Huddle01 🎤 Pioneering RTC - Susmit Lavania

foreign [Applause] [Music] so I'm here uh quite excited uh to give a talk uh so I would be talking about a technology right or something that that we use in our everyday life uh but something that we don't talk about much and it is something that we require as a human to build connection with other humans out there uh so that's called real-time communication technology right it's a part of our daily life uh so during our prehistic time what used to happen was that it was the communication that helped us coordinate to each other and when we were able to coordinate we were able to increase our survival chances and we were also able to bring more food to our table and distribute it much more better uh this is also the core reason one of the major factor of the evolution of ours we also know that the money and the crypto that we know is basically a tool a tool to capture the value and this value comes from the consensus which is basically people coming over an agreement a group of people coming over and everyone this makes a real-time communication technology much more critical than ever right now if we see in our web to World there are a lot of applications like Facebook Instagram Tick Tock Twitter all of them combined the social media combined there are 4.48 billion active users social media user and if you compare them to web3 there are only 84 million active wallet users and these are basically combined with centralized exchanges also addresses also under decentralized exchanges also so you can see the ratio like we are still a lot of problems the social app that has to be solved the web 3 adoption is still an unsolved problem now now to get this real-time communication basically how do we bring it over the internet right so real-time communication is basically many to many audio video data transferred synchronized at sub latency second but like the current protocol that we have which is like HTTP TCP UDP uh they are not optimized to do so because many of them operates on request response mechanism and all of those things so there was a protocol there is a protocol called webrtc that is actually used to give this RTC functionality in the today's world but there are a lot of problems with the webrtc because it was not designed to scale out there and actually this led to a lot of problems out there uh if you see this led to like the custom servers to scale out there and since it requires custom servers so what happened was that the corporate started accumulating those servers out there and even on the P2P level you required this centralized signaling mechanism to actually establish those called out there which is the problem now this what's happened with this is that basically the user's data goes to a server which is practically a black box so you don't know what's happening with your data you don't know what's happening with your privacy uh out there and also you'd never know like if it's the best performance that they are getting uh out of the box and on top of it the users are actually exposed to rent extractor mechanisms and even the vendor lock-in systems out there and there are a lot of other problems but if we go into that I think this we can go so now we have drtc like uh uh this is something that we were actually working on at huddle uh so huddle is not just a normal audio video conference it's incubating this RTC infrastructure of the future it's called Uh decentralized real-time communication infrastructure uh so what is it uh I will give you a visualization so basically like how polygon and other are like basically the scalability solutions for layer ones the drtc is this real-time communication layer for l1s out there and what this would enable is that now you can see this web3 social kind of uh dabs dabs which required this uh live streaming which required this audio video calls or which require their avatars to chat with each other at real time those can be possible uh with these things now at huddle what we did was that basically uh to get this real-time infrastructure up and started we took this product first approach and with this product first approach you can actually boot boot start this RTC infrastructure layer out there so yeah how did we basically and another thing that uh point to note is that uh what you see is that uh the token uh like what happens in generally the web world is that people make token and they dump the token on the community itself uh the other thing is that like they make the technology and then they dump the technology on the users itself it's the same thing right but what we need to do is that we need to make a product and we need to make that product valuable and then once that is valuable we need to distribute that value to the users and the community now once we are able to distribute wealth uh this higher development among that community members out there so yeah at title what we are able to do is that within a single click you can see you can access all of the best of web3 uh like you can live stream you can record you can have your ens name you have can have a Unstoppable name and lens protocol out there you can even live stream to YouTube twitch everything out there all at a single click so we are also be able to bring this experience to your web application and to the mobile applications also and these are live you can just check it out from your app store and Play Store and start playing around with it so all of this infrastructure is actually powered by a media infrastructure uh this media infrastructure if you see talks directly to the filecoin network to store on ipfs and filecoin uh it it has its live streams to live here protocol uh to live stream which is basically one to many uh broadcasting scene and it has this media routers uh which is made on open source libraries but on top of this media servers there are custom algorithms that we have written such as last 10 uh such as dominant speaker identification uh codec optimization and all of those things so this makes uh the infrastructure much more robust and this is a sufficiently decentralized architecture as of now foreign now if you see like we have spent like a year or two year uh making this architecture much more robust and all of those things but what we don't want is that the developers out there to go into that same Rabbit Hole so what we did is that we packaged all of this infrastructure into our sdks Now using our sdks you can have the same capacity that the huddles app has so is basically you can Fork the Huddle application and make it on your own you can just white label it and you can make other kind of application uh you can make Twitter spaces you can make forecaster spaces you can make lens uh lens calling one to one all of them you can also make like a wallet address which is calling other wallet addresses out there and it opens like a huge door out there if you're building a metaverse in the metaverse like the avatars need to chat with the other avatars so here is where you can use the huddle sdks out there and there are other things also like the imagination can go well wherever you require this real-time communication infrastructure you can just plug in the hurdle sdks and get that up and rolling uh these are load aware reason aware globally distributed server and I would also like to take you the quality uh that we have actually what the Huddle this in infrastructure is capable of first is quality so we can see like it does 1080P and higher quality out there so it's much better for the screen share we have been able to minimize the packet loss days let's digital buffer uh this Dynamic bandwidth uh adaptation and all of this thing we are trying to do in less than 100 millisecond because uh that's the best VC experience that we get when we are doing this 100 milliseconds uh so yes we are end-to-end encrypted also uh so if you can see uh there is a call going on so this is a version that a user sees on the left side and uh the other version is what's if the hacker breaks down into the media infrastructure that is what he sees right it's an end-to-end encryption call and even this is how Huddle's media infrastructure actually see so your data is actually secured uh out there even like the Huddle servers cannot see uh other routers they cannot see the data that you are doing they can only process the meeting at a super high quality out there scalability so these end-to-end calls we were able to scale to up to 30 people out there and here you can see it live out there were further activating the last an algorithm this helped us scale to more than 400 people out there and this is basically the robust infrastructure that we have got and further going on we have also can be horizontal scale so which means such meetings can happen parallely out there so can handle any amount of skill so the progressive roadmap uh so what we believe is that uh the decentralization is not a switch which can be like zero and one it's Progressive in nature uh so huddle was born in eighth Global hack FS uh 2019 which was sponsored by the file coin and protocol lab we went into techon accelerator and we have which and this is the current product that we have right now uh so the drtc we talked about is the basically the phase two part of plans and there is an Afra crypto economics uh which is our phase three part-up plan and they both would be working parallely uh so drtc is basically uh this will make the all the central component into this nodes architecture and the afro crypto economics would be basically this incentive mechanism for all the apps to work together what we will realize end of the day where we actually want to get to is basically an open neutral bondless decentralized sensitive privacy and government Community owned and government real-time communication infrastructure yes uh so would love to share some of the learnings that we had while building huddle uh so in in hacker first we actually made it as a P2P machine but we realized that there's a lot of problems with this approach uh so the P2P mess was not scalable for after four people like the experience on the user end was not something that can go as a product itself there's a lot of less security because if in in P2P mess if any one of the peer is malicious the whole thing gets compromised uh the data procession you cannot purchase the data in this typical fashion out there uh this is the whole reason like the ipfs actually made the file coin which is basically the same ipfs mechanism with an incentive layer out there uh and the other problem is that in this p2b mesh the other peers can send the malicious or illegal files to other peers because if they are contributing to the network so again that's a problem and we cannot always rely on p2bms because they might be available or they might not be available so they are highly unreliable what we can do is basically convert this p2b mesh to peer to node to node architecture once we are able to do so what we can do is that these nodes can become uh have this macro level feedback uh and wherein these macro properties can be decentralized uh and they have this uh consensus algorithm taking care of all the micro level uh competition that is going on so this is like uh individual node which are optimizing their self game but they're competing with each other but on the overall macro level they're collaborating with each other uh so this makes it very highly scalable out there and you can always have this uh security and slicing algorithm this maintains the quality of service the network has and even the security uh so I have for crypto economics uh so this is something uh a network flywheel uh many people ask my question that how the value approval would happen on the token itself so if you can see there are end users like and there's a Community Network effect starts to happen once that's happened your volume of artisan starts to rise and once that starts to write what happened is that people start building application on top of the hotel infrastructure or or the RTC infrastructure out there this makes your volume of RTC transaction rise and once that's rise you're basically the token value starts to raise and investor generally sees wants an upside exposure to any value that is going up so they put in the financial Capital once there is like a financial capital and the token has this both monetary value uh the people start to put forward their computational and bandwidth Capital the Huddle nodes the nodes provider and in return they get the token itself so you can see on a global level like the people are putting in more notes so what happens is that the quality increases the security increases and the pricing goes down and the scalability increases with the more the nodes that get acted out there the app for crypto economics uh so in the recent like what used to happen was that the incentive layer was only on the infrastructure layer but on the application layer the only incentives you have is basically bounties uh or some other forms of making people to make the apps out there I think we can have a protocol that can work that has this both incentive layer on the application layer and this has this incentive layer on the infrastructure layer and there is a dynamic between two uh that collaborately makes the whole ecosystem grow yes uh so would also love to take you to the medical fellows and social network Theory because real-time communication has this uh value in uh social network out there so metcalfilos states that that in a telecommunication Network the value of that system Grows by n square if another node comes into the picture out there so here you can see in the graph basically like first it's like one or two user but as as the node starts to increase the value of the whole network start to increase and the whole of the value can be captured into this crypto economics uh out there uh so the perfect example of explaining the Metcalfe law is that suppose you have a mobile right uh so in this mobile if other person has mobile the mobile this mobile gains value right and uh so if you're the more the number of people you are able to call within this with this mobile the more value that the mobile the whole network has so for end users the value increases if number of people and the nodes into that infrastructure start to increase and this is like a what what the Metcalfe and The Social Network Theory actually summarize the and this is how actually the networks like Network effects product like uh Facebook Instagram and all of those things are actually valuated on the network effects again uh so zoomware was able to overtake Skype because what they enabled was that uh so Zoom only you had to have this email ID to get to a Skype call but what Zoom did was that even without a Gmail ID you can do a Skype you can do a zoom call so this led to more Network effects out there uh but at huddle what we are doing is that you can have this like non-zoom account and you can also bring in this crypto Primitives out there which is on the rise and this is again leads to this network effects out there yes uh so again uh I would highlight some of the problems in the webrtcs out there uh for a webrtc to take there's like a fourth thing that goes on one is uh the first step is signaling uh other step is uh like you need to uh connect the secure out there first a signaling then is connecting then it's securing that uh whole infrastructure and then you start communicating out there uh but there for signaling you need to have this servers out there which are centralized in nature uh so again uh so in this if you can see like the signaling is centralized uh but if you can convert this servers into a nodes and these can be like multiple nodes which are like chatting with each other so your sdps can you can now send to this node and these nodes can gossip around this sdp and whichever peers required that STP it can subscribe to that stps out there so this is how you can like remove the single point of failure out there now we have realized that like you can you can use this not just for like putting out the sdps you can use it for both synchronous communication and you can use it for asynchronous communication right uh so what is synchronous communication uh so if you had to use slack right they just slack button out there so it is like many people coming at call at that point of time so Zoom meeting is a synchronous out there but Discord and slack their asynchronous communication so with this mechanism of Divergent right wherein your diversion node wherein you can use this infrastructure for signaling to for the to attain the async communication and simultaneous industry you can use an infrastructure to have this this async communication capacities out there so again what we did was that uh this is the end-to-end encryption that we had solved uh out there it used NaCl algorithms out there from the private public keys of the user itself so even this signaling mechanism is end-to-end encrypted and yes so we have packaged this Divergent node into the Divergent sdks now you can see that this is a WhatsApp clone made on a lens protocol and you can see a starter hurdle that uses huddles media infrastructure huddles sdks and all the chat SDK is like the Divergent sdks out there so you can see we have achieved both asynchronous capacity of communication and the synchronous capacity of communication and developers can have this imagination and can build this any kind of applications out there and the dids that is is basically a lens protocol out there yes uh uh so like now coming back to drtc like what is actually a drtc infrastructure so if you see like UPI right the UPI is a fintech infrastructure wherein this vmap phone pay app and Google pay app was made uh similarly this drtc layer is like enables this web is the spine of this web3 social lab upcoming web for social app so now you can have this decentralized Tick Tock you can have this decentralized Twitter uh and you can have uh another kind of New Primitives app which I think would be possible which is like the social app and the financial merger kind of apps out there because think of like a social a Twitter leveraging the defi infrastructure out there and this could this can blow up like uh we just need to like productize that very well out there but to start for the start I think the cloning of the existing Vapor infrastructure would start to happen like a clone of Discord a clone of WhatsApp clone of Twitter uh clone of LinkedIn they would start to happen and they would require this sync known as capacity and asynchronous capacity also and even in the meta was like in the metaverse like all the nfts they have these games right the MMORPG system the massive multiplayer online system all of this requires this real-time infrastructure yep uh so the huddles infrastructure is already powering a lot of uh other applications out there uh so this is the whole ecosystem that huddle is actually powering so there's like schedulers there's accelerators there's like metaverse spaces there's messagings there's web 3 attack there's wallets nft communities social and I think this is a blank canvas which is like yet to be filled by other entrepreneurs and other people making new Innovations out there yep thank you uh this is and this is my barcode if someone wants to follow yep