Saving Web3 Protocols from the Fate of Web2 Companies

ETHCC 6 · July 2023 · 26:13

Maggie Love examines the dependence of web3 protocols on conventional cloud infrastructure. The speaker argues for more open, decentralized infrastructure to reduce vendor lock-in, censorship and deplatforming risks.

Speakers

Topics

From ETHCC 6, hosted by Ethereum France.

Transcript

foreign [Music] how are you good uh yes love to hear it so my name is Maggie love I'm going to be giving an infrastructure talk titled saving web 3 protocols from the faith of web 2 companies I'm super honored and excited to be on the ECC main stage today I've been coming to ECC since 2018 when it was in one room you'd listen to a talk and then walk to the common area and

chat everyone up so just incredible shout out to the organizers and everyone who's been a part of scaling each CC and making France a hub for the ethereum and broader blockchain community so let's get started uh just a bit about me I'll be quick I went to undergraduate at Northwestern University and I started my career at IBM and I actually did product strategy at IBM Watson for about two and a half years before I heard about blockchain and a corporate strategy meeting and said oh I have to go I have to go do that right now found my way to consensus in 2017 and then in 2018 I co-founded web3 cloud and web3 Cloud provides compute and storage

infrastructure for blockchain protocols we're a joint venture between consensus the ethereum company and Advanced Micro Devices the chip company so we've been around building infrastructure working with protocols helping them scale so they can focus on software for about five years now then in 2019 I saw the revolution that was going on in decentralized finance and decided I wanted more women in non-binary folks to be part of it so I built shifai and that's turned into a 13-week course on all the top use cases 1700 members from 50 plus countries so you could say I spend all my time at different foundational levels of the ecosystem building it up and really the goal of this talk oops

is convincing web 3 to make decentralized infrastructure a priority and a first-class metric and to get you to start thinking about infrastructure providers is not just Commodities where you go and rent compute but actually figuring out who can be a partner with you to keep your infrastructure decentralized and scalable and I'm going to take a bit of a different route I'm not going to sit here right now and just say like oh ethereum is centralized here and here and here or here's where all these other networks are centralized I'm going to take us through the lens of what's going on in the web 2 space and then bring us into what's going on in the web 3 space how many people are familiar with this article

software is eating the World by Mark Andreessen good a couple people if you're not familiar I would definitely go check it out um I guess then probably not many of you are familiar with that 10 years later his firm wrote an article called the cost of the cloud a trillion dollar Paradox so 10 years prior we're just talking about software software software thinking about scaling there and 10 years later we have some challenges because of how much we focused on software and left compute to be outsourced and there's no doubt that cloud is one of the most significant platform shifts in the history of computing it helps cultivate Innovation as company

resources are freed up to focus on products and New Growth but um I want you to put a number in your head how much do you think of the market value is being lost across the top public software companies due to Cloud impact on margins can you raise your hand if you think it's in the thousands across the top 50 Public software companies raise your hand if you think it's in the Millions raise your hand if you think it's in the billions okay so we are aware of what's going on in the cloud space so an estimated 100 billion dollars of market value is being lost across the top 50 software public companies due to Cloud impact on margins so once again

um the this leads to what we are calling the cloud oligarchy right and part of the reason that Amazon Google Microsoft uh represent a combined five trillion dollar market cap and they are because it's because they're all buffeted from Market competition and they have high profit margins driven in part by running their own infrastructure enabling even greater reinvestment into product and talent while buoying their own share prices right so cloud is an amazing step function and how we were able to build software we no longer had to run our personal uh large computers in order to make the internet work but at the same

time now we face this challenge of this Cloud oligarchy who can drive huge profits and have great margins because they own all the infrastructure and so I want to talk a bit more about the worldwide Enterprise spending on cloud and data centers I.T spending on cloud is growing at a base of over a hundred billion dollars annually and of course at early stage Cloud drives efficiencies and operations in economics so it's a wise choice for startups and they can spend their Capital elsewhere building their team building their product but as companies mature and they get to the growth and late growth stage the cost of cloud drives lower profit margins and an estimated 8 billion in

aggregate is spent on Cloud for the top 50 publicly traded software companies so this is what a big challenge that web2 is facing today and what web2 is doing right now is something called Cloud repatriation which just means bringing your infrastructure in home it's a challenge it's expensive and it's costly but between moving it back in-house or to third-party Solutions or building special purpose Hardware these companies can save billions of dollars and so for these top 50 Public software companies repatriation drives a 50 reduction in Cloud spend and a net Savings of around 4 billion and so moving infrastructure back on premise saves drives one-third

to one-half costs of running workloads in the cloud and so what we can take away from web 2 the tldr here is that infrastructure spend should be a first class metric that companies need to prioritize early often and outside of the cloud which I think all of us here in web 3 would definitely agree with that but I know of many web3 companies spending quite a massive amount on their Cloud infrastructure because they wanted to scale and then think about decentralization second now we're going to go into you know what web3 kind of offers but first uh how many of you know dhh the inventor of Ruby on Rails he has a company 37 signals and he's recently

been uh being very public about his process of taking his company off the public Cloud because they're spending over three million dollars on cloud so while you may have scaled and have all these customers and users I think about 80 of revenues uh somewhere around there are actually going to Cloud spend and so as I mentioned we now are in the hands of three main clouds that we are very dependent on for our services for our scaling and it's very hard to move on from them and so what are some of the problems that these Cloud oligarchy drives uh cost so we've spent a lot of time

talking about cost already outages and downtime you may have been aware of the AWS outage that U.S Virginia East like one of the biggest data centers went offline for three hours a couple weeks ago privacy and security so we know that our data we would prefer to be private and secure and traditional Cloud infrastructure does not really provide the right safety and then there's a lot of challenges for these Cloud providers with data sovereignty and Regulatory issues where in the land of gdpr here but most importantly what we have from centralized Cloud infrastructure and building on it is Extreme vendor lock-in it is very hard if you've built all of your infrastructure software apis into

AWS to have it ported to somewhere else you have to basically start all over so this extreme vendor lock-in has led to This Cloud oligarchy and if we look at the evolution of the cloud we have web 2 we're aware the three main players yes there are smaller bare metal providers and players out there as well then we have the hybrid Cloud right these other big players who wanted to get into this space but provide both on-prem and on-prem solutions and then we have this idea of a web 3 cloud or a web 3 decentralized compute and it was honestly started by the Bitcoin miners who are still out there today uh providing compute and storage for the Bitcoin Network or compute we have this idea around decentralized

physical infrastructure or d-pin a growing class of protocols that are using the economic incentives of tokens to bootstrap physical networks we have Solo validators in many of the blockchain protocols and of course we have node infrastructure providers so this is like the ideal case of what web3 looks like right independent parties coming together being bootstrapped by token economics to support infrastructure moving us away from being dependent on some big cloud providers and to take a step back Bitcoin really pioneered this Bitcoin invented this idea of a global distributed infrastructure Network through proof of work it proved that with the right incentives one can coordinate Capital

machines and label without a centralized entity right nobody was hired at the Bitcoin Network to scale it to what it was today the network itself distributed tokens in order to Galvanize people infrastructure providers developers to all be a part of it and so what was really exciting about Bitcoin and blockchain networks is that they offer this blank canvas for token distribution and crypto economics to facilitate coordination we call blockchains coordination machines and the ideal here is that you're able to really scale your network with the right incentives uh without Too Many Humans behind it and so what we hope with decentralized infrastructure is the right token

rewards allow a network to bootstrap the supply side to a point where users find its services attractive to use and we know that has been a bit of a challenge as well and so you might have heard this that web3 is cryptos attempt to reinvent the internet and it's an internet owned by the users and Builders orchestrated with tokens so if we look at web 2 to web 3 we're looking at shifting from an Internet that is closed off and focused on platforms to one that is more open and decentralized as it's made up of blockchain protocols run by a global community of participants incentivized and coordinated through blockchain protocol tokens and the defining feature

of moving from web 2 to web 3 is this idea that we can distribute value through these networks and it's why it's captured the hearts and Minds in so many of so many Builders as we are no longer just transferring bits across the internet but bits attached with value and the fact that we have this value to help bootstrap our networks and drive many coordinated activities has moved us really from a read write internet to a read write own and so if we look at the differences between web 2 and web 3 web 2 is centralized there are a few large players they have multi-trillion Dollar corporations they can censor they're expensive for developers you really many

times have to be invited into that company or ecosystem to build and really the corporations retain all the value and can drive a lot of profit from those margins what we imagine in web 3 and what we're all trying to build towards is a decentralized ecosystem distributed in long tail built of crypto incentivized protocols sensorless where developers can easily participate through open source and it has native value distributed to all of the builders and participants even the infrastructure providers so web 2 we're familiar with a lot of these companies if we're thinking about the cloud it's Azure Google and AWS and then in the web 3 bucket we have a lot of different protocols at the

infrastructure layer attempting to distribute and decentralize the internet and its infrastructure through tokens and these are four buckets of where we as an infrastructure provider are really excited about seeing decentralized infrastructure come to life so access and infrastructure decentralized RPC protocols how we actually read and write information from the blockchain compute and infrastructure so a lot of the layer ones out there as well as zero knowledge protocols storage infrastructure and staking infrastructure as well so all of these different protocols can really harness many infrastructure providers

and should be thinking about not just going straight to AWS or a cloud provider but really how to get the incentives right over a long term to excite infrastructure providers solo validators uh node Runners of all sizes to get involved and so really what is the value of decentralized infrastructure especially across those four buckets we talk a lot about credible neutrality in the blockchain space building these networks that are credibly neutral and in order to do that we are going to need to not be dependent on a single cloud provider resilience is a key factor right as different geographies have different regulatory bodies and policies we want

our networks to withstand what one jurisdiction does performance so we know that the performance of the overall network is enhanced especially an ethereum if the nodes are not centralized and one cloud provider or entity or Geth is not centralized we have better performance of the networks uh cost so you can really drive your costs down and Outsource your costs right by incentivizing infrastructure providers rewarding them really get away from those Cloud costs that you have if you have the right incentives and most importantly there's going to be this concept of no vendor lock-in so even today if you are doing staking

infrastructure you sign up with a bare metal provider let's say maybe like ovh you're likely going to be locked into a three-month six-month x-month contract and even if you're not getting the best pricing there of a new provider comes up it's actually hard to move your stake from that provider to a new one so what we want to see is infrastructure providers and technology that allows you to keep your tokens keep your keys and seamlessly move from move from one provider to another at a much more easy way without the costs and then we have this idea of open compute marketplaces as I mentioned so different protocols different applications can seamlessly move if there's an open market for decentralized

compute and get the best costs and efficiency and performance advantages by not being locked into one cloud provider on and at the end of the day that's offering all the protocols all the participants in the ecosystem greater freedom and control over how their network works how their infrastructure is working and then finally Agreements are code right we have this concept in the blockchain space where we're operating with math and algorithms and they're becoming the way that we form agreements and uh outside of parties in the traditional legal system so how do we invent protocols that also enforce the right types of agreements between multiple ecosystem participants in the

infrastructure space and so now that we know a bit you know what's going on in the web 2 space what's the ideal in the web 3 space from an infrastructure perspective what are the types of participants what are the types of use cases and protocols that would benefit from decentralized infrastructure I just want to check in on the ethereum ecosystem you could probably apply this more broadly to other ecosystems as well but you know what's kind of going on right now in terms of infrastructure in the space so if we look at node distribution right now and this was just pulled oop pulled from ethernodes.org about over 60 percent of ethereum nodes are being hosted and then if we break that down amazon.com has over 60 of the nodes on

their hosting infrastructure uh it's actually better than it was a few months ago at East Denver it was around 70 percent but still this is definitely uh should raise some eyebrows uh get people thinking about uh what that means I'm not here to cause alarm or say ethereum is centralized I've done that on Twitter it didn't go well um and so just being aware as somebody who might be participating in the node ecosystem you should know where your stake is being is being located and it is a bit of a concern that over um 60 of nodes are on AWS because if we look at web3 today what does that mean

for the cloud space we have Bitcoin miners we have dpin networks solo validators node infrastructure providers and we have AWS Google and Azure and we're quite dependent on them and so what's the reality for ethereum why is this happening Uh Kevin owaki tweeted my eat staking setup makes a hundred dollars a month and costs are one thousand dollars a month to host on AWS I'm going to move to home staking for cost decentralization reasons show me your setup please so not only is uh at a risk that more than half of the nodes are on these hosting providers but also it's expensive as well and you may not be aware of that

it's easy you can go to AWS and click spin up a node today but you may be thinking you're making a great return but it's actually costing you a lot more another point of centralization in the infrastructure space this has been talked a lot a lot about recently is Lido right so Lido is a liquid staking provider it's really easy to stake any amount of eat you want with them to participate in consensus and securing the network but they have uh you know I believe like 70 percent share of the ethereum staked which Rings The Alarm bells for many reasons and just an example of what could happen but what didn't happen Amazon web services briefly hit by a wide rating

outage impacting major websites Evan Van Ness he has it he's like the ethereum decentralization Maxi in his profile I don't know if that's a joke or not but you put AWS is back seems like an outage limited to us East Zero Effect on ethereum I imagine if there was some effect if AWS went down in Europe given how much Lido is in the cloud don't stake with pools who run in the cloud and we'll talk about really what the consequences of that are next in contrast and this isn't picking on a different blockchain protocol Phantom was uh hit by the AWS outage and currently impacting our services your assets are safe and we will keep you updated as things are resolved and really the idea of decentralized

infrastructure is that these outages don't happen that you never try to log into your wallet and you can't and you can't see your assets and you can't get a hold of your assets right self-sovereignty starts with always being able to access your assets and so what are the risks of cloud infrastructure for web3 if it's centralized a continued dependency and growth of web2 cloud providers in the space cost it is very costly to be running all of your software on AWS outages censorship and interestingly enough regulatory body claiming jurisdiction over a network whether we believe Gary or not he tweeted one day that you know most of ethereum nodes were hosted in

the United States therefore he gets to regulate the ethereum network and then there can be an increased attack surface if something like a cloud provider goes down and it does impact the network if nodes start being slashed uh we'll we'll have performance issues as well as you know with over one-third of nodes are hosted on a certain provider and uh they start to not be able to finalize for some reason let's say there is an outage their stake will be slashed until they're less than one third of the network so the network can go back to performing as it should and so you might lose some assets items goodbye nfts especially for nfts are actually hosted on a cloud provider and

not with ipfs or decentralized storage option and so how can we examine decentralization then if we know the risks how can we take this information and understand you know how decentralized is a network what could I be looking at you can be looking at the hosting provider so 60 plus percent of ethereum nodes are on centralized services you can look at pool distribution so look at Lido look at the exchanges look at stake us rocket pool I believe swell is a new Marketplace or provider node clients as well so actually having a large amount of the client on something like guess is also a risk for the ethereum network not being able to

finalize you can look at Geographic node distribution and coin ownership distribution so all of these things can be indicators of how decentralized a network is and how resilient and performant it will be because of that and so um you might be thinking like what can I do maybe you work at a protocol maybe it's launching soon you're thinking about how to incentivize more infrastructure providers you definitely want to be doing that there's a lot of compute out there looking for ways to put its resources to work and what the right Roi they'll be happy to do so a support Solo validators figure out how to make it easier educate people make it easier there's also distributed validator technology now better tokenomics in crypto economic incentives a lot of

people thought our protocols thought inflating until Infinity was the way to go but it's not really sustainable for the infrastructure providers I'll work with multiple providers and do it early and often it's really easy to just get spin up on AWS be the first to Market get all those vanity metrics but it will cost you in dollars in decentralization on the road uh collaborate with Partners on design and decentralize early and often and really quickly I just want to end with there's a lot of use cases that are driving the need for compute in this space we have nfts we have the explosion of the metaverse enhanced by what apple is doing and many web 2 companies are doing there we have an explosion in the zero knowledge space that you know web3

Cloud the infrastructure company I work at is partnering with many companies today to provide application specific Hardware a new infrastructure and then we have ai we know that AI is almost driving a shortage in the compute market and all of this could either lead to more centralization and infrastructure but we can work together to keep it decentralized as well and so you know what I'm really excited about and what I hope you took away from this talk today is there's many reasons to have uh decentralized infrastructure networks of computational resources and data that are distributed across multiple nodes or devices rather than being concentrated in a single point or the three clouds and so we believe at web3 Cloud that decentralized software

needs decentralized hardware and through the right economic incentives you can have a robust decentralized Network supported by many infrastructure providers and not dependent just on one and we're going to help you scale in a way that doesn't cost you billions of dollars when you're a wildly successful big protocol thank you very much my name is Maggie love [Applause] uh and I'm the co-founder of web3 cloud I would love to talk to you if you're thinking about how to get your network incentives correct if you're in the xero knowledge space distributed storage space if you're doing you know sprinkling AI on top of whatever you do

and you can find me at Twitter at Maggie love underscore have a great rest of your ECC [Applause] thank you [Music]

foreign [Music]