Success in media streaming open-source : a view after the OSMART#3 workshop
The OSMART initiative (Open-Source Media Application Reference Tools) is a collaborative effort aimed at fostering discussion and collaboration in the streaming media industry. It provides a platform for open-source maintainers, developers and stakeholders to share ideas, tackle challenges and explore synergies. Organized by Eoghan O’Sullivan with key contributors like Jordi Gimenez (Head of Technology at 5G-MAG), Daniel Silhavy (Project Manager at Fraunhofer Fokus), and Thomas Stockhammer (Senior Director at Qualcomm), the OSMART workshops are critical for advancing open-source projects in this field.
The 3rd edition of OSMART took place on December 5th 2024. Motion Spell had previously participated as a speaker in both the previous editions (see our blog post about the first edition). In this 3rd edition I delivered two presentations: one on JCCP conformance for CTA-WAVE (slides) and another on GPAC MABR support for DVB (slides). Nils Duval also represented Motion Spell, contributing to the 5G-MAG XR activity (slides). Motion Spell has established itself as a key partner in these events, thanks to our strong commitment to open-source initiatives and industry standards.
OSMART is a dynamic event, and I particularly appreciate how contributions are documented in an open format. Questions from the audience are often addressed in writing by the speakers after their presentations, adding clarity and depth. The organizers are actively exploring ways to build momentum through a collaborative platform based on github tools.
Another organization working to create momentum is SVTA, with CEO Jason Thibeault promoting the concept of the “SVTA Sandbox”. The idea is to create a cloud-based platform that can spin-up streaming workflows for testing open-source (nginx/Varnish/FFmpeg/GPAC/dash.js being mentioned) and proprietary workflows. This initiative appears to overlap with Eyevinn’s Open-Source Cloud project. It will be interesting to see if these efforts can succeed in advancing the ecosystem in terms of interoperability or acceleration in new technology deployments.
As an open-source software maintainer, I find maintaining third-party cloud platforms hard work because of both the initial bootstrapping effort and the long run cost. This idea resonated with Jordi Gimenez from 5G-MAG: Jordi also mentioned that 5G-MAG developments have only been supported by its members, while the 5G-MAG organization only covers coordination costs. This highlights a recurring issue: corporate communities often fail to contribute sufficiently to existing projects, even when funding is provided.
This edition of OSMART emphasized synergies within the industry. It became evident that some projects, such as dash.js and livesim, are foundational and heavily relied upon. They could be considered part of the industry's infrastructure. Similarly, SVTA’s Common Media Library (CML) aspires to occupy this category.
I always feel there’s an elephant in the room when we fail to acknowledge the omnipresence and influence of established open-source projects and communities like VideoLan, FFmpeg, MediaInfo, GPAC, and TSDuck. I can imagine a OSMART#4 dedicated to the contribution of these projects to the streaming media industry.
In my view, the success of an open-source project can be defined by the following criteria:
-
Standardization Success:
- Meeting conformance or interoperability objectives (e.g., JCCP, DVB MABR, ComplianceWarden, CTA-WAVE DPCTF, DVB-I Reference Application).
- Fostering standard adoption (e.g., MPEG-DASH through dash.js, GPAC, JCCP).
-
User Popularity
- Measured by engagement metrics, such as discussions and open issues. Some projects target a niche audience of media streaming developers, while others, like dash.js, have broader appeal, serving JavaScript developers and appearing on numerous websites.
-
Developer Popularity
- Based on GitHub stars: Most of the projects showcased had between 4 and 50 stars, reflecting limited interest from the developer community—a notable challenge. For context, GPAC, despite being a niche project, has garnered nearly 3,000 stars.
- Number of contributors: Some projects face difficulties after being open-sourced without a clear plan for long-term maintenance or reaching a critical mass of contributors. As a result, many industry projects unfortunately falter and struggle to sustain themselves once initial funding runs out.
-
Long-Term Viability
- Availability of comprehensive documentation.
- Sustainability beyond the initial funding effort.
- Robust testing frameworks.
- Social recognition through awards, online presence, and references.
In my opinion, dash.js stands out as the flagship of corporate open-source projects. Daniel Silhavy mentioned that dash.js risks stagnating without industry funding - a concern I raised above, emphasizing the importance of working with and leveraging established projects like GPAC. The reliance on diverse communities builds resilience, and we should anticipate the end-of-life phase for projects to ensure they are left in a reusable state.
The community aspect was another under-represented area. Long-time observers recognize that the same individuals often shoulder the responsibility of maintaining these projects. Talent retention and broader impact are crucial. Funding those who actively contribute beyond their funded commitments should be encouraged and incentivized.
Toward the end of the event, Thomas Stockhammer raised a question about the need for synergies and the creation of a physical event in addition to the online workshops. A key takeaway of the discussion that followed, joined by Remo Vogel and Daniel Silhavy, with responses from Jordi Gimenez and myself, was the need for better coordination among organizations and more open discussions about the business value of these developments.
The key issue is that when a company or consortium funds the development of a project, releases it publicly (e.g., as open source), and then assumes it will maintain itself without further effort, this is a recipe for failure. Open source projects need to achieve a critical level of popularity and community engagement. Unfortunately, this is often overlooked.
While I suggested success factors earlier, the most critical ones remain 1) the fulfillment of clear short-term objectives for the funding organizations (such as the validation of a standard) and 2) long-term objectives with project analytics tracking to ensure enough popularity and self-maintenance - the other option being to fund an existing open-source project which already reached that point and would maintain external contributions as part of their usual maintenance.
