🚀 What is Web3?

Hello everyone, so we're live. I'm happy to be with you today. Today we'll be discussing Web3 in the context of the web revolution and how Cloudflare might help support it.

This is definitely like a very broad topic and like discussions online are usually like very involved, sometimes polarized on this topic.

And today it was In-Young who wanted to provide a nuanced version for what we think is Web3, which is a term that like has crystallized some of the current evolution for the web, which we think is a more decentralized and more private web.

So yeah, as mentioned, joining with me today is In-Young.

Hi everyone. Yep. Good to be here. Thanks Thibault.

Excited to talk Web3. Great. So I thought maybe In-Young you could walk us briefly before we start diving into this future trend for the web.

Maybe you could like walk us through the history of the web as we know today.

Yeah, sure. So just as a quick caveat, I'm definitely not an expert, but I have been doing research in this area over the past few months to help with some of the announcements that have come out today around our distributed web gateway products.

But yeah, I can definitely provide a high level perspective on this.

And Web3 is certainly a very interesting topic that's constantly evolving.

So if you kind of look at the evolution of the overall web in its 30 or so year history, it's really interesting.

And I'm sure you and everyone watching is very aware of how quickly things evolve.

And that's no different during birthday week at Cloudflare.

And so I think before diving in, though, it's probably important to kind of distinguish between what the Internet is and what the web is, because these are terms that are thrown around often interchangeably, but they do have some slight nuances.

So when you think about the Internet, it's really about the technology that helps computers talk to each other.

So whether that's kind of sharing data or computational resources. And it's this concept that's been around since the 60s with the development of things like ARPANET, which was a very early computer network.

On the other hand, you might think of the web as more of a network of information systems that's kind of connected together through hypertext links.

And this was a concept that was developed by Sir Tim Berners-Lee in the late 1980s.

And it was really around trying to solve this problem of existing information systems organizing things in a very rigid hierarchical manner.

And one of the ways that this initial version of the web or what people call Web 1.0 was around decentralization at its core.

And this was really necessary in order for links between different information systems to evolve more kind of organically as organizations kind of grew in size or merged together, things like that.

So an example of probably a Web 1.0 solution was Yahoo.

And in its kind of inception, it was just like this long directory of links to other websites.

And this kind of captured the essence of Web 1.0, which is kind of static and read-only.

Now, yeah. Maybe one thing to crystallize is Web 1.0 wasn't defined as Web 1.0 until we saw further trend.

And so it's kind of like the early web got categorized as one version.

It's definitely something to describe a trend and not something that got built as a web that we upgrade step after step.

That's a great point. These things happen very organically. There was no one who said we're going to release Web 2.0 at this date and Web 3.0 at this date.

It's the phases of the web are really characterizing kind of broader trends in how the web has evolved over time.

So the way that I think about it right now is like Web 1.0 was the static web.

Web 2.0 kind of marks the period when there was more adoption happening and people were really trying to derive more value from the web.

And one of the ways that kind of materialized was enabling interactions. So like some examples are search engines that kind of like move from just static links on a page to enabling people to search more intelligently for things that they're trying to find.

Another example is social networks that let users like all over the world interact with each other.

And yeah I mean another kind of like key characteristic that I think defines this transition from Web 1.0 to 2.0 is also like the data management that needed to happen in the back end.

So you know in Web 1.0 you had content that was stored on the website files themselves whereas like in Web 2.0 just because of the sheer volume and the amount of interactions that were happening there was just a lot of data that needed to be processed.

And so those those two like the website front end and back end kind of got separated and you had this whole industry that was focused around developing technologies to support kind of like the data the networking and data storage that needed to take place to support this new paradigm for the web.

There's been like a lot of content in a short amount of time and so like now that like so you can you characterized Web 1.0 as static and like Web 2.0 as interactive like what is a sign that like something is about to change and like what is a sign that like a new web may emerge or like something new that we haven't seen before?

Yeah so I think you know Web 2.0 solved a lot of problems that were present in kind of like the initial versions of the web you know making it more interactive helping connect people more organically but I think some unintended consequences from Web 2.0 were kind of around centralization.

You saw these kind of trends emerge where data was controlled by kind of central organizations so you know there would be websites that collect user data and sell it to monetize their websites.

And there's also an issue around like centralized digital identities so you have all of these disparate platforms on the Internet that each have their own kind of login portals and varying levels of security for those portals and so this creates a lot of different points where potential data breaches can happen, private information can get leaked and no one on the Internet really wants that to happen.

So there's this kind of shift now that we're seeing towards bringing decentralization back into the forefront of the web and this is kind of like what's characterizing Web 3.0 right and so it's solving a lot of these issues with Web 2.0 and one of the key technologies that's kind of underlying it is blockchain which is a term that's being thrown around a lot these days but and you know it's really technically fascinating and complex but you know just to quickly define the term it's just a distributed database it's kept up to date through kind of like collaborative collaboration and synchronization between nodes in a network as opposed to kind of all of that being controlled by a centralized entity.

So that's what we're seeing there's a lot of really interesting use cases as well for the technology as well.

Yeah I mean like it makes sense maybe more than like blockchain technology like as you mentioned like being like quite of a buzzword and like sometime having like a package into more like what it could mean there's also like this thing of like having like more trust like in like having a different approach to the trust assumption we had rather than having like the trust you have in a certificate for instance you have on your machine etc trying to have something in a public setting and like that's what for blockchain might have provide it also went like on the other systems such as I don't know certificate transparency logs or etc that went on to challenging these trust assumptions about as you mentioned having one central point for like going to to make sure like the data from the given user is actually like this user so for the login for instance.

So yeah definitely something I would think to transform it in that way.

Yeah no I think that's a good point for clarification right like a lot of times web3 is basically synonymous for a blockchain is going to take over the world and that's not necessarily the case I think the way that I think about it at least is like blockchain is one technology that's kind of helping like drive this shift towards decentralization better security better privacy on the web and so there are a lot of interesting developments that are happening outside of even kind of like the realm of blockchain that are kind of helping support this movement.

I mean and pretty definitely as like we like as Cloudflare like we were in CDN and definitely like try to make things greener because there's like less data exchange etc blockchain is definitely one option and like even in the blockchain space are like multiple heavy debate on like should you use like proof of work or like you would like heavily use your computer hardware or do you want to move to proof of stake do you want to move to like something else and like once again like that's different trust assumptions you have do you want to trust like people that have like money to like power blockchain do you want people to stake money or like all those things which is definitely something in nuance that that yeah like that is very different than what we've seen.

Yeah no that's that's interesting as well like yeah blockchain is also not going to be this cure all for all the problems of web 2.0 and there are definitely all these other considerations around scaling and environmental impact that that need to be you know considered but I think you know we're there have been a lot of advancements in the space you know if you want to get technical about it there are like layer two scaling technologies that are coming coming into motion and there are ways that people are thinking about handling this and we're still I feel like very early days in kind of like the development of web 3.0 with a lot of experimentation taking place ideas being proved and disproved so it's just a really exciting time to be involved.

Yeah no I feel sort of great.

So yeah I'd actually be curious kind of like from your perspective kind of as a research engineering and more technically savvy than I am like what do you see as things that are changing today and kind of helping bolster like just this web-free paradigm?

Well that's a good question definitely not to be like taken like word for word but like one of the things that I do think like in the like current highlight we made like today on like Cloudflare blogs is there's definitely like a new like web primitives and like cryptographic primitives that like is also like shaping these decentralizations so like for instance on like now like modern web browsers like it isn't ever to have like cryptographic keys which like can be used for instance by web or send can be used like by the user like to prove that like at some point they were able to identify again some certain services and I mean this is like really interesting because when you think about it like for quite some time you had like this like server model where the server had to provide like a certificate and like this certificate was trusted by like certificate authority which like verified that yes the server had the right to to provide certificates and all this was like kind of low-level primitive that were embedded in browsers or software and now these seem to be like available like on the user space so that's what I mentioned before like powers web will send and that's also what has been like leveraged by blockchain to a certain extent to provide some kind of public trust so of course one like I mean like you and me we like can definitely meet like in a cafe even though we're like quite sometimes on the paths but if we were to meet in a cafe I could go and like verify the public key and say yes I mean then next time when like I will discuss with you I will like I would be able to make sure that like with the same entity and that's like what for instance messaging application have been using we have like this QR code where you can like trust the person in front of you like you make like kind of a verifying party where you agree that this person from now is not safe to do this publicly however it doesn't suit every application and that's where like one of these for blockchain has been like in terms of like decentralized identity was to have like a setting which at some point you claim that like you're like a certain person and of course the more people like can verify that like these certain persons you the more like you have confidence in like the public key or your controlling is actually like you it's like one source of trust definitely there are like other ways you can publish like publicize um and like as mentioned before like blockchain may not be sufficient for everything but definitely that's an interesting setting to provide some kind of transparency um to that yeah that's that's really interesting and you know one of the questions I have is like like there it's clearly it's clearly a case that something like this um like kind of a trustless transparent um way of verifying identities online or even in person like that could be pretty transformative like what's why isn't this kind of like all over the place right now um and like what or maybe like taking a step back are there any kind of like technical challenges to web3 that we still need to overcome in a way for for adoption to to really take place I mean um rolling like a pki system and at least like I mean I'm definitely not an expert on that I definitely like much more competent people but like um as far as I know as I've seen uh rolling like public pki system uh like a public key infrastructure is technical and challenging um and it's even more technical than challenging than like for some time you didn't even have those um like low level cryptographic primitive easily accessible um in the user space on like the web browser which is like the main um like one of the main way for users like to access content and so that got really challenging you still have like this trust assumption like how do you make certain certificate reach like the user how do you make sure like this is trusted and so this is something like that has been like very challenging um blockchain in a way saw this way as in we don't trust any participants in like the network uh but like we will trust the output that they have and so the output that they have like for instance on the like on bitcoin based on the paper with sachi nakamoto was if you can prove that like you've committed a certain amount of work then yes I trust you to have like a certain amount of bitcoin um then in like staking system if you can commit that you prove and like that's a cool thing with cryptography if you're proving that you are able to lock a certain amount of money for a certain amount of time and you're committed to do that and know like you cryptographically if you try to cheat well you will be seen um so this is also something we do reward on like the good and maintenance of the system and so that's something changing a bit in some of like you have more technical opportunities and like you had before to roll such a system um and at the same time there like has been different assumption about how we like these nodes and my computers need to communicate together to like make a model that was sustainable um so of course like um certificate authorities are definitely like great work on that um in terms of like making a web more secure than like it used to be there might be other options we're not sure if like this is a good way to move forward to but these provide different like technical technical setup um that could be leveraged i'm not sure if it's not like too abstract or not uh but that's more something down yeah i mean there there's definitely a lot of terminology um like that could be unpacked like what is exactly is proof of work and what is proof of stake like what does that how does that actually work and you know like what are consensus mechanisms in general um i mean that this is a big question and definitely people want to dig into them um i mean there's like a fantastic thing on the Internet uh which is like either like a directory like yahoo where you have a taxonomy and you can like go to the consensus category and like likely we have a lot of stuff uh but you can also go to like like services like google search engines where you would be able to find a lot of things like on the topic um which is like a magic of the Internet like maybe it's something you can do today like the amount of knowledge available is kind of crazy yeah awesome well um you know we've we've spent a lot of time going over kind of like the trust identity kind of like use case um for blockchain there's the obvious kind of like um transformation around like financial services that's taking place with things like bitcoin ethereum and all these cryptocurrencies that we um hear about day to day um are there any like other use cases of um maybe not just blockchain but just like web3 um technologies overall that are particularly interesting to you right now yeah um i mean there will be multiple ones i do say if we like we start with like what we invested it um at cloudflow is like um ipfs is something interesting in a way that um it provide like kind of a global um indexing layer um and very fireable layer for content um the great thing about that is um like ipfs doesn't like magically make your file like available forever or um that doesn't make um storage permanent but like one of the thing ipfs is great at is like if you put a file on ipfs um you would have a hash which is based on the content so it's kind of like an identifier of the content um and like if i give you an identifier of a file i do provide an ipfs um you will be able first to like retrieve the file over like the distributed network but on top of that you would be able to verify that the file you receive is actually the one uh you wanted um and that's great because that provide the common interface and somehow i won't say standard because it's not standard but definitely like a protocol for everyone that everyone would understand um to maybe store multiple small chunks um of files um across multiple multiple nodes and participants in the network while still having this ability and like property to like property to verify that the content you receive is the one you expected um so in a way it's like very simple of having this indexing layer where you have the content and then it's just like file retrieval uh i mean when i say just it's like still a complex problem but like it it's like file retrieval of your distributed network um so that is very simple but like very powerful um because you would not have to trust who is providing the content to you um you just can verify as you receive the content it's actually the one you expected which in a way is very similar to like what torrent uh like wanted to provide except that you do not have like the like centralized tracker in order to retrieve the list of blocks you have to to get right yeah i found ipfs or you know the interplanetary file system a really interesting use case as well because it it kind of draws from like um these these concepts of peer-to -peer file sharing that have existed for a long time now like you mentioned bit torrent has been around for ages um but and and also kind of like maybe almost like dropbox in a way where but but essentially you're just enabling file sharing um for people in in this trustless way um that um you know it overcomes a lot of like the challenges with some of the initial peer-to-peer networks with um like creating incentives for people to actually run nodes on the network and use their like computers to not only store files but help deliver them to the end users um so it's it's just a really interesting confluence of like all these concepts um that are kind of coming together like maybe one thing is like there's no i mean there's not much apart from like social incentive for like people to store content on ipfs so like if you do not want to provide some content you're like perfectly free to not provide it um that's like what some projects such as like should be file coin which is like by the way using another very cool technology which is um zk-snark but we may not dive into it uh but like like filecoin or cyno etc are also leveraging to make sure this storage is permanent that's definitely like some say far more cutting edge than what ipfs is doing yeah awesome well i know we're we have about six minutes left i would love to also hear more about kind of like what Cloudflare is doing because um we've obviously had some blog posts come out um this morning um about kind of like our distributed gateway products um but yeah we would love to hear your take on kind of like what we've done and uh what we're kind of uh envisioning doing going forward yes um so once again a big question and definitely you've been also like very like very modest in like asking this question because you've been the key part uh to making this happen uh but uh maybe in a short uh sentence what Cloudflare is doing so like the first like main major thing we've announced today is uh we're um improving uh like our ethereum and ipfs gateway uh making it uh like fully flat product uh for Cloudflare and taking it out of the research team where they've been like slowly growing over the past uh the last couple of years um and i think that's amazing because like that provides a bridge between um the the current distributed networks such as ipfs and ethereum and the web as we know it today um so that kind of paves the way to to having something more like decentralized while being interoperable um with the web we have today um and maybe some of the things which um i hope like uh transpire in the um the work we've done um so far we've announced today um uh is like kind of the commitment we have to um to having more privacy and having like more more decentralized network um so for instance we we're now starting privacy week which was like earlier as ecl um that we're committed to provide the tools to people to have uh provide the content more localized um like have content kept in like certain jurisdictions that like customers um can choose and like in a way this is something decentralized because you're like discussing about the user owning data the businesses being able like to decide where the content would be um while having like maybe at some point enable and scale up um to have the content enable wherever they are um it's also like very key to to privacy because um if the user are able to have their own data then you don't have to care about or you have to care less about like which jurisdiction the data is because you know it's with the user and so definitely that's something uh like very powerful we've been working on um in the research team and and clarifies it all yeah oh it's super exciting to see and i guess um do you do you have any sense for like why we started off offering gateways for ipfs and ethereum i mean there's there's so many different networks that are that are out there um but yeah i mean what's what's your take on that um definitely like that's a great question um so we started with ikfs in 2018 um and once again like was like an ask for for research project um because at the time we do seem like we could learn about like um how to operate um such like an emerging technology and the technology and the community might uh benefit from having um like like an easy and like fast access um to the content stored on the ipfs network um at the time and like um it has grown like a lot um since then there was um like only like a very limited number of like gateways available um and definitely providing something like on top of like now on top of capital workers um which is fast and secure like like offering um sss certificate for um each of those domain was definitely like something big uh that like we really wanted for the community and that allowed people to access content on ip that was stored on ipfs without having like to run the node on their own which comes with different uh like privacy implication because you would have to participate in a p2p network share your ip maybe like broadcast the content that you're willing to like that you want um having it with like a gateway like Cloudflare makes this like a lot more privacy preserving while benefiting from like the speed and security we have um poised the case is like slightly different like at the time we like assume was like one of the community we like really like and helped um to a nurse like the to expand like this um distributed web gateway um we are still looking like because the ecosystem is like very early but very active um at like all the options and all the gateway that like we might run um but that's something like we would focus on as we like graduate and natural yeah that definitely makes sense and i guess when when i see it i also think about like the ipfs and ethereum are supporting kind of like two very important concepts of like um putin networking overall right like ipfs is kind of handling distributed file storage and like storage in general whereas ethereum is yeah it's a cryptocurrency but it's also a way of distributed compute and like people are building like all of these decentralized applications on top of the ethereum network and so like by combining those you almost get like a full platform for enabling development on um on the blockchain or like doing things in a centralized way which is kind of like mirroring almost um like the way public cloud has emerged and like kind of the the kind of components that are required to support that um but just you know doing it doing it in a decentralized manner so super exciting to see that's very great um because we have only a few seconds uh we would like to stop our conversation there um so thank you thanks all for staying with us um and like right um after this segment you have even johnson for hacker time so like definitely exciting like he's um handing the show like on a weekly basis and sort of like every time like it's a pleasure thank you