment process, it can be hard to track, so here is a list of some of the introductions and major improvements that caught my attention:

  • 2015 (since April)
    • Firefox support – work with the Mozilla team on multistream and renegotiation
  • 2016
    • Simulcast – switch between different stream qualities
    • YouTube LiveStreaming – livestream Meet sessions
    • React Native – SDK for building apps with React
    • Jibri – new Jitsi recording and broadcasting tool
    • Meet iFrame API – embed Meet inside any webpage
    • Introduction of lib-jitsi-meet – A low-level JS video API for making your own GUI
  • 2017
    • Meet Electron app (jitsi-meet-electron) – make a Meet desktop app
    • Meet apps for Android and iOS
    • Microsoft Edge support
    • Speaker times in Meet
    • Adaptivity – better video quality in variable bandwidth conditions
    • Peer-to-peer for 1-1 calls – improve performance for 1:1 calls
    • Jitsi Meet Mobile SDK for developing your own apps
    • PSTN calling added to meet.jit.si
    • Speech to text transcription
  • 2018
    • Meet UI redesign
    • Jibri rewrite in Kotlin
    • Send-side bandwidth estimation

The Community Keeps Growing

Meanwhile, the Jitsi community has grown substantially. I did a check to see how their github contributions have grown using this methodology. The number of distinct github users publicly referencing a Jitsi project in a given month jumped from less than around 150 in the months before the acquisition to 424 in March. The number of distinct developers publicly pushing and reviewing Jitsi code the month of the acquisition was 64. That count was 153 in February.

Graph showing growing Jitsi developers on Github

Number of unique Github users referencing a Jitsi term (jitsi, jigasi, jicofo, jidesha, jicoco, jigmi, jirecon, or jiloin) each month. Code-users made a push, pull request, or did a pull request review that month. See here for the data.

It’s not just github developers either. The number of sessions on meet.jit.si – Jitsi’s free Meet instance that anyone can use, has jumped from 24K in April 2015 to 185K in March. NSA whistleblower Edward Snowden even recommended Jitsi in Wired and used Meet during a security conference, exposing it to a more mainstream audience.

The Community Loves Jitsi

I have been helping Atlassian review the Jitsi user community. As part of this project, we put together a Jitsi user survey to get feedback on the kinds of users, their impressions of Jitsi, and find out get some ideas on how to grow the community.

Perhaps it shouldn’t be a surprise that those who track Jitsi’s messaging are a fan of Jitsi, but clearly Jitsi is liked:

77% of Jitsi user survey respondents rated 8 or above on their likelihood of recommending Jitsi

Responses from a recent Jitsi user survey asking the likelihood of recommending Jitsi

What’s in this for Atlassian?

So Jitsi has done great, but is there anything in this for Atlassian? Is Atlassian just being benevolent here and taking a loss supporting Jitsi’s open source projects?

I had some discussions with the Atlassian team to uncover this. It turns out most of the Jitsi roadmap is at least indirectly tied to Atlassian product needs – most notably Stride, through Stride Meetings. Stride is Atlassian’s team communications solution. It uses Jitsi for its scheduled and ad-hoc video calls. You’ll see the user interface is very similar to Meet. Like Meet, it’s a streamlined experience with the usual calling and screen share experience but also has additional hooks into Stride. These hooks help Stride users move from messaging to video more fluidly and take advantage of Stride’s other collaboration features.

picture of Emil on mobile during a Stride call

Stride call with Emil and team.

I recently asked Evan Michner, Principal Product Manager for Stride in charge of the communications features if Jitsi’s open source approach really is a benefit. He sees several benefits:

  • The best video engine – without a large community behind it, Jitsi could not maintain a lead technology position
  • Feature validation – new features and capabilities can be test and validated rapidly before being rolled into Stride
  • Cross-sell – they have not taken advantage of this yet, but for Meet users that don’t want to setup their own infrastructure and need additional chat and collaboration features with an SLA they can pay for, Stride is a natural upsell

With all the great momentum in the community and Atlassian getting the products it needs, it seems Jitsi team generally gets what it asked for.

What’s Next for Jitsi?

Talking to the Jitsi and Atlassian team, it does not look like there will be any major changes to their market approach. The focus is clearly about improving the Meet user experience. As Stride explores new capabilities, the Meet experience will need to follow or drive it in the case of video conferencing features.

Overall, this is a great case study on how to acquire and open source community and how to nurture it instead of killing it. We should see many more great things out of Jitsi as this trend continues.

Guest Author
Chad Hart of webrtcHacks

Disclaimer: The statements above are my own opinions and I do not represent Atlassian. This post was produced as part of a paid project for Atlassian to help assess the Jitsi community, advise on improving outreach, and suggest methods for growth

Share this!