WebRTC disconnections are quite common, but you can “fix” many of them just by careful planning and proper development.
Years ago, I developed the H.323 Protocol Stack at RADVISION (later turned Avaya, turned Spirent turned Softil). I was there as a developer, R&D manager and then the product manager. My code is probably still in that codebase, lovingly causing products around the globe to crash from time to time – as any other developer, I have my share of bugs left behind.
Anyways, why am I mentioning this?
I had a client asking me recently about disconnections in WebRTC. And it kinda reminded me of a similar issue (or set of issues) we had with the H.323 stack and protocol years back.
If you bear with me a bit – I promise it will be worth your while.
I am starting this week the office hours for my WebRTC course. The next office hour (after the initial “hi everyone”) will cover WebRTC disconnections.
Check out the course – and maybe go over the first module for free:
H.323 is like SIP just better and more complex. At least for me, who started his way in VoIP with H.323 (I will always have a soft spot for it). For many years, the way H.323 worked is by opening two separate TCP connections for transporting its signaling. The first for passing what is called Q.931 protocol and the next for passing H.245 protocol.
If you would like to compare it to the way WebRTC handles things, then Q.931 is how you setup the connection – have the users find each other. H.245 is similar to what SDP and JSEP are for (I am blatantly ignoring H.225 here, another protocol in H.323 which takes care of registration and authoentication).
Once Q.931 and H.245 get connected, you start adding the RTP/RTCP stuff over UDP, which gets you quite a lot of connections.
Add to that complexities like tunneling H.245 over Q.931, using something called faststart instead of H.245 (or before H.245), then sprinkle a dash of “parallel H.245” and then a bit of NAT traversal and/or security and you get a lot of places that require testing and a huge number of edge cases.Where can H.323 get “stuck” or disconnected?
With so many connections, there are a lot of places that things can go wrong. There are multiple state machines (one for Q.931 state, one for H.245 state) and there are different connections that can get severed for one reason or another.
Oh – and in H.323 (at least in its earlier specifications that I had the joy to work with), when the Q.931 or H.245 connections get severed – the whole session is considered as disconnected, so you go and kill the RTP/RTCP sessions.
At the time, we suffered a lot from zombie sessions due to different edge cases. We ended up with solutions that were either based on the H.323 specification itself or best practices we created along the way.
Here are a few of these:
H.323 existed before smartphones. Systems were usually tethered to an ethernet cable or at most over WiFi in a static location at a time. There was no notion of roaming or moving between networks. Which meant that there was no need to ask yourself if a connection got severed because of a switch in the network or because there’s a real issue.
Life was simple:
And if you were really insistent then maybe this:
(in real life scenarios, these two simplistic state machines were a lot bigger and complicated, but their essence was based on these concepts)Back to WebRTC signaling and transport
WebRTC is simpler and more complicated than H.323 at the same thing.
It is simpler, as there is only SRTP. There’s no signaling that is standardized or preselected for WebRTC. And for the most part, the one you use will probably require only a single connection (as opposed to the two in H.323). It also has a lot less alternatives built into the specification itself that H.323 has.
It is more complicated, as you own the signaling part. You make that selection, so you better make a good one. And while at it, implement it reasonably well and handle all of its edge cases. This is never a simple task even for simple signaling protocols. And it’s now on you.
Then there’s the fact that networks today are more complex. User expect to move around while communicating, and you should expect such scenarios where users switch networks in mid-session.
If you use WebRTC in a browser, then you get these interesting aspects associated with your implementation:
A lot of dying taking place on the browser, and the server, or the other client, will need to “sniff” these scenarios as they might not be gracefully disconnected, and decide what to do about them.Where can WebRTC get “stuck” or disconnected?
We can split disconnections of WebRTC into 3 broad categories:
In each, there will be multiple scenarios, defining the reasons for failure as well as how to handle and overcome such issues.
In broad strokes, here’s what I’d do in each of these 3 categories:#1 – Failure to connect at all
There’s a decent amount of failures happening when trying to connect WebRTC sessions. They start from not being able to even send out an SDP, through interoperability issues across browsers and devices to ICE negotiation failing to connect media.
In many of these cases, better configuration of the service as well as focus on edge cases would improve the situation.
If you experience connection failures for 10% or more of the sessions – you’re doing something wrong. Some can get it as low as 1% or less, but oftentimes that depends on the type of users your service attracts.
This leads to another very important aspect of using WebRTC:
Measure what you can if you want to be able to improve it in the future#2 – Media disconnections
Sometimes, your sessions will simply disconnect.
There are many reasons why that can happen:
Each of these requires different handling – some in the code while others some manual handling (think customer support working out the configuration with a customer to resolve the firewall issue).#3 – Signaling disconnections
Unlike H.323, if signaling gets disconnected, WebRTC doesn’t even know about it, so it won’t immediately cause the session itself to disconnect.
First thing you’ll need to do is make a decision how you want to proceed in such cases – do you treat this as session failure/disconnection or do you let the show go on.
If you treat these as failures, then I suggest killing peer connections based on the status of your websocket connection to the server. If you are on the server side, then once a connection is lost, you should probably go ahead and kill the media paths – either from your media server towards the “dead” session leg or from the other participant on a P2P connection/session.
If you want to make sure the show goes on, you will need to try and reconnect the peer connection towards the same user/session somehow. In which case, additional signaling logic in your connection state machine along with additional timers to manage it will be necessary.Announcing the WebRTC course snippets module
Here’s the thing.
My online WebRTC training has everything in it already. Well… not everything, but it is rather complete. What I’ve noticed is that I get repeat questions from different students and clients on very specific topics. They are mostly covered within lessons of the course, but they sometimes feel as being “buried” within the hours and hours of content.
This is why I decided to start creating course snippets. These are “lessons” that are 3-5 minutes long (as opposed to 20-40 minutes long), with a purpose to give an answer to one specific question at a time. Most of the snippets will be actionable and may contain additional materials to assist you in your development. This library of snippets will make up a new course module.
Here are the first 3 snippets that will be added:
While we’re at it, office hours for the course start today. If you want to learn WebRTC, now is the best time to enroll.
CPaaS differentiation seems to be revolving around tackling niches.
Time for another look at the world of CPaaS – Communication Platform as a Service. In January 2018, a bit over a year ago, I’ve looked at CPaaS trends for 2018. The ones there were:
I’d like to look at what’s happening in CPaaS this time from a slightly different angle, which alludes itself to trends as well, but in a more nuanced way. From briefings I’ve been given this past few weeks and the announcements and stories coming out of Enterprise Connect 2019, it looks like different CPaaS vendors are settling on different target audiences and catering to different use cases and market niches.
Today CPaaS is almost synonymous to Twilio. Every player looks at what Twilio does in order to plot its own route in the market, at times, with the intended aim of disrupting Twilio and then mostly with lower price points. In other times, with trying to offer something more/better.
Then there are external players who add APIs to their platform. Usually UCaaS (Unified Communications as a Service) platform. They don’t directly compete with CPaaS, but if you are purchasing a “phone system” for your enterprise from a UCaaS player, then why not use its APIs and services instead of opting for another vendor (a CPaaS vendor in this case)?
Planning on selecting a CPaaS vendor? Check out this shortlist of CPaaS vendor selection metrics:
Get the shortlist
Here are how some of the vendors in this space are trying to differentiate, pivot and/or find their niche within the CPaaS market.Agora.io – Gaming
If you look at Agora’s blog, what you’ll find out there is a slew of posts around gaming and gaming related frameworks (Unity to be exact):
Gaming is an untapped market for CPaaS.
There’s communications there of all kinds – collaboration or communications across gamers inside a game, talking before the game, streaming the game to viewers, etc.
All this communications is either developed by the gaming companies (not a lot), being catered for by specialized VoIP gaming vendors, done out of scope (using Discord, Skype, …). Rarely is it covered by a CPaaS vendor.
Somehow, for CPaaS cracking this market is really tough. Agora.io is trying to do just that, along with its other focus areas – live broadcast and social (two other tough nuts).ECLWebRTC – Media Pipeline
The Japanese platform from NTT Communications – ECLWebRTC.
Like many of the WebRTC-first/only platforms out there, ECLWebRTC had an SFU implementation and support for various devices and browsers.
When you get to that point, one approach is to go after voice and PSTN. Another one is to add more features and increase the sizes of meetings and live broadcasts that can be supported.
ECLWebRTC decided to go after machine learning here, with the intent of letting its customers integrate and connect its media paths directly to cloud APIs. This is done using what they call Media Pipeline Factory, which feels from the looks of it like a general purpose media server.
ECLWebRTC is less known in Europe and the US, and probably not much outside of Japan either. With the Japanese market focus on automation, it makes sense that media pipeline would be a focus area for ECLWebRTC. This type of a capability is relevant elsewhere as well, but it doesn’t seem to be a priority for others yet.Infobip – Omnichannel
I’ve had the opportunity to fiddle around with Infobip Flow recently, something that turned out to be a very pleasant experience. From Flow, it became apparent that Infobip is working hard on offering its customers an omnichannel experience. Compared to other CPaaS vendors, they seem to have the most coverage of channels:
To the above, you can add SMS and RCS and email.
Infobip Flow has another nice quality – it is built for both inbound and outbound communications. Most of its competitors do inbound flows only.
In a world where competition may force price wars on CPaaS basic offerings of voice and SMS, adding support for omnichannel seems like a good way to limit attrition and churn and increase vendor lock-in.RingCentral – Embeddables
RingCentral isn’t a CPaaS vendor. They offer a communication service for the enterprise. You got a company and need a way to communicate? There’s RingCentral.
What they’ve done in the past couple of years was add an API layer to some of their services. Things like pushing messages into Glip, handling phone calls, etc.
The idea is that if you need something done in an automated fashion in RingCentral you can use the API for it. In many simple cases, this might be used instead of adopting CPaaS APIs. in other cases, it is about using a single vendor or having specific integrations relevant to the RingCentral platform.
What RingCentral did was add what they call Embeddable:
“With RingCentral Embeddable, you can embed a full-featured softphone into your favorite web application for an integrated communications experience that drives productivity and ease of use without lengthy development time“
This concept of embedding a piece of code isn’t new – YouTube videos offer such a capability as well as a slew of other services out there. When it comes to communications, it is similar in nature to what TokBox has in the form of Video Chat Embeds but done at the level of users and their user accounts on RingCentral.
This definitely makes integrations of RingCentral with CRM tools a lot easier to get done, and makes it easier to non-developers to engage with them – similar to how Flow type offerings make it easier for non-developers to handle communication flows.SignalWire – Price and Flexibility
SignalWire is an interesting proposition. It comes from the team that created and is maintaining FreeSWITCH, the leading open source framework used today by many communication providers, including some of the CPaaS vendors.
The FreeSWITCH team decided to build their own managed service (=CPaaS in this case), calling it SignalWire. Here’s a few examples of the punchy copy they have on their website:
What they seem to be aiming for are two things: price and flexibility
They offer close to whole-sale price points (at least based on the website – I haven’t gone to a price comparison on this one, though their sample pricing for the US does seem low).
To make things easier, they are targeting Twilio customers, doing that by offering TwiML support (similar to what Pilvo did/is doing). TwiML is a markup language for Twilio, which can be used to control what happens on connected calls. Continuing with the blunt approach, SignalWire calls this LāML – Legacy Antiquated Markup Language.
While this may fit a certain type of Twilio customers, it certainly doesn’t cover the whole gamut of Twilio services today.
On the flexibility front, there’s mostly marketing messages today and not any real announced products on the SignalWire website.
Besides LāML there’s a WebSocket based client API/SDK, not so different than what you’ll find elsewhere.
They can probably get away with it in the sales process by saying “we give you FreeSWITCH from the source”, but I am not sure what happens when developers want to configure that elastic cloud service the way they are used to be doing with their own FreeSWITCH installation.
All in all, this is an interesting offering and an interesting approach and go to market.TeleSign – Security and Data Analytics
TeleSign is focused on SMS. And a bit of voice. As their website states: “APIs Delivering User Verification, Data Insights & Communications”
Since security, verification and fraud prevention these days rely heavily on analytics, TeleSign are “horeding” data about phone numbers, using it for these use cases. It isn’t that others don’t do it (there’s Twilio Authy, nexmo Number Insight and others), but this is what they are putting front and center.
Since their acquisition by BICS, a wholesale operator for wireline and wireless carries, that has grown even further, as they gain access to more and more data.
It will be interesting to see how TeleSign grows their business from security to additional communication domains, or will they try to focus on security and expand from the telecom space to adjacent areas.Twilio – Adjacencies
Talking about adjacencies, that’s what Twilio is doing. Now that they are a public company, there is even more insatiability for growth within Twilio, in an effort to find more revenue streams. So far, this has worked great for Twilio.
Here are two areas we’ve seen Twilio going into:
How email fits into the Twilio communication APIs is still an open question, though I can see a few interesting initiatives there.
And then there’s the wireless offering of Twilio, which resembles a more flexible M2M play.
But where would Twilio go next?
UCaaS, going after unified communications vendors and competing with them head to head?
Maybe try to jump towards an Intercom like service of its own? Or purchase Intercom?
Or find another market of developers that is growing nicely – similar maybe to its recent Stripe integration of Twilio Pay.
Twilio in a way has been defining and redefining what CPaaS is for the past several years. They need to continue doing that to stay in the lead and well ahead of their competition.VoIP Innovations – Marketplace
VoIP Innovations came out with what they call Showroom.
Here’s a short video of the explanation of what that is exactly:
Many of the CPaaS vendors offer a partner program of sorts. This is where vendors who develop stuff for others or build tooling and apps on top of the CPaaS vendor’s APIs can go and showcase their work. The programs vary from CPaaS company to another.
What makes it different a bit is the target audience associated with it:
While there isn’t much documentation to go about, I am assuming that the whole intent behind the marketplace is to offer direct monetization opportunities for developers and resellers by taking care of customer acquisition as well as payment on behalf of the developer and reseller.
A concept taken from other marketplaces (think mobile app stores). It will be intersting to see how successful this will be.Vonage – UCaaS+CPaaS
Vonage is interesting. Started as consumer VoIP, turned cloud UC vendor (=enterprise communications) through acquisitions, turned to acquire Nexmo and then TokBox to add CPaaS, continued with NewVoiceMedia acquisition to cover contact center space.
How does one differentiate in such a way? Probably by leveraging synergies across its product offerings and markets.
What Vonage recently did was bring number programmability from its Nexmo/CPaaS offering to its VBC/UCaaS platform.
What do they gain?
Is this good for Nexmo customers and partners? Yap. They can now reach out to the Vonage business customers as an additional target market.
Is this good for Vonage customers and partners? Yap. They can now do more, and more customized communications solutions with this added flexibility.Voximplant – Flow
Voximplant is one of the lesser known CPaaS vendors. Its whole platform is built on the concept of an App Engine, where you write the communications logic right onto their platform using Java Script. It is serverless from the ground up. A year or two ago, Voximplant added Smartcalls. A product that enables you to sketch out call flows for outbound interactions – marketing, sales, etc. These interactions would be played out across a large number of phone numbers and get automated, making it really easy and flexible to drive phone based campaigns.
Now? Voximplant took the next step of adding inbound interactions, covering the IVR and contact center types of scenarios.
Twilio, MessageBird and Plivo offer inbound visual flow products. These allow developers to drag and drop communication widgets to build a flow – a customer interaction through the system.
Voximplant and Infobip offer inbound and outbound flows, where you can also plot company/agent based initiatives with greater ease as well as the customer initiated interactions.Why aren’t you listed here?
The CPaaS market is large and varied. It is hard to see everyone all the time. It is also hard to innovate and differentiate every year. The vendors here are the ones I had briefings with or ones who promoted their products in ways that were visible to me. But more than anything, these are the ones that I felt have changed their offerings in the past year in a differentiating manner.
BTW – if you think that differentiation here means that it is a functionality that other vendors don’t have then you are wrong. Doing that is close to impossible today. Differentiation is simply where each vendor is putting his focus and trying to attract customers and carve his niche within the broader market. It is the stories each vendor tells about his product.
If you feel like a vendor needs to be here, or did something meaningful and interesting, just contact me. I am always happy to learn more about what is happening in the market.Who is missing in my WebRTC PaaS report?
Later this month, I will be releasing my latest update of the WebRTC PaaS report.
There are changes taking place in the market, and what vendors are offering in the WebRTC space as a managed API service is also changing. This report is there to guide buyers and sellers in the market on what to do.
For buyers, it is about which platform to pick for their project – or in some cases, in which of the platform vendors to invest.
For sellers, it is about what to add to their roadmap. To understand how they are viewed from the outside and how do they compare to their peers.
Here’s who’s been in the last update of the report:
Think you should be there? Contact me.
Want to purchase the report? There’s a 30% discount on it from today and until the update gets published (and yes – you will be receiving the update once it gets published for no additional fee).
There will be a new appendix in the report, covering the topic of Flow and Embeddable trends in the market. Something that will become more important as we move forward.
This demo of the Microsoft Surface Hub 2 is pretty damn cool…
I don’t run a lot of Microsoft product anymore, switched to mac when the intel chip landed + Apple moved to a unix underpinning. That said, I have seen much better quality in products coming from Microsoft in the last few years, so maybe they deserve a second look.
Surface Hub 2 sort of reminds me of a product called Perch, built a by a local Vancouver team which was meant to serve as a portal into disparate global offices. Perch was way before it’s time. WebRTC was still in its infancy and personal device video conferencing had not really crossed the chasm, which is a shame considering where we are today.
Now there are many of video conferencing companies and products, and plenty of alternatives / platforms for developers to build on. It certainly seems plausible now that we could see the Microsoft Surface Hub 2 in boardrooms across the globe. Apparently it will be interoperable with WebRTC endpoints as well, which could make this a powerful work tool indeed. That would enable collaboration with peers over IP on various endpoints including laptops, tablets and mobile, regardless of the OS. Sharing product ideas, riffing on concepts and polishing final features on a product release using the Microsoft Surface Hub 2 as a tool, could be a refreshing new way to work.
It will be interesting to see what developments come about from the Microsoft press event in NYC in April, as reported by The Verge.
I haven’t blogged here in some time, so I figured that since the topic is relevant this would be good a good opportunity to dust off the old blog (webrtc.is / sipthat.com) and post something we have been working on at SignalWire. I am quite passionate about WebRTC and real-time communications so it’s great to be helping bring it to life at SignalWire!
We all know and love <cough> SIP, so we decided we would enable the use of SIP over WebSockets at SignalWire. This new offer also enables functionality like WebRTC with SIP over WebSockets.
This means our customers can now use off the shelf JS libraries, like JSSIP to create basic web experiences for their users, powered by SignalWire. It used to be a bit of a PITA, to create services that provided users with seamless online communications. Now it’s a breeze, and when using SignalWire it’s also very affordable.
For now, we are enabling basic calling and video capabilities, the advanced functionality (including video conferencing) will come in conjunction with a future release of a SignalWire RELAY JS library.
Personally, I can’t wait to see what creative minds will build using this technology with SignalWire on the backend.
WebRTC doesn’t really connect people, but the way you think about it signaling is important to your WebRTC application.
WebRTC is… still just a little confusing…Tsahi, i’m reading the book recommended by Loreto & Romano but the examples are outdated. With regards to the SDP signal – if peer A is on a webRTC application, but peer B is surfing youtube – How does peer B get notified of an offer? It would have to go to peer B’s email address right? — because there is no way of knowing peer B’s IP address. Please help.
A few quick things before I dig deeper into this WebRTC connectivity thing:
How well do you know WebRTC? Check it out in my online WebRTC quiz.
I’ll try to use a kind of a bad comparison here to try to explain this.
Let’s say you are the proud owners of a Pilates studio. You’re the instructor there (#truestory – at least for my wife).
My wife gives Pilates lessons at different hours of the day. These are private lessons so it is rather flexible on both sides. But let me ask you this – how do people know when to come for a lesson?
This being Israel, they usually communicate with my wife via Whatsapp to decide together on the date and time. Usually, people stick to the day of week and time and start communicating only if they can’t make it, want to reschedule or just make sure the lesson is still taking place.
Back to WebRTC.
WebRTC is that Pilates studio. It does one thing – enables live media to flow from one browser to another. Sometimes also non-browsers, but let’s stick to the basics here.
How do the people who need to share or receive that live media connect to each other? That’s not what WebRTC does – it happens somewhere else. And that somewhere is the signaling mechanism that you pick for your own application. I am calling it a mechanism and not a protocol, since it is going to be a tad more confusing in a second.
Now let’s go back to WebRTC, signaling and connecting people and look at it from a point of view of different scenarios.Scheduled Meeting
We’ll start with a scheduled meeting. At any given point in time, I have a few of those coming up. Meetings with clients, partners and potential clients. Here’s one such calendar invitation:
This one happens to take place using Google Meet. Who’s calling who? No one really. I’ll just click that link in the invite when the time comes and magically find myself in the same conference with the other participants.
In most scheduled conferences, you just join a WebRTC link
Where do you get that link to use?
Some of these services allow inviting people from inside the meeting. That ends up being sent to them via email or an SMS as a link or just dialing their phone (without WebRTC).Ad-hoc “upgrade” of text chat to video conference
There are ad-hoc calls. These usually start from a chat message.
Often times, I’d rather text chat than do a voice or a video call. It has to do with the speed and asynchronous nature of text. Which means that I’ll be chatting with someone over whatever instant messaging service we select, and at some point, I might want to switch medium – move from text to something a bit more synchronous like video:
Like this example with Philipp – most of our conversations start in Hangouts (that’s where he is most reachable to me) and when needed, we’ll just jump on a call, without planning it first.
Who is calling whom here? Does it matter?
What happens here is that both of us are already “inside” the communications app, so we both have a direct link to the service. Passing that information from one side to the other is a no brainer at this point.
So how will that get signaled? However you see fit. Probably on top of a Websocket or over HTTPS.I am calling you on the “phone”
What if there’s nothing pre-planned, so it isn’t a scheduled meeting. And we haven’t really been on a text chat to warm things up towards a call. How do you reach me now?
How do you “dial”?
Puneet is one of our support/testing engineers at testRTC. While he will usually text me over slack to start a call, he might just try calling directly from time to time.
What happens then?
I am not in front of my laptop with the Slack app opened. My phone is on standby mode. How does it start ringing on me? What does WebRTC do to get my attention?
The phone starts dialing because it received a mobile push notification. I’ve got the Slack app installed, so it can receive push notifications. Slack invoked a push notification to wake up the app and make it “ring” for me.
The same can be done with web notifications. And there are probably other means to do similar things in IOT devices. The thing is – this is out of scope for WebRTC, but something that is doable with the signaling technologies available to you.Contact center agent answering calls
When a contact center adopts WebRTC to be able to migrate its agents from using desktop phones or installed softphone towards WebRTC, calls will end up being received in the browser.
This happens by integrating callbars inside CRMs or just by having the CRM implement the contact center part of the equation as well.
What happens then? How do calls get dialed? (the above is a screenshot taken from Talkdesk’s support site)
They go through PSTN towards a PBX. More often than not, that PBX will be based on Asterisk or FreeSWITCH, though other alternatives exist. PBXs usually base themselves around the SIP protocol, which will lead to two alternatives on the signaling protocol that will be used by WebRTC in the browser:
In both cases, the contact center agent is registered in advance. It is also marked as “available” in most contact center software logic – this means that incoming calls waiting in the call center queue can be routed to that agent. So it is sitting and waiting for incoming calls. In some ways, this is similar to the upgrade from text chat scenario.Connecting? WebRTC?
When it comes to actual users, WebRTC doesn’t get them “connected”. At least not from a signaling point of view.
What WebRTC does is negotiate the paths that the media will use throughout the session. That’s the “offer-answer” (or JSEP) messages that pass between one WebRTC entity to another. And even that isn’t sent by WebRTC itself – WebRTC creates the blob of data it wants to send and lets your application send it in any way you see fit.
Still confused? There’s a course for that – my online WebRTC training. The first module (out of eight modules) is free, so go learn about WebRTC.
WebRTC wins over competition because there is no competition – browsers offer only WebRTC as a technology for web developers.
It was raining and miserable this last Saturday. I had lost of ideas for articles to write for BlogGeek.me in my backlog, but none of them really inspired me to action. The 8yo went to his cousin. The wife had her own things to do. My 11yo daughter was bored to death. She comes to me and says: “Can we do a trip outside to the park? I need some fresh air.” How could I answer besides saying yes?
The rain stopped a bit, so we went outside. What she really wanted wasn’t fresh air, but a chaperone to the closest candy vending machine. They are having a game at school for Purim, where she needs to bring small presents and candies to another kid in her class without her knowing who is pampering her. She needed an extra candy.
How is this related to WebRTC? It isn’t.
When I asked her about her plans for this game, she mentioned the trinket she planned on giving today –
2 mechanical pencils.
And that’s definitely WebRTC related.
A quick conversation ensued between me and my daughter – are these 0.5 mm or 0.7 mm point type? My daughter went to explain that it might even be 0.9 mm.
So many alternatives.Competing standards
It got me thinking:
With analog video recording we had VHS and Betamax.
Paper size? A4 and Letter.
Power frequency? 50 Hz and 60 Hz.
With VoIP signaling we had H.323 and SIP. And also XMPP.
Audio and video codecs? A shopping mall of alternatives.
Web browser streaming? HLS and MPEG-DASH.
Inches and Meters. Left side vs right side driver in cars.
The list is endless.WebRTC standard
But browser based real time media communications?
There. Is. No. Other. Alternative.
We had that short romance around ORTC, which ended with ORTC dead and its main concepts just wrapped back into WebRTC.
What other technology would you use or could you use inside a browser to do a video call?
The other alternatives just don’t cure it (including what Zoom is presumably doing).
What does that mean exactly? It gives us a kind of a virtuous circle.
For the most part, there’s no question if you should select WebRTC these days. There’s also no question what are the alternatives (there usually are none). It isn’t a question if WebRTC is getting adopted, used, growing or popular.
When our window to the world is the browser, then WebRTC is what you use.
For mobile apps or other devices, the need for browsers or just having an ecosystem around the technology picked translates again to WebRTC.
Thinking of using real time media technology? That’s synonymous to WebRTC.
Want to learn more about WebRTC? Check out the first module of my online course – it is free.
I know I am. I am constantly surprised what people are doing with WebRTC.
Here’s something I hear a lot:
How do you make a call with WebRTC?
Well… you don’t. Not really. And in many scenarios – that term call, or dialing, or answering – has no real meaning.
Here’s a funny opposite for you:
Kids in front of old phones don’t know what to do. It isn’t “natural”. Guess what? Nothing is. The things that are natural to you are things you’ve learned, and are now used to. They are a set of rules in your upbringing.
If you come from a VoIP background, then WebRTC brings with it quite a challenge to your world. I know – I had 13 years of VoIP background before WebRTC was announced. Since that announcement, I’ve been surprised time and again by what people are doing with WebRTC. Especially people who shouldn’t be able to even use it because they don’t know VoIP enough.
Coming from VoIP? Interested in streaming? Broadcasting? Some other communication use cases? Tomorrow I am hosting a free webinar – Google Does Gaming: WebRTC Man-to-Machine Use Cases
When we all first started out in this adventure called WebRTC, what we’ve seen was video calling. It was all about face to face meetings. It took time to think about WebRTC in other settings and for other use cases.
And here we are. Years later, dealing with WebRTC in the aid of cloud gaming. Google used WebRTC in Project Stream, where they showcased playing the game Spartan through a web browser – the game itself was rendered in Google’s cloud.
Who would have thought WebRTC would be used for that?
Anyways, if you come from a VoIP background, here are some aspects of WebRTC you’ll need to unlearn and relearn – I am still grappling with them myself every once in awhile:Signaling? What’s “Signaling”?
With any other VoIP protocol out there, it seems like we’re starting off with signaling.
SIP? That’s signaling.
WebRTC? Nope. No signaling. Sorry.
What does that mean exactly? That you can use whatever signaling mechanism/protocol you see fit. That’s assuming you can get it to run inside a web browser or wherever it is your application needs to operate.
SIP, which is the most popular VoIP signaling protocol out there, is probably an overkill for a lot of WebRTC services. I tend to look at it as a hindrance when I see it in architectures – I often ask time and again why is it there to make sure there’s a real need other than saying someone needed signaling for his WebRTC application.You. Don’t. Answer. Calls.
There’s no such thing as a call while we’re at it.
I remember doing a live WebRTC training a couple of years back. I had to hammer out of the people the need to ask incessant questions about dial, answer, mute, hold and a bunch of other paradigms they thought are golden rules in communications.
If you feel that way too, then look at that video at the top of this article again. What made sense 20 years ago doesn’t hold water today.
WebRTC isn’t fixed in any specific concept of how “calls” are made. I prefer using the term session and deal with the initiation part of it on a case by case basis.
If there’s no need for dialing or answering – just don’t force it on your WebRTC solution.It isn’t only Google
Most days of the week, I like thinking of WebRTC as the source code that resides on webrtc.org. That’s the codebase Google is maintaining and putting inside its Chrome browser.
The thing is, many end up modifying it for their own needs. They:
There are some really interesting “mods” to the vinyl WebRTC implementations out there, usually held privately for internal use of companies. In many ways, this is a shortcut to building your own media engine from scratch.There’s more than one way
What I like about WebRTC is that usually, there’s a single way of doing things with it: everything is encrypted – you can’t override that; it defaults to multiplex and bundle its media connections; the list goes on.
How you use it is a totally different story.
Each SFU implementation is different than the other. There are different ways to record a session. Different ideas and approaches to broadcasting at low latency.
The “right” answer differs a lot not only based on the use case, but also on the business model, the developers available, the DNA of the company, etc.Wasteful can be just fine
There’s also a school of thought that never really existed with VoIP: the “good enough” approach – one where we’re just fine with not optimizing everything and leaving things it a kind of a mediocre stage that is good enough for what we’re trying to do. It may eat up to much bandwidth or tax on the CPU. Or just not be how things are done around here. But it works. Good enough.
Heck – the default WebRTC implementation does it on its own, deciding to waste 1.7Mbps for a VGA resolution encoding instead of limiting it to 800kbps or less. Such a waste of good resources.
I learned to love this approach (and then try to optimize it with my clients).How do you think about WebRTC?
What about you?
What mistakes you see people make when thinking about WebRTC that fits the web or VoIP better?
What things do you need to unlearn about WebRTC?
Coming from VoIP? Interested in streaming? Broadcasting? Some other communication use cases? Tomorrow I am hosting a free webinar – Google Does Gaming: WebRTC Man-to-Machine Use Cases
The post Are you blocked by the rules of your upbringing in your WebRTC application? appeared first on BlogGeek.me.
Thanks to work initiated by Google Project Zero, fuzzing has become a popular topic within WebRTC since late last year. It was clear WebRTC was lacking in this area. However, the community has shown its strength by giving this topic an immense amount of focus and resolving many issues. In a previous post, we showed how to break the Janus Server RTCP parser. The Meetecho team behind Janus did not take that lightly. They got to the bottom of what turned out to be quite a big project.
Some believe WebRTC isn’t ready. I think it is ready. But when will WebRTC 1.0 be available?
Ready or not, WebRTC is here. The thing is, we still don’t have a closed standard specification we can all print and take on a plane to read for our enjoyment. There are drafts – but nothing that is final.
And once final, does it mean that it is available?
There are 3 parts that needs to be addressed to answer this question. I’ll deal with only two of them (skipping the IETF one):
Want to learn more about WebRTC, the various components in its specification and what compute power you need for each WebRTC server? Try out my free video course:
Learn about WebRTC servers
Want to learn more about WebRTC, the various components in its specification and what compute power you need for each WebRTC server? Try out my free video course:WebRTC standardization
WebRTC as a standard is built out of two components:
Most of the industry is already viewing WebRTC as a done deal – so much so that the IETF already has an RFC for SIP over WebSocket. The only reason to have such an RFC is to be able to use SIP inside a browser, and the only way to use SIP inside a browser with media being sent or received would be by way of WebRTC. The people working at the IETF were so certain WebRTC will get an RFC of its own in 2014 already (5 years ago!).
Each of these organizations has its own set of rules, policies, governance and flow.
I’ve tried to keep the standardization of WebRTC at arm’s length. In the past I’ve been part of standardization processes related to H.323 and 3G-324M, going to ITU-T and 3GPP standardization meetings as well as acting as a co-chair of the 3G-324M activity group at the IMTC (dealing with interoperability). It is a tedious work that combines technology with politics. As fun as it is (at times at least), dealing with it as an employee of a company is different than doing it as a consultant. The value for me just wasn’t there.
For vendors? If you want to take a driver’s seat at this, and decide what gets more attention, then you should invest time in it.
But where are we with WebRTC then?W3C WebRTC status
I’ve asked Dominique Hazael-Massieux about WebRTC’s status. He works as a W3C Staff dealing with WebRTC. Here’s what I got –
When it comes to W3C, where the browser WebRTC APIs are being defined, WebRTC is considered to be at the CR stage.
CR means a Candidate Recommendation. We’ve moved from a Working Draft (WD) towards a Candidate Recommendation.
Next up would be PR – Proposed Recommendation, and from there, a Recommendation.
How do we move to the next step?
That first one is “easy”. Get the people writing the spec into a room. Have them agree. Then have someone write down the agreement on “paper”. Get everyone to read it. And agree again. Rinse and repeat. It’s never easy.
That second one of implementing in browsers? That’s also not easy. They have other things on their minds as well. And WebRTC is pretty darn complex to implement. But we’re getting there.
That third one of interoperability testing? With a test suite. That tests for the various features? This is downright suicidal. And daunting.
All that work needs to be done for “free”. There’s no direct money to be made out of it. But lost of hours needs to be spent by many people to get it done. We’re getting there, but we’re not there yet.WebRTC 1.0 browser implementation
And then there are the browser implementations.
The specification is as good as its implementations. People always complain when I suggest following the Chrome behavior in WebRTC as opposed to implementing against the specification. That’s where theory and expectations meets reality.
At the end of the day, your service will need to:
In the first case, Chrome wins on market share; Microsoft Edge will be migrating to Chromium. And for most use cases, Chrome is the first browser to target anyway.
In the second case, if you are using the code in webrtc.org for your app, then you are effectively basing your app on Chrome’s WebRTC implementation.
Better go with what’s available now than what will be ready some time in the future.
In the past, the changes we’ve seen in browser implementations of WebRTC revolved a lot around media optimizations and interoperability across browsers. What we are seeing now a lot more is changes in the API layer, where browsers are shifting towards the WebRTC 1.0 specification. This is necessary because:
These changes mean one sad thing though. You can be certain in one thing – during 2019, WebRTC implementations in browsers is going to break existing apps multiple times. This is due to the changes taking place. We are seeing migration from Plan B towards Unified Plan, modifications to the connection state machine, and an experimental implementation of mDNS. There’s more that I probably forgot and more ahead of us still.
The only certainty is that nothing is certain. You’ll need to continue investing in aligning with the browser implementations with each and every browser version release.When then?
The current intent is to be able to get to the PR stage for WebRTC somewhere in Q3 2019. Will it be postponed further? I don’t really know.
Interestingly, work has started in parallel about WebRTC NV – what comes next. I’ve covered the WebAssembly in WebRTC part of it in the past.
Want to learn more about WebRTC, the various components in its specification and what compute power you need for each WebRTC server? Try out my free video course:
Learn about WebRTC servers
WebRTC is a great piece of technology, assuming you can develop a coherent strategy on how you plan on using it.
There are two extremes happening in the enterprise communication space, and they are quite opposite in nature. On one hand, companies are striving towards more automation and this is coming to their contact centers by way of machine learning and bots “replacing” humans. On the other hand, many of us are striving for better and more meaningful communications. Be it for long distance relationships (personal as well as business ones) or by the use of machine learning (again) and context, to guide us through an interaction – being able to know beforehand the intents of people for example.
Enter WebRTC, which enables communications to take place anywhere – be it a mobile application, a physical device or a modern web browser. What WebRTC brings with it is better context of sessions and lower barrier of entry for enterprises to make use if this technology. Some enterprises use it to improve business agility or lower their operating costs. Others use it to create new businesses never before seen or to improve the communications with their customers or peers in the industry.
We are now 7-8 years since the announcement of WebRTC (depends on who’s doing the counting and from which date), but in many ways, a lot of enterprises (I don’t want to say most) have failed in to capture the value they initially envisioned from using WebRTC. In many cases, the lack of any thoughtful strategy created a rush towards initiatives that never really matured.
Through my work with many clients on their WebRTC initiatives along with discussions with many others on their projects and services – failed as well as successful ones, I’ve seen a few challenges that crop up consistently across such initiatives.#1 – Where to begin?
WebRTC is a versatile and powerful building block in your arsenal. This means that you can do a lot with it. That range of utility can be overwhelming, oftentimes leading to wasted resources. The other problem is that WebRTC can’t do everything, while the expectations of it are rather high. This leads to requirements and plans that are often not grounded in what can be done in reality or within the allocated budget and resources.
Deciding what to build using WebRTC requires an understanding of the capabilities and limitations of WebRTC coupled with a clear view of the communication problems you are trying to solve for your customers. There’s a lot of feature creep happening when it comes to WebRTC. I find myself asked about a simple video chat service for 2 people, but once you dig a layer deeper, you see requirements for group video calls, recording and even broadcasts as part of the project. Being able to see the full picture, and map it back into requirements and a roadmap comprised out of multiple phases is an important first step in any WebRTC initiative.
There are a few other things to keep in mind –Integration with existing infrastructure
Oftentimes, you’d be planning on adding WebRTC to an existing service. This can happen in many ways:
This requires extra care in how WebRTC gets introduced as it isn’t going into a green field where anything you pick immediately fits your needs.Cloud migration and transformation
WebRTC was born in the cloud era. Many of its deployments are cloud based.
Most of its uses in non-cloud environments are actually enabling guest access from the public cloud towards the internal communications infrastructure. In other cases, it just needs to integrate with on premise data centers for things like users database and policies.
This places an additional strain on enterprises who are just starting out their migration towards the cloud.Not your regular web application
WebRTC is different than other web technologies. It has a lot more moving parts to get to a minimal viable product, and then there’s that media quality issue to contend with. Its deployment needs to start as a global one for many of the use cases.
What are the server side components needed for WebRTC? Learn that in my free online mini video course.
Register now#2 – Who should I have on my team?
Putting a team of developers on a WebRTC initiative is a daunting task. There are multiple disciplines they need to come from and the myth of a full stack developer that can do it all gets stretched even further here, as that superhero needs to also know about media processing, WebRTC APIs, browser changes and standardization processes.
Here’s what i wrote a while back about WebRTC developers after discussing the topic with a few people who manage/hire them.
Some other aspects you’ll need to decide on:Internal vs External
Will you be relying on your existing engineering team or will you be outsourcing some/most of the project to an external vendor? Assuming you decide to go for an external vendor, who will maintain the service on an ongoing basis?Multidisciplinary
The team in question needs to be multidisciplinary, capable of handling anything from media processing, to mobile app development, to backend integration work and ongoing DevOps and maintenance.
There needs to be a skilled product manager and a system architect who understand WebRTC enough to know what is possible and what’s… less possible. What incurs risk and where quick wins can be found.Which new skills are needed?
Your teams. Do they have the necessary skills?
Here it goes to a lot more than just developers. There are product managers, testers, DevOps people, support staff.Do I need to enhance some in-house capabilities?
What skills are you missing? If you operate everything on premise and WebRTC is forcing you to start using cloud services, then this is an in-house capability you will need to start contending with.
The same goes for mobile application development, going global in how you deploy servers, etc.
Looking to beef up the WebRTC experience and skills of your team? Check out my WebRTC training (the first module is free).
Different companies have different DNA to them. That often dictates what their technology stack will look like and how they’d prefer to partner/hire.
There are three main aspects that need to be taken into account when picking a WebRTC technology stack:Open source / commercial
You might favor open source components and frameworks for your WebRTC service or you might be someone who prefers a commercial offering with a company focused on that product development.
Both alternatives can come with support contracts but companies seem to prefer one or the other.
Which alternative will it be for you?Hosted or on prem?
These two approaches means different technology stacks, levels of expertise and staffing on your end.
Are you planning on hosting this on your own, in your data centers, on bare metal or in the cloud? Or are you going to have someone else host the service for you? Which parts of it will be managed and which will be self managed?Acquisitions
WebRTC is still relatively new, with the vendors ecosystem dynamically shifting. There have been quite a few acquisitions in this space. These acquisitions sometimes removed solutions from the market, made them weaker or made them stronger.
When selecting a technology stack, the potential acquisition scenario of the vendors in question needs to be taken into consideration as well.Fit for the requirements
This one seems silly but it is highly relevant and important.
Are you sure the technology stack you’ve selected can do the things you want it to do?
I’ve seen too many cases where the framework used wasn’t up for the task. Things like taking signaling when media servers needs to be used, picking a CPaaS vendor when the scenario requires too much control of media processing, etc.
Just look at what WebRTC signaling alternatives people have these days.#4 – How do I know it is working?
You built it. Tested it in the lab. Did a call or two with your colleagues. Went home and showed it to a friend.
Does it scale? Will it work properly?
I had a customer recently who is developing a group video calling feature. He wanted to test the service with around 20 people in a single room. It wasn’t easy to find 20 people to run that one scenario. And when he did – things broke and needed fixing. So he had to find 20 people to run it again once a fix was put in place.
Testing is often neglected when it comes to WebRTC applications and it shouldn’t be. Take this one seriously. You can cobble up a testing environment on your own (there are even a few open source projects that can help you out here) or you can just use testRTC (I am a co-founder there) and start running tests within a couple of hours.#5 – What do I track?
Tracking websites is rather “easy” these days. Use Nagios, Cacti, Zabbix or any other open source tool that sounds like a disease. Or use something like New Relic or DataDog to do it managed in the cloud.
Problem is, these tools only cover the machines metrics and performance and they don’t really watch for the media and its quality (or even if a session got connected for that matter). There’s no end to end monitoring/tracking.
You will need to collect WebRTC related metrics from either the backend or the devices (or both). You’ll need to track it for quality.
You’ll need to monitor your service (we’re doing a webinar on WebRTC monitoring next more @ testRTC – register to join).How can I get help?
There are various ways in which you can get some help for what you are doing.
The best approach is probably to get some external assistance in what you are doing as part of your research and planning – even before you go outsourcing the whole project (if that’s the path you are going to take).
You can contact me for that, or go to other consultants. Some of the outsourcing vendors offer such consultancy service as well. Whatever you do – don’t go it alone. At least not in the planning stages.
The post The five make-or-break WebRTC challenges you need to address appeared first on BlogGeek.me.
Is QUIC in WebRTC a solution looking for a problem or a real requirement?
QUIC is the next evolution of browser transport protocols. I’ve written about it in 2015, when Google started experimenting with the idea of replacing SCTP with QUIC for data channels. Three and a half years later, and we still don’t really have QUIC in WebRTC – at least not until last month. Google decided to come out with a new RTCQUICTransport for WebRTC in Chrome and written a post about it on their Chrome Developers site.
UDP, TCP, SCTP & QUIC. How do these transport protocols compare?
Download my free Transport Comparison TableWhat is QUIC again?
I am not going to go into the technical details – I’ve done that in the past already, and there are other places for that. I want to focus here on the bigger picture.
If you look at the timeline of web transport protocols, it looks something like this:
We had TCP and UDP for some 40 years now. HTTP 1.1 is defunct, but runs most of the internet at the moment. HTTP/2 is growing nicely in adoption. According to W3Techs, we’re standing on ~33% adoption for HTTP/2 (Feb 2019):
HTTP/2 came to be after Google came out with SPDY, a “fix” for HTTP and got parts (most?) of it wrapped into HTTP/2 to get it standardized.
HTTP 1.0, 1.1 and HTTP/2 are all built on top of TCP. Signaling, which requires reliability and causality won’t work on top of UDP without adding these characteristics. After around 40 years, it is time for a refresh. Enter QUIC. It uses UDP and works in ways that are better than TCP for signaling purposes.
QUIC follows a similar path – Google created it to “fix” the ailments of HTTP over TCP. the end goal here is to turn it into HTTP/3.
Since QUIC is built on top of UDP, it can handle a lot more than just HTTP signaling. Which is why it is becoming an interesting topic for WebRTC –Where QUIC in WebRTC fits exactly?
This is the real question. My answer to it in 2015 was this:
There are two places where QUIC fits in WebRTC:
1. In the signaling, which is out of scope of WebRTC, but interesting, as it enables faster connection of the initial call (theoretically at least)
2. In the data channel, by replacing SCTP with QUIC wholesale
Google’s answer in their post on Chrome Developers blog?
A powerful low level data transport API can enable applications (like real time communications) to do new things on the web. You can build on top of the API, creating your own solutions, pushing the limits of what can be done with peer to peer connections, […] WebRTC’s NV effort is to move towards lower level APIs, and experimenting early with this is valuable.
The QUIC protocol is desirable for real time communications. It is built on top of UDP, has built in encryption, congestion control and is multiplexed without head of line blocking.
Hmm… somehow they lost me in that explanation somewhere. This is about real time communications. It is about doing stuff on top of UDP. And it is about low level APIs. Great. Why do I need it again? For voice and video I already have SRTP in WebRTC. The SCTP data channel works quite well. So where exactly do I need this great thing called QUIC in WebRTC?
I think there’s merit, but it is in totally different places.
QUIC is about having a single, modern, common transport protocol for the web.
Here’s what we do today with WebRTC in terms of transport protocols:
There’s this popular drawing from the High Performance Browser Networking book that shows this amalgamation of protocols:
So many transport protocols in a single standard. This makes implementations of the backend more complex, as they need to be able to understand all these transport protocols as well. One can say that this is already common enough and widely used already that it is a solution looking for a problem, but the developer in me can appreciate unifying all these functionality over a single transport protocol.
Here’s how life will look like with QUIC in WebRTC:
Putting it into an architecture diagram of my own, we get this:
Much simpler.What do we gain?
Theoretically, we can multiplex signaling, voice, video and low latency data in a single QUIC connection. That’s powerful:
This isn’t going to happen in a day. Getting there is going to be a journey of multiple years and people will complain and whine about it along the way. Similar to what is happening today with WebRTC – whenever something is modified or something new is added – things tend to break (either because APIs get deprecated, behavior changes or just pure bugs).
Moving to a QUIC based stack is a huge undertaking – for the WebRTC stack, browser vendors and all the related internet infrastructure vendors.
Connecting to other realms such as SIP? That’s going to get even harder, as we move away from the domain of SRTP towards QUIC, more translations and protocol interworking will be required.
The question then becomes – is it worth all the fuss? Are we gaining enough to make this effort worthwhile?Can you use QUIC in WebRTC now?
To some extent you can. Check out the recent post on QUIC @ webrtcHacks for that.
I will be adding a new dedicated lesson to my online WebRTC course about QUIC – my goal is to have the most up to date and relevant WebRTC training curriculum in the market, so keeping up with these changes comes with the territory.
Interested in WebRTC? Check out my WebRTC course.