TADSummit 2023: Open Source

Open source telecom software is the foundation upon which programmable telecoms / communications is built. It is also a critical technology for telcos to make the leap to Techco. Put simply, if telcos continue to use their legacy closed source vendors, the Techco vision will not happen. This advice comes from the fixed line telcos, that have already had their 5G moment with 1Gbps internet access.

In this post we’ll review the following sessions:

Maintaining Communications Apps, Arin Sime, Founder/CEO WebRTC.ventures.

Slides can be downloaded here.

WebRTC.Ventures are the people behind many popular communication applications and features. Their client list is impressive, including Amazon Chime, /daily, Twilio, Vonage, Agora, Janus, Liveswitch, Janus and Jitsi. WebRTC.Ventures also sponsored our Day 1 evening soiree.

At TADSummit we tend to focus on the bleeding edge aspects of programmable communications. However, we should not understate the complete Product Lifecycle:

• Prior to Deployment
• Deployment
• Support Teams in General
• Support of Comms Apps Specifically
• Ongoing Feature Development

In this presentation Arin reviews their experiences with an EdTech application. Some of the features include:

• Video chat between instructors and students
• Ability for instructors to move between rooms
• Whiteboards (3rd party tool)
• Text Chat
• Screensharing
• Login, Security, Privacy

Arin then provides a step by step guide to Design and Maintenance process. along with tools used and methods used.

  • Design for Security – Security should be a team concern, and developers need to think about it from the start
  • Design for Observability – logging, monitoring, alerts
  • Design for Testability – We will discuss load testing specifically
  • Design for Resilience – Build in scalability and error handling to respond gracefully to failure
  • Design for Change – Make sure your processes and communication encourage and support change management

It’s a great recipe for the industry to follow.

Chrome extension to monitor WebRTC outbound flows, troubleshoot and learn. Philippe Sultan.

Slides can be downloaded here.

Philippe was one of the original cast members of TADSummit back in 2013, along with Jean Deruelle.

Philippe brings a massive breadth of experience in implementing RTC and WebRTC. In this presentation he reviews how WebRTC is implemented at Livestorm.

Philippe highlights the critical remaining issue for video comms, is not WYSIWYG (What You See is What You Get), even after one decade. Often calls begin with ‘Can you see me?’, or ‘Can you hear me?’, or ‘Can you see me?’ Hence quality measurements are critical, and here Livestorm uses getstats Chrome extension: https://github.com/livestorm/webrtc-publisher-stats

Philippe then runs through an excellent list of resources and highlights some tricks, trips and discoveries, especially in multicast. WebRTC is still maturing, hence why in-house, like Philippe, or for-hire resources like WebRTC.Ventures remain essential.

WebRTC broadcasting: standardization, challenges and opportunities. Lorenzo Miniero.

Slides can be downloaded here.

I was really happy when Lorenzo was able to join us at TADSummit. He attended the first TADHack in 2014, when WebRTC was sparkly new.

Lorenzo reviews WebRTC broadcast. Sharing his frustration of watching a ‘live’ stream over the internet, while the pub next door was a few second ahead, impacting his experience!

The broadcast TV industry is a mature market, where quality is everything. I worked with the BBC in the early days of MPEG, their requirements created a digital TV silo.

But 20+ years later the worlds of broadcast TV and WebRTC are starting to converge. Lorenzo highlight the 2 important issues of ingestion, distribution and of course scaling for WebRTC broadcasting.

Here the new WebRTC group defining WISH (WebRTC Ingestion Signalling over HTTPS) and WHIP (WebRTC-HTTP ingestion protocol) are creating the groundwork for this next category. The standards are evolving as SFUs use their own APIs, while a standard method would make more sense, and hence the emergence of WHEP (WebRTC-HTTP egress protocol).

This incremental, implementation / demonstration approach across experts on this topic, Lorenzo’s PhD thesis was on this every topic, shows the power of the IETF in creating standards that are market centric and get adopted.

While the broadcast industry will initially see slow adoption as it is mature and quality is everything, direct to consumer has grown in importance. So we make see a rapid jump onto WebRTC broadcast once the industry see’s its success. Watch this space!

And finally a shout out to Januscon, April 29-30, 2024, Napoli — https://januscon.it!

Intro to WebTrit and Call to Action to join the WebTrit Open Source Community. Andriy Zhylenko.

Slides can be downloaded here.

Hugh Goldstein, a longtime TADS member recently joined WebTrit as CEO. It was great Andriy was able to join us to explain WebTrit, a white-label, multi-platform, open-source softphone app for telcos, hosted PBX providers, and SaaS developers. Rather than the services approach of WebRTC.Ventures, WebTrit packages up a white-label solution.

Simply with WebTrit:

  • Deployed in a cloud / private cloud
  • Client <-> WebTrit: WebRTC / WebTrit <->VoIP system: SIP
  • Auto-provisioning, push notifications, contact sync, etc.
  • Publish your own app to Apple/Google (visual editor for basic branding)
  • Source code for mobile & web app is available on GitHub (MIT license)

Depending on your needs WebTrit could be a fit, as always I recommend having no-BS discussion of your requirements with a range of experts, to understand the pros and cons of each approach.

WebRTC extravaganza. Romain Vailleux. Apizee DevRel & Partnership Manager.

Slides can be downloaded here.

We wrapped up the Open Source section for the main onference with a fun review of the many innovative applications of WebRTC. When technology is made easy enough, so developers outside then core telecoms / real time communications industry can play with them. Really cool, new, and innovative applications are discovered. As Romain reviewed, and I’ve witnessed at TADHack many times over the years.

RESULTS: Open Source Telecom Software Survey 2023. Alan Quayle.

The slides can be downloaded here.

We hd many hardy souls stay until the very end of TADSummit. Every year since 2019 I’ve run the Open Source Telecom Software Survey. This is just a quick summary. I will do a more indepth review in November, as the results reveal many insights on the current projects, and important developments for the industry.

For 2023 we had 168 responses (2023) versus 120 (2022), lots of insightful opinions. I consider it a excellent barometer for the industry.

Highlighting a few of the insights:

  • 58% of respondents have little to no awareness of the EU Cyber Resilience Act, while 42% have some to full awareness. However, assumptions on its applicability and that passing ISO 27001 covers compliance need testing. General opinion is delay or revision are considered most likely.
  • “Which projects do you use, and how do you contribute” generated the broadest response and we’ll make 2024’s survey a review across the projects. Please contribute what questions you’d like asked.
  • On Representation, the first 4 in the list came from women; and are actions/issues all projects can tackle.
    • Misguided attitudes must be stopped by leadership
    • Projects must mentor, e.g. partner with GirlCode
    • Projects are exclusive (self-selecting)
    • Be Genuine, not lip-service / PR
    • I agree the companies with the biggest revenues should lead the way in supporting representation for the FOSS industry. But even small projects can tackle the above 4 actions/issues.
  • On will the desktop phone every disappear. I think disappear is too strong / simplistic. There will likely be desk top phones as long as humans are around. Rather, when will desktop phones be <20% of employees?
    • 2040 was considered the most likely date when desktop phones will disappear. But like SS7, it may hang around for a few more decades.
  • Voice CPaaS Evolution shared some interesting perspectives
    • FOSS voice CPaaS is considered a viable alterative
    • A fundamental problem in replacing Twilio is the cost to change CPaaS outweighs any possible savings.
    • FOSS works for new build, with geeky organizations that want to control their roadmap, or incremental voice gateway projects. So could be limited to 20% of the market. Which I think most FOSS CPaaS projects will be more than happy with.
  • SMS CPaaS is a great review of the current status and issues.
    • People seem to be a little down on SMS CPaaS. I’ve never seen this, a multi-billion dollar market that’s been growing at 30+% for years seems to have hit a wall.
    • Given what’s happening with SMS, IP Messaging is considered greater competition than in the past
    • Some markets (Europe and South America) are seeing strong demand.
  • Fragmentation across IP Messaging, in-app, and PSTN looks likely over next 3 years.
    • Platform control remains the primary concern for IP Messaging.
  • Fraud and Identity CPaaS
    • The Fraud and Identity market is far from defined and will remain turbulent / fragmented.
      • Some see Fraud / Identity becoming a standardized feature in CPaaS that will feed into risk / identity aggregators.
      • Some mentioned Passkeys as the likely winner. However, SIM swap can not be determined through passkey.
      • Many alluded to the telecom industry’s inability to work together to deliver solutions for specific industries.
  • Slide 39 lists many great open source projects that can complete a full open source UCaaS/CCaaS/CPaaS. The only gap appeared to be on Regulatory Compliance.
  • WebRTC
    • Browser compatibility and the permission prompts still suck after all these years!
    • Respondents are even split on whether they will upgrade their telephony infrastructure to WebRTC in the next 2 years.
    • On open-source media servers, Freeswitch, Asterisk and Jitsi are roughly neck and neck, with Janus further back.
    • Common use cases of AI/ML in WebRTC are: audio transcription, background blur / replacement, background noise reduction, keyword detection, and summaries / note taking.
    • With voicebots, most are experimenting

Thank you to everyone who took part in TADSummit 2023, see you again next year.