Jitsi: 20 years of Real Time Communications
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
Title |
| |
Subtitle |
| |
Title of Series | ||
Number of Parts | 287 | |
Author | ||
License | CC Attribution 2.0 Belgium: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
Identifiers | 10.5446/57068 (DOI) | |
Publisher | ||
Release Date | ||
Language |
Content Metadata
Subject Area | ||
Genre | ||
Abstract |
|
FOSDEM 2022266 / 287
2
4
6
8
12
17
21
23
31
35
37
41
44
45
46
47
50
62
65
66
67
68
71
73
81
84
85
86
90
92
94
100
102
105
111
114
115
116
117
118
121
122
124
127
131
133
135
137
139
140
141
142
145
149
150
156
164
165
167
169
170
171
172
174
176
178
180
183
184
189
190
192
194
198
205
206
207
208
210
218
220
224
225
229
230
232
235
236
238
239
240
242
243
244
245
246
249
250
253
260
262
264
267
273
274
277
282
283
287
00:00
DiagramEngineering drawing
00:41
Projective planeStandard deviationInternetworkingJava appletBefehlsprozessorBitMultiplication signUniform resource locatorTelecommunicationReal-time operating systemSemiconductor memoryAuthorizationWritingOpen sourcePoint (geometry)Position operatorOpen setVideoconferencingUniverse (mathematics)SpacetimeCASE <Informatik>Process (computing)Product (business)Moore's lawRing (mathematics)Order (biology)Descriptive statisticsServer (computing)Revision controlSession Initiation ProtocolNumberInternettelefonieVolume (thermodynamics)Limit (category theory)Constraint (mathematics)Service (economics)Generic programmingData conversionExistenceLibrary (computing)SineQuicksortAxiom of choicePartial derivativeMereologyRule of inferenceFacebookYouTubeSurvival analysisSoftwareVideo gameGame theoryImage resolutionMeeting/Interview
10:24
Table (information)FamilyOpen sourceFrictionProduct (business)Maxima and minimaDependent and independent variablesMultiplication signRule of inferenceProjective planeEuler anglesBitLevel (video gaming)Student's t-testDifferent (Kate Ryan album)Ring (mathematics)Cartesian coordinate systemCollaborationismFigurate numberInheritance (object-oriented programming)SineForm (programming)Goodness of fitMobile appEquivalence relationMachine visionQuicksortRight angleOrder (biology)Computer programmingCodeClosed setDesign by contractTask (computing)Operating systemVapor barrierNormal (geometry)Standard deviationLocal ringOpen setVideo gameAngleBuildingSource codeHand fanMoment (mathematics)TelecommunicationArithmetic meanSoftware developerRegular graphPoint (geometry)MereologyBootstrap aggregatingMeeting/Interview
20:04
Real-time operating systemDirection (geometry)VotingService (economics)Product (business)Latent heatQuicksortNeuroinformatikOpen sourceTelecommunicationComputer configurationSoftwareGame controllerProjective planeConfiguration spaceClient (computing)Traffic reportingBusiness modelMultiplication signRight angleStandard deviationIntegrated development environmentCodeShift operatorSlide ruleGroup actionWeb browserFocus (optics)Process (computing)Rule of inferenceOrder (biology)Session Initiation ProtocolLine (geometry)CASE <Informatik>Presentation of a groupRing (mathematics)Connected spaceInternet service providerVideo gameOpen setDifferent (Kate Ryan album)Task (computing)System callCellular automatonMereologyReal numberPoint (geometry)Likelihood functionPivot elementDegree (graph theory)Interactive televisionSelectivity (electronic)Meeting/Interview
29:43
System callAddress spaceACIDEmailServer (computing)Connectivity (graph theory)Group actionGoodness of fitFunctional (mathematics)Library (computing)Rule of inferenceDifferent (Kate Ryan album)VideoconferencingMathematicsProduct (business)System callShift operatorWeb 2.0Multiplication signHypermediaMereologyService (economics)Open sourceCodeJava appletCore dumpBuildingBlock (periodic table)Client (computing)Internet service providerPoint cloudPay televisionFocus (optics)Meta elementRoutingHuman migrationTelecommunicationPoint (geometry)Mixed realityGame theoryStandard deviationConnected spaceBefehlsprozessorGeneric programmingSoftwareQuicksortInternetworkingSoftware developerComputer hardwareBinary codeDependent and independent variablesLatent heatFront and back endsPerspective (visual)Entire functionBridging (networking)TwitterInstance (computer science)BitField (computer science)Web pageLine (geometry)Computer animationMeeting/Interview
39:00
FrictionArithmetic meanParallel portCodeElement (mathematics)Term (mathematics)FeedbackSingle-precision floating-point formatVideoconferencingQuicksortNP-hardLine (geometry)Data conversionPredictabilityImage resolutionMachine visionClient (computing)Point (geometry)Moment (mathematics)Connectivity (graph theory)Physical systemPosition operatorSoftware developerProjective planeMathematicsMultiplication signCore dumpOrder (biology)1 (number)MereologyServer (computing)Open sourceParameter (computer programming)Software bugWebsiteGoodness of fitFunctional (mathematics)TelecommunicationHypermediaFiber bundleCollaborationismTwitterArithmetic progressionDifferent (Kate Ryan album)Universe (mathematics)Standard deviationProduct (business)BuildingPolarization (waves)Machine learningInteractive televisionGame theoryArtificial neural networkMeeting/Interview
48:11
Computer animation
Transcript: English(auto-generated)
00:42
Well, hello everyone Saul here from the GT project today. We have a special one for you. It's a GT's 20th anniversary and Today I have with me Emil above the original founder of the project Welcome Emil It's great to have this conversation with you. So it's going to be a great trip down memory lane. I
01:10
Like to share three stories with you three things we've learned and how they have accompanied us along this journey We're going to talk about our relationship with open source a relationship with open standards and our relationship with money
01:25
Amongst across all the all the lifetime of GT and yeah, and and and I'm sorry to jump in I I think the reason why we want to do that is because you know Our We live short lives like we don't have time to learn everything by ourselves
01:43
And we think we have learned some things during these past 20 years as we were working on GT and we'd love to share that, you know, hopefully people would find that useful in their own journeys and They would have an easier time of it that way So first just let's try to rewind time and think in what world
02:07
You'd see was sort of came into existence was 2002 2003 We didn't really have we had no YouTube. No Facebook. No Gmail You know the standard resolution was
02:23
1024 by 768 so no HD sort of fancy things for many people and it was a world where The instant messaging choices were pretty pretty limited. You got like MSN. I am I think I'm not sure exactly Skype was around the corner or already there and
02:43
There was definitely a lot happening in the open source open standards world as well. IRC was their Java at the time later XMPP and this was how the universe was at that time and At that point a younger mill working on his PhD at a strategy. How did that happen?
03:05
Yeah, so You know, I think the primary reason it happened was a Passion and interest and the desire to just go and walk a journey
03:22
That has to probably precede everything III think you You've had that yearning yourself. You were the author of an open source project yourself You just felt like I don't see of all the paths available to me I don't see one that actually fits what I want to do I just want to go, you know trace my own route and see where it takes us
03:40
I think you and I share a lot there What's interesting is, you know, why exactly do what what did our starting position look like? And why did we find ourselves there? I like your description of how the world was 20 years ago one thing that we should also mention is that Computing power was very far away from where it is today and it was specifically
04:05
quite insufficient For the way in order to handle any, you know that the the amount of video Volume that we need in order to have a decent conversation It was you know, just Jenna generic CPUs were not capable of doing that
04:23
They weren't even capable of properly doing Audio, it was the same even in earlier times with the same with the actual internet networks international circuitry That was available for us to transport audio, although by the 2000s the internet was about good enough
04:41
To do decent audio. So so let's look at the situation Where we found ourselves one one of the things that we both shared is that we were very early attracted to open standards And it's interesting, you know, why why would we land there? Why is this? Why would this appear as the obvious starting position for both of us?
05:01
I believe the answer has to do with that limited CPU and it what it meant was that You know a lot of the real-time communications issues had to be solved by Dedicated product specifically audio and video processing had to be done in dedicated circuitry
05:24
you had to build cords that were able to process audio and video which is a very very heavy lift to do and That meant that people who concentrate on doing one of these things they could only do these things People who were doing then video servers
05:41
the internet wasn't powerful enough for the video routing technologies that we have and they Could then therefore only do video service that was that was they had to focus There was so much work to do there and there's a limit to how much work we can take on ourselves But you know, you've got a you've got to make choices at a certain point so you ended up with these dedicated phone builders and dedicated server builders and
06:09
Then finally there wasn't a lot of pre-existing work on the topic that just you couldn't really stand on the shoulder of Giants because these giants hadn't lived yet So you couldn't incorporate a bunch of libraries into what you were building. They didn't exist
06:23
So again, we're in this situation where everyone had to focus on their one thing if this was going to work we had to pick our roles well, and so If that is going to be the case, though then we have to make sure that we all agree on how we're going to talk to each other and
06:45
This is why I believe in real-time communication 20 years ago standards were paramount because that was the only way the thing could work now some companies Might have been able to get away without that, you know companies
07:00
I'm talking like the really big players of the time There was Cisco back then there was a via and a few others and you know, they they had full solutions It's you know, um, it wasn't sufficient because some of their customers, you know, they wanted to facilitate Acquisition of new customers which meant that if they had someone working with partial Avaya stuff
07:23
They could come in and sell them Cisco for the other part of their networking start there and hope to win them over completely With time so even for the big players Standardization was kind of important but for the small players. It was a matter of life and death It was a matter of survival. So for people like us
07:42
Adopting open standards was really the only way to be and if there were no open standards that meant what we couldn't be in There in that space. I think this is one of the interesting this led to one of the interesting connotations of open standards back then open standards became synonymous with this it kind of they kind of acquired this
08:03
moral ring to them and Open standards meant good Proprietary productless started meaning evil kind of which which I think is is interesting I don't believe that that's the situation that we're in today. What do you think about? Definitely not but but I think I think you're right in in the past. So at the time for example
08:25
what we what is kind of the Lingua Franca of VoIP today, which is sip this the second version actually was was being written I believe is from 2002 or 2001. So at that point it was all kind of in the air and Cisco had
08:43
skinny or SCCP Protocol and whatnot and it ended up, you know, they end up abandoning it and going for the open standards But I remember actually one of the bigger proponents of sin indeed But I remember way back when it was reader used the standard thing or there was this other thing But then you could only use these devices and then the others you had to like scrap them or something
09:03
So that was indeed some some cost also associated with it Like for my personal perspective, I think I landed there kind of by accident I just didn't know any better. I think I like what are the rules and then it's like oh this book is the rules All right, then. Let's see where it takes us and it was the only way you could play
09:24
Yeah, you could play the game Yeah, and then the other situation the other the other place where we kind of made ended up the same In the same location was L as well was our relationship to open source now, I The the situation there is a little bit different you could it was possible for you to write proprietary
09:45
Right, but then it's not like the CPUs don't support that like it was the case with video you you could write things yourselves But then it meant going down you know a Much more lonely
10:02
It meant that well if it's going to be proprietary, what am I going to do with it? How am I going to find users to it? I'll probably have to sell it if I'm going to have to sell it then it had it has to uphold a certain You know Satisfy a certain number of constraints which really means I need money and a team which means that I'm probably going to have to Go and build a company There were a lot of steps
10:21
Because the way that you used to make money back then and do proprietary things was you sell copies of the software and in the code That you give out essentially is is the product Open source, I think what we what we both got attracted by you you tell us for for yourself
10:42
But personally what attracted me was the low Barrier of entry you you just go in and you start and then if someone wants to work with you They just come in and start they just send you the code There was like the maximum amount of friction removed from collaboration
11:00
so That was amazing now, I don't know Entirely why But open source to had a very moral ring to it And and it was similar if you're doing open source for many people you're being good if you're not doing open source
11:20
You are sinning somehow, you know, and there were different levels of sin You know people would tell you that it's okay to use an occasional proprietary program That's fine. As long as you're creating something in open source and a certain people will tell you know everything has to be open source, I starting from You know not only the stuff that you write
11:42
But also the the the apps that you use in the operating system that you use so we had this entire you know a panoply of Of adoption of the concept but the concept was there open source good Closed source evil just do more or less
12:03
Does that resonate with you too is that did you experience that he does it does resonate a lot even though I started from a different angle like I think so, of course way back when a lot younger and I Kind of like to be a bit of a contrarian I think And the norm back then like unlike right now where you you would kind of have to justify yourself for doing something
12:26
Proprietary it seems these days. Everything is open source But way back when it wasn't so doing open source was the contrarian part let's say so I like that that attitude that kind of like it was the rebels that worked in open source and the
12:42
You know the more bigger established evil players where we're doing The regular sort software development is that is true and there's something and you know something inherent to youth that you know The attitude comes with youth you're kind of given this world that is already pre-established And it has a bunch of rules in it as you alluded to in some rules
13:03
You can choose to follow because they make sense to you and some rules just appear like this Huge hindrance that you don't want to go through and open source was there was a lot of that for me as well I remember when I was doing my The time when I decided that I wanted to be an open source was a time when I was doing a an internship in a very big company at the time and
13:25
And I was doing that internship and I was looking at what I was doing which was which was an interesting project But it was nowhere near finding itself in the hands of users, and I don't think it ever did It was just a research thing that a bunch of people did and I was thinking okay if I want my work to actually matter
13:42
To be a significant project, what are all the steps that I have to go through in this company You know in this cubicle that I'm in That in order for this to happen and I was realizing it'll probably mean at least 10 years Today, it's faster actually for various reasons back then it meant at least 10 years of you know
14:01
quote-unquote making coffee and Or the equivalent of making coffee just doing boring Tedious tasks before I potentially find myself in a situation where I'm able to impact my destiny that was really unappealing and then at the same time there was this
14:22
Possibility of just going and do it whatever you want to do Now if you were going to do whatever you were going to do You do have to answer one question What are you going to eat in the meantime and we were young back then we didn't have a lot of responsibilities It's we didn't have children to worry about and family tables to put bread on
14:44
so we had less of A problem in that regard, but there was a problem Nonetheless which takes us to our you know final axis that we'll be pursuing here our relationship to money. So For me that relationship started with
15:02
You know, I was just a student I did Started I was actually still in my master's when I started jizzi so my parents were still helping I had a you know, Some sort of a little bit of money coming through in the form of a scholarship And then when I did enter my PhD there was contract it was contract funded. So
15:25
Some amount of money enough to keep me in life was was coming in and I was able to do that although we it became More complicated when we started to grow so as jizzi was born, which was 2003 was quotes if communicator back then
15:46
I Clearly remembered this moment when I Came back from a home from Christmas vacation once and I did a CVS update because back then there was no CVS then we had to go through subversion and only then we arrived at kid
16:00
But I did a CVS update after coming back from vacation and it struck me that I didn't know why I was doing that because it was only me working on the project and What am I going to find that the project wrote itself? Wasn't there and this thought was also, you know only slightly amusing but mostly it was depressing because I thought
16:22
This isn't going to go far if I'm the only one that's working on it. So You know because back then even when there were contributions they were just occasional and still it was me that had to do the work of actually Exactly, so It became very important to me at that point
16:42
That Well, there has to be a team of people working on this if this entire effort Is is to mean anything and the interesting thing is that you know It wasn't exactly clear to me what it was supposed to mean. It wasn't exactly I didn't have You know a very clear exact precise vision of where the product had to land. I just knew
17:04
That I wanted it to be something meaningful when I kind of expected that Hey, we'll figure that out along the way as we go which is kind of different from the start of vision right because the start of vision is you are expected to provide an answer to Exactly how you are going to change the world now. I'm not
17:22
Throwing stones at any of those of these approaches. I think sometimes people do have a vision and it is easy It is good to support such visions, but I don't think that That is the only way we can and we even we do arrive at solutions I'm also a big fan of the idea that incrementally building and just looking around you and
17:43
Seeing what's necessary paying attention to the problems that people have and just ever always adapt And and try to provide better solutions to people's problems. So I subconsciously, I think that's where My mind was adding but so again that meant that we had to
18:03
We had to Form a team of people who were working on this now this started with Yana was the first person to join that team So she too was an engineer at the University of Strasbourg and she too started contributing some of her time to this and then
18:24
We really had to find money in order to allow for more people to be full full time on it so Our initial approach to money was through public funding Public funding meant funding from University there were these local funds in Strasbourg in France that were helping at least startups and
18:44
Went and applied and and we had a couple of grants that helped us that helped us get started So This is it. This is our situation at the beginning of the project. We were fiercely pro open standards We were fiercely pro open source
19:00
and we rely on Public funding as a means to start It's a pretty idealistic path that you ended up in what you think the start was was filled with ideas I don't know that that's even well, and yeah, absolutely, I mean if you look If you look at how the world is today
19:23
You would argue that what you did was a boring Bootstrapped a business, right? You started scratching this agent slowly just no super grand vision Just going for the next step and then the next step and then the next step is well I need someone to do some UI so I'll hire someone well, we need to pay them
19:42
so well, let's find some money there and Sort of like that or organically growing that's that's that's exactly it so now Even though we had a little bit of money to get started it wasn't enough to keep us going
20:03
for forever so We had to start finding real customers now I really love how things turned out for us because what we ended up doing as a business model is something that People often do in the open source world. We were providing engineering services around the GT project
20:24
So you come in you see that it does 90% of what you wanted to do and you want it to be tweaked in a Certain way you want extra configuration options or new features in order for it to To suit your needs and by you in this case I mean usually it's some voice service provider someone selling, you know telephony solutions to people
20:45
Or messaging solutions to people and they want to have their own client, you know Starting with an open source project that's almost there and then tweaking it a certain way is a good way to solve your problem
21:03
It definitely was back then I think Ever since the world has moved to a more efficient way and we're going to talk about that in a second But that was the way that it was supposed to be done Back then today people do that with API essentially chaining APIs on top of each other and and all that sort of stuff so
21:21
That's what we were giving customers what kids customers were giving us and and I think this is where my relationship to money Started changing so There was this ring to money in the open source world in our communities that you know The suspiciousness of money everyone was acknowledging. Hey, we need to put bread on the table But at the same time, you know too much money was always viewed suspiciously as something that would corrupt you that
21:46
That is not a legitimate, you know focusing on every single not any single metric at any point corrupts you because it makes you blind to the rest of the world and you're only looking at one thing and it's very easy to only look at money because It just enables you to do so many other things having money that it's very easy to think
22:04
Well, I'm just going to focus on getting as much money as I can and I'm going to worry about everything else later And then later you realize that oh I've actually put my cells on paths where it's even impossible to to not to get out of however, the positive side of
22:20
working with actual real customer money was that That really changed the accent Of what we were doing. So to give you an example the way that public funding works is that you know, you have Some institution somewhere that allocates a certain amount of money for a certain topic that
22:41
Resonates with them now why that resonates with them is is different, but it usually some sort of political reasons You know, it's these people were usually elected and their voters Expect them to kind of direct channel money into specific topics, you know, and it might be communication
23:00
It might be environment or whatever now what this doesn't really Address especially in our case is that you know It kind of says we want work to be put in that way in that direction It doesn't say We want these specific problems solved and by specific. I mean really specific
23:24
I mean this person gets up every day and tries to do a task and that task takes them 10 hours it should start taking them 10 minutes because other people are waiting and and it's we just need these things to happen faster or All these sort of things that focus
23:42
You know that that difference in focus was was very palpable Between public funding and and and Contras that you get from customers another difference is that it when you when you work with a public entity what you deliver to them It's not even the product what you deliver to them is a report a description That that that has to describe and kind of prove that what you did matches the original intention
24:08
so there was a lot of I don't want to minimize it but it wasn't again Let me put it this way We've had customers who come to us and they ask us to solve a problem and you know send
24:23
Let's say handle voicemail a certain way or something like that And then we start working on it and then we realize that it's not going to work that way and that customers weren't needing exactly That so we're going to change mid-course and go and solve it a different way And maybe we started with DTMF control But it actually had to be with voice commands or with specific buttons and it keeps changing and at the end of the day
24:42
That's what matters. Did we make the situation better for someone who is a customer of our customer in funding? That that is a bad thing when you're doing public funding What is perceived as good as you did what you said you would do and it didn't it doesn't really matter what you learned along the way
25:00
If you just said well, we actually decided to do a completely different thing that that is a very bad thing in public funding Which is completely understandable We don't want you to take everyone's money and run away with it and not do what you said he would do But but it does create this tension. It's like look I thought This was going to be the right thing to do and as I started doing it, I realized that it wasn't
25:22
How am I going to convince you you cannot just keep having committees Reapproving the product right? It's it's an impossible bureaucracy to navigate. So we just end up trying to Present things in the line that they were meant. It's not a it's not a great process to be in I have a little Story in the in the in the topic of public funding as well
25:43
With a segue into open standards actually Yeah, remember was in Germany and we had received some some public money for a project that was around, you know Real-time communication stuff and and we had a little presentation to sort of show indeed like a report we've been working on these technologies blah blah blah, but the night before the iPhone 4 launched and
26:05
There is this famous slide with where Steve Jobs is doing this and here it shows that technologies that FaceTime has it was like SDP h to 64 AC codes black RTP and then We just replaced our presentation with that's like what have you been working on this thing all of them except, you know
26:24
Connected with open standards instead of in FaceTime and that was all they needed to hear, right? Right. All right. Great. Great job moving on. Yeah, and you know it makes sense for them to do that because you know People in these institutions they want to make sure that we you know, we abide by certain moral rules and whatnot
26:45
but again What you lose completely is the place where the rubber hits the road and at the end of the day Does that make anyone's life somewhat easier that connection is lost in public funding I'm not saying this with denigration for public funding. I do think it is completely
27:03
required for many things but over rely on some on that is Can be bad just as over reliance on on VC money can lock people into these burnout startup efforts You know truth is truth is out there the truth is out there in the middle. So there we are
27:21
That was our I think our first transition just realizing that at the end of the day money Wasn't simply a tool that can corrupt which it can but it was also a tool to confirm value, you know I like how the Evolutionary psychologists or our biologists refer to to these things called the havian signals and these are
27:43
essentially the difference imagine the difference between Me telling you that something is important to me. This is important to me or Me telling you that something important is important to me and proven it in a very costly way to me Which of the two has a higher likelihood of being a knownest signal?
28:02
Obviously the one that was costly to me is is it's like I wouldn't engage in it if I didn't really mean it so All sorts of animals do that in there and it's part of sexual selection And I think it also matters in interactions between companies It's you know, where when you're dealing with with public funding, you're always in this fear of saying
28:25
What you what is it what is supposed to be important for you? But it's not always clear that you really mean it and there's a lot of abuse of public funds for that reason Where on the other hand if you're putting your money where your mouth is then that at least You know has the merit of being completely honest if you were willing to put it then then and I know you mean it
28:46
At least to some to some degree. So that was a good lesson, I think Yeah, absolutely Now fast forward in a few years Suddenly there was a pivot in the industry, I guess Where we shifted away from having everything done in the like a lot of done in the client side
29:05
So in this, you know a piece of software we ran in our computers to the browser because WebRTC happened and Around that time as well as it's made another transition
29:21
In how the project was run there was indeed a big shift and it was a tremendous shift and This this takes us to I think one of the the main lessons that we've learned from these three stories And let me let me take you a little bit back GT started as
29:43
A sip cell phone. It was built for audio-video calls it then morphed into a messenger that almost didn't even have it's It's it's it's audio-video functionality But we kind of moved into this combined meta messenger concept that we completely
30:05
Almost completely wrote everything Then our focus to sip audio-video came back So we reintegrated the code that that we had in the beginning It was still mostly sip then we went in, you know focused on
30:22
supporting audio-video for Jabber because we felt that that was the good way to Provide a fully functional Communicator we started thinking more at that point by the way, not Not about playing a role defined by standards, but really solving people's issues. So We started, you know
30:42
Thinking about XMPP. Hey This is a good way to actually provide the proper audio-video communicator to to people Then we thought well Just a communicator isn't enough And people need things like conferencing functionality and they cannot always rely on their server
31:01
that's one of the problems with standardization that you can be talking to all sorts of different servers different components that have different functionality and You have to handle these differences the way we thought we would handle the need for Conferencing and it not being available in every for every person was well We're just going to implement conferencing in the client and we started by implementing audio conferencing
31:22
We started by making that more interactive Then we thought oh well audio conferencing is not enough. Let's make sure that we also add video again So the client became this conferencing server for audio and video Which is one of the reasons why we immediately went to video routing rather than video mixing because you can't mean it Mixing up on the client's computer, but you could route video
31:44
And and that looked that looked pretty good. Then we thought well actually Clients even though they're capable of routing video Are not necessarily or connections internet connections that are capable of writing video and even if they are You cannot have one user carry all the responsibility of the entire conference because they might just want to leave the call at some point
32:04
So we thought let's just take all that functionality and put it in the server And that's how GT video bridge was born It was called from GT some of the lines of code were from the very first GT But now it became this this server product then came the shift that you're talking about web RTC
32:21
And which was such a tremendous change in our industry But what what what it allowed us to do is to completely get rid of the client-side code that was written in Java which had its own challenges and to replace it with JavaScript So we the GT core continued to live in the server But now we had this complete new front-end and now here's a very interesting piece
32:44
When we did that it was the first step of our new building in which we were no longer One player playing a specific role in the entire communication solution We were now the entire communication solution. And this is a change that happened. Thanks to web RTC. So
33:04
Let's talk a little bit about that web RTC. Most people know the story Was essentially a group of technologies that Google acquired over time and then open-sourced and it made available to everyone now I think that the way that This really changed the industry. Was that remember how we talked about?
33:21
Everyone played their standard defined role and that's how we're all able to collaborate this now changed with You know, we're going to start giving you the libraries That would make playing any specific role much easier than it was before So if you now want to be a provider that provides the service the the servers the the clients and the service itself
33:45
it is now way easier to do that because any one of these components now is much easier to build and Part of that was CPUs are faster. So we could just do these things on generic hardware But also you now have these giants that have that have given us access to things like web RTC
34:01
So we can incorporate them and and and really build Callistic Solutions as opposed to being cornered in a single role I think that was really the main the main change of the industry because because having the standards means that you kind of put yourself in this corner where
34:23
You you you not only you're playing a specific role It's not the only problem with standards The second problem which is potentially even bigger with standards is that you're all playing a specific game You know, you are doing a telephone networking solution that kind of does the same things as a telephone network, you know
34:43
it's you you Sit wasn't built for notaries to sign doc to notarize documents for you over webpage It wasn't built for people to play Dungeons & Dragons together it wasn't you could use it for these things, but that's not why it was invented and with this new approach
35:02
where look we're going to let you define the game that you want to play and we're going to give you all the building blocks So that you could build it so that you can define whatever game you need This is what web RTC did for telecommunications and it's actually what the web did for everything, right? It's you can you can have yeah find similar trends in absolutely every every business and in every industry
35:24
I think one of the more important things from that from the kind of technical perspective was that is Really really hard to have a good media engine build from scratch. Yeah And whether it is he gave it to us it I like state-of-the-art deco cancelling and we didn't even have to work for it
35:42
I just got from the browser and with a new update it just gets better and Building on top of it. It really for instance that it did level the playing field so you could Everybody was building on top of this and there was no oh That's it client is proprietary but it has access to that proprietary echo-canceller, that's so good. Well here is everybody's playing by the same
36:05
by those same rules and It's a great building block So I remember when we we just stopped worrying about that that media part is like, ah, so now it all just works So now we can all focus on actually solving user problems
36:22
And not simply technological problems. So So this was this was quite Quite the experience for us as well. And then around that time. We also got acquired as GT By at last and it was in 2015, which I think coincides
36:42
with really the time When companies moved to the cloud Amazon started existing around 2008 But but really the cloud the massive migration to cloud was somewhere in the mid 2015 Where every company just figured out no, that's that's how we got to be now This is this leads us to our learnings about open source because I think that massively changed the game about open source
37:07
with the cloud Game came a certain amount of flexibility now, right you you I don't need you to pay me money if I'm writing if I'm writing software I don't need you to pay me money that would
37:22
Kind of pay for the entire development of the product that I'm giving you right because remember we're talking about binaries I'm giving you the thing and I want you to give me the money That pays for this entire thing or at least the compound earnings pay for this entire thing That's no longer necessary because I don't need to install a bunch of servers now I can go in on-demand get things from the cloud when I need them, which means that I'm more flexible
37:45
Which means that I can give you more flexibility Which is why a lot of the you know move started happening to two subscriptions So, you know you want to be a customer this month and not the next one. No problem. We can do that
38:01
now this move and this incorporated flexibility now will now change the Relationships to the actual code that the companies had to the actual code remember how companies like Microsoft had this very hostile relationship to open source and Google back in the day were you know very unique
38:22
For fostering open source probably because they kind of were already living the first days of the cloud When you are in the cloud, the money is no longer being given for the code It is now being given for the service which means that the obsession of who gets their hands on the product and and the code became
38:42
It disappeared And I've seen this a lot with with companies the more a company has been Dealing with selling licenses in the past the more it has it Had a hard time of embracing open source any companies that is cloud-native
39:00
Realizes completely that Keeping your code private is just an expense that takes effort and it doesn't really give you anything It that's no longer a thing that's worth protecting That much in fact It's much better for you to just not spend the cost on keeping it secure and put it out there and potentially get a feedback
39:24
For it's not guaranteed to happen. Sometimes we oversell the open source story It says like we just put code out there and people come in and contribute that doesn't happen in order for people to do That it has to solve a problem for them And no effort needs to be spent in Fostering the contributions and taking them over to finish line
39:43
But it I would say this is this is the lesser thing The first thing that has to be true is that if you want people to come and help you The whole thing has to be helping them Yeah, they need to to find your thing useful, which means I think about what we've done with Jitsi
40:01
we've always Made sure that Jitsi in itself is a solution to a problem not just a piece of another bigger thing Which is what many companies do now, you know some companies would only open source their API to a proprietary system the Incentive for someone to come and collaborate with that is much lower because they don't have the system that this fronts
40:21
You know, they might be a developer that uses it They might have kept throwing some sort of a trivial bug fix, but they're not going to come in and add functionality Around that if you want someone to do that They have to feel that they're adding functionality to their own thing as well, right? So that's I think a main difference like the main requirement for open source project Just to frame successful product. It has to solve people problems if you want people to put their time and effort in it
40:48
so that Change with with with open source was was you know Pretty pretty key
41:01
When I think about this, I think that another key element of this entire thing is that? of the way that open source changed was I Think people sometimes don't give enough credit other than the hard core open source idealists I think people in the proprietary world and the today's business would don't give enough credit to people like Richard's
41:21
Talman for what for what they did and how much they helped including business now one of the key contributions that Richard Stalman did I think was that when he insisted that You know every project ships out with a specific license, you know, most people focused on the fact that you know
41:40
He was promoting the new licenses but but but what they Didn't really realize is that with that came the notion that everything has to have a license Even if it's not the new one Everything has to have a license and I think we I Really believe that people like Richard Stalman and the ones that that
42:03
Established that notion even though he meant it for a specific one again. What ended up happening is that they everything Licenses are kind of expected and found in most projects They're critical in a couple of years ago There was this thing with the react license that there was a huge turn off for the community and they ended up having to change
42:26
so For many people like a good license. It's critical that and the reason it's critical is because Remember one of the main points of open source is that it facilitates collaboration But in order for any collaboration to happen it turns out and I think that was proven here was that hey
42:46
We need to know upfront Under what terms we're collaborating. I used to participate in in European Commission projects and and they were universities and private companies big and small participating in and
43:02
The way that these things got decided there You know a consortium would be formed and people would start negotiating But what is it exactly that they'll do and under what terms would it be available to which partners and those? Discussions were taking months before a single line of code was written and at that point
43:21
That was when I had my moment. I realized that The fact that you don't have to deal with that in open source that going in you already know What the terms are going to be? Is is so precious it is It is absolutely necessary it actually having some terms Matters way more than having any terms because if you don't have any terms now now we have you've opened room for
43:44
argument and then people start On the details and it actually turns out that doesn't really matter because you choose as long as your licenses are within Well-known licenses, yeah, it's People actually don't really get hung up on the details. So I think that was a major realization about open source as well
44:05
So I think that pretty much is our journey and if I have to Go into that for the for the predictions, you know, so people usually expect from these from these talks is what's in the next 20 years?
44:21
And I Don't think I'm good at those and I Think there's obviously plenty of exciting stuff that will be happening What I could say is that I believe the trend where? We Work together by stepping on each other's
44:41
Shoulders rather than on forming these consortiums whether it's through standardization or through European projects or something like that I think building on layers is going to be an approach that that matters That that works a lot better That's why we when Jitsi for example decided that Jitsi the open source project is not only going to be a client
45:03
Solution for having video conversations But it's also first going to be an embeddable component that can be part of other solutions And and that's what our business is about. But it's also what the open source project is about now It's an embeddable component that others can Can step on so I think that this enables tremendous innovation
45:24
and you you like this this way of putting it the other day that We're now in a position where Jitsi can be the perfect solution for your second most important feature like you can focus You know You can just the way that we take the media engine from WebRTC and then we focus on exactly the details of the video
45:41
Conferencing someone can take our video conferencing component and you know, don't worry about these interactions and then focus on you know not arising documents or playing Dungeons and Dragons and how exactly the Game I think What this is enabled is an incredibly fast pace of innovation we're all aware of
46:05
The new technologies that are coming out with regard to machine learning and artificial intelligence There's tremendous Development there and I don't intend to you know Make any predictions about that, but what I will say is that we are very well prepared
46:24
You know to quickly build solutions that that embed such new technologies and I'm very excited to see what it will be Absolutely, I'd say that like without actually planning for it as you said there was no grand vision the one thing and this is actually how your path and my path kind of went parallel is we worked on things to communicate with others and
46:48
That I don't know why but that fills me with joy sort of and I like a lot These paths in the end the only constant doesn't matter you build a desktop client then you build a server then you do a website for it because that's where
47:04
Where the path takes you but in the end it's about making a tool to communicate people So we don't know what's going to happen five ten twenty years from now But it's probably going to revolve around communicating people at least of that. I'm pretty sure That I I completely agree with you. I think that communication is is key to our progress as a society
47:25
I think I'm actually very happy that more and more communication is happening Even though some people perceive that as polarization what I actually believe that No, it's just that more people are being able to voice their opinions and and ultimately that's good Even if sometime the friction might appear too hot
47:42
So I completely agree with you there knowing that what we're doing is helping people talk to each other is Also something that gives me you know, a deep sense of meaning and so I guess we we do indeed run parallel on that one Yeah, and let's hope that these three stories these three lessons these three things that we learned across this journey
48:02
Well help others in their own journeys That's the that's the best we can hope for until the next one Bye everyone
Recommendations
Series of 20 media