Intro to Ethereum Protocol | Mario Havel

[Music] all right Welcome to the uh special edition of the uh protocol Fellowship study group The Redux version of the intro to ethereum uh by Mario here um so I think I don't know Mario do you wna you want to do a little introduction like we did to the last one or just get right into it I think we can say a few words for about the the study group quickly um for people who are watching the recording and and have don't have a chance to catch the first one uh just to reiterate what we said in town hall I guess [Music] um uh yeah welcome to the study group welcome to the protocol Fellowship study group everyone and um uh yeah before we dive into the first talk uh just to say um well we are very happy to have you here we are glad that you want to learn about it here eum uh we created this um program based on our experience with protocol Fellowship eum protocol Fellowship which is a um more about handson development and as people are starting to work on etherium as contribute to the actual software parts of it uh we see that there is a big information Gap big education Gap it's a pretty steep learning curve even if you are coming from a programming computer science background to understand the ethereum context so we created this program to bridge the gap to provide some uh educational materials and also uh just help you navigate on a certain path like we still try to mimic this experience of a core developer of somebody who is diving into the protocol we don't provide all the resources because it's up to you what you are interested in and what you what you want to learn about um we uh we have some basic structure of the program uh for the first uh five weeks uh we will have uh one call a week one one talk a week uh starting with the general introduction to the protocol today uh then uh the uh we will cover the execution layer the consensus layer testing and security and finally research and a road map overview uh this the first five weeks and in the second half in the in the other five weeks we will have two calls a week uh divided in two tracks which are uh development and research so there are they two different domains and you can uh you can choose one of those or just you know switch between them or just uh follow both um and those will cover uh on the development s more uh deeper dive into the actual client code bases and the specification and we'll show you how you can actually participate uh as a programmer on the research side uh you will learn about the future of ethereum about uh the uh Improvement being researched developed right now so uh yeah so that's this overview of the program uh and with all of this we provide with these presentations we provide certain resources but our goal is to uh inspire you and and and show you all of these various domains and uh give you give you some starting points so you can actually Deep dive yourself um there is uh Viki EPF Wiki uh website coming with this program where as you notice this this program is coordinated we publish these calls there together with some resources but what we are hoping for is to create this Vicki as an actual um resource uh uh actual actual knowledge base uh with resources gathered from all around the ecosystem because uh if you already started diving into some topics you might notice that it's really scattered all around there are various hmds research posts and papers and so on it's there is not like a single place to go for this Hands-On um uh protocol topics and of course there are many great other vickys which we don't want to copy we just want to um have a context and links to the all available resources for somebody who is learning about protocol wants to become a protocol core developer so uh yeah feel free to contribute if you get uh interested in one of these topics that I mentioned today or we will um we will learn about in coming weeks uh dive into it uh you know D go about it Google about it uh find more details on it and uh uh contribute it back to the back to the vikii yeah uh that's that's that's overview of the program uh took like five minutes hope it's not too long um yeah uh let's let's so what did I forgot Josh should I should I add something do you want to add something uh no I think that is pretty sufficient um just want to uh reiterate that uh most of the uh conversation is going to be happening on our Discord server so if you aren't a part of that please join the Discord server uh during these calls we want to try to keep as much of the conversation to that server as possible there is a thread in the study group Channel there will be one for each uh of these different sessions that uh we can keep the conversations and questions and commentary on that thread so that everything uh is in one sort of centralized place and we don't lose it to the chat in in the streaming platform um yeah and outside of that uh Mario and I are are here to to assist you guys through this process and so if you have questions or comments please uh feel free to uh ping Us in that server um but yeah looking forward to this presentation from Mario I've seen it a couple times and it just gets better every time yeah just seen it so many times the next time it should be you doing it man probably yeah yeah thank you so much Josh and yeah uh you can use the please use the Discord thread as Josh said because I might ask some question during the presentation and uh you should ask me questions so feel free to drop them in there so we have it all in one place um yeah and I should share my screen and uh uh show you my presentation oh it's here okay let me know if you can see though well I suppose so I believe that this works better than the jitu before um yeah Again by the way if you are here after Monday Fiasco I'm so sorry about that uh the problem there was not the server uh was not the software but the server uh because we just didn't scale the bandwidth and the CPU of the machine to so many people we are very thrilled to see um so many people coming to this um stud group so yeah excited to have you guys and excited to dive into the protocol 101 with you so and this talk uh I will give you an overview of the etherum protocol uh of its technical Parts its design uh and uh the way it's developed uh the way we work on this with with uh uh colleagues across many protocol teams and uh but I believe that to understand uh the protocal development what is happening on ethereum now and the way it was developed the way it was designed you need to understand certain philosophy and uh fundamentals of it so uh I title you can see the prehistory because I want to talk also about the history about the way not just ethereum evolved but actually where it's coming from where um uh where uh these these values these uh these fundamentals uh uh have been have been created have been built um these ideas are not new uh they are uh going all the way back many many years so uh to uh to have good context and to be able able to grasp what is really important in the in the ecosystem I would like to dive all the way back to like 60s to you starting with Unix creation then free software movement and uh uh of course the uh the dawn of dawn of modern cryptography Cypher Punk movement because I believe that uh all of this and much more of course had impact on ethereum design on what we how we use ethereum today and uh how we see the crypto currency landscape today so yeah let's let's start with that I guess let me just uh show you the presentation because my face yeah you don't need to see my my face it's early morning here uh and yeah so let's start with Unix uh for those who uh don't recall what Unix is it's uh well it's basically two things uh first is an operating system but it's also sort of a philosophy so uh Unix has been developed in uh Bell labs in uh This research institute uh since 60s 70s and it was created by uh Richie and Ken Thompson uh if you recall this name then is Richi K Thompson they are creators of the sea language uh and even before c the B language uh but today the C of course povs like like everything around us like the the operating system you are using the browser you are using it's all uh deep down SE it's the today considered lowlevel language which uh uh which became very the most important building parts of of the of the modern software and uh together with the sea language uh these two people created the Unix operating system not just them actually it was many many developers researchers in the Bell labs uh working together but they the the Unix was designed together with the SE language uh in a certain way it uh uh when we talk about this Unix philosophy uh what we mostly mean what is the uh major part of it uh is a modularity because the operate this operating system is doesn't work in a way that it's a single program that tries to do everything uh including the kitchen sink as we say um but uh rather than that it's uh separate modules separate programs each of the program does what is best at and has a clean interface which can connect to others uh if you are Linux user today if you are maybe Mech user today uh this is what is what is uh the core of your system uh if you use piping uh taking output you can take output of one program and use directly as an input to the other this gives you very very powerful tools uh it's really empowering and for the user perspective it's a strong tooling but for for the developer perspective it's very practical and secure because having the system divided in these modules rather than just one program uh it gives you uh advantages so if there is an issue uh it's in one module and you can just fix that and it doesn't break rest of the system um and you can like encapsulate the complexity in these in certain modules um the documentary that I link here uh you can you can find it in the Vicki um it's it's phenomenal I really recommend watching it it's just amazing how these uh old uh at8 archive documentaries are able to articulate some maybe complicated topics but really clean really elegantly explained and you can see the people there Ken Thompson ex typing into pipes in in the early Unix explaining how it works it's it's really beautiful to see um and it really catches like this also the Spirit of uh this uh this environment in early' 70s I believe uh because at a time uh this this uh uh this institute the the B Labs but also like MIT or similar institutions all around the world actually like the early programmers uh worked in a way that well they collaborated also the way that Unix is created with these different modules and different people working on their part they need to talk to each other and and collaborate they had uh always you know open door to their office they could meet and chat and and uh they uh it was a strong collaborative environment where uh people would been building basically like one big thing together but ultimately it was uh everybody's uh doing their part collaborating with each other so I I see big part of this culture also reflected in the ethereum development today on a much different scale so yeah I really recommend this documentary uh it's like yeah uh it really well explains the uni like it's a paradigm of computing that was invented more than 50 years ago and we still use it today it's the way that we use computers today we didn't get anything better like from from the teletype writer to TTY that you had you have in your shell today it's literally the same thing um okay and uh continuing with this story uh we go to 80s um with uh uh the dawn of free software so uh as I mentioned like people early programmers they they collaborated uh their software was well open but it was because well to be able to Brun the software to compile it to fix it for your machine everybody just you know got the source code and um uh these early hackers as they call themselves um uh this was their their way of of operating this was the the way they collaborated but of course then in the 80s uh the commercialization of software comes the military comes and they create proprietary software they close the software uh from from its users and uh now this has certain consequences uh if you proprietary software is the is well uh the that which limits your freedom as a user so now um I would compare it to a uh to a to a house that you buy uh or have built but to change something to you know open Windows go to the basement and uh fix the heating whatnot you need to ask permission you need to get keys from the basement from the owner or the Builder or or so on and and the question is like whether you really own the house if you buy something like this and now you don't have permission to do what you want with it but the the Creator is is controlling it like whether whether it's your um whether it's really yours because this what is happening with computers today the computer you have in front of you the phone you have in front of you is um a multi-purpose device general purpose device it's a calculator which can calculate any number of operations in any way and you should be able to control it you should be able to tell it what you want to do with this machine uh however if you uh run the proprietary of software uh it's the developers the creators of the software controlling your machine and you don't really have an input you don't even know what's happening you don't know what the proprietary operating system is coming whether it's calling home sending your private data or or calculating uh uh Pi on infinite decimal numbers as I believe Windows wi was doing or why it was so slow anyway uh like we we uh when when we don't have we don't have the freedom uh we are uh very limited and this this this what this what Richard stalman realized so Richard stowman is often called the last Hecker of MIT because he for he dedicated his life to this he's now in his 70s and all his life he only used free software he dedicated uh his life to to keeping up this original hacker culture of of uh being able to share the software he in uh September 198 three he started a g project in the free software movement uh which defined uh the four basic freedoms of software the ability to run it any way you want to study the source code to modify it and redistribute it uh so he he gave freedom to users and he been uh he been advocating based on this all his life uh the the project that he started the G project uh might sound familiar to you uh if you are using Linux you might see g/ Linux because that's the operating system uh the goal of guo was to recreate Unix experience but in fully free software because even though that uh people somewhat openly collaborated on Unix it it wasn't truly truly free um uh the G project created licenses if you go to GitHub if you look for some of your favorite programs there you might see that it's licensed under GPL or the G general public license which was created by stallman uh to have give programmers a simple way how to how to uh license their program as a as a free software um and G project created many core utilities uh maybe if you learn C you used GCC the G compiler collection which is probably the most used um uh C compiler uh these days um and uh uh many of these core utilities is what we call uh shell the shell the operating system but the core of the operating system was a big big task which was taking a long time until in early '90s um lonus tals release uh the Linux uh as a kernel of an operating system under the GPL license and this could be coupled with the G shell creating the g/ Linux operating system you use today so all of this was was really started by by stalman I really much recommend the uh the stalk from from tedex Geneva where he really nicely explains it uh puts it to context why it's important for software to be free and in ethereum in all cryptocurrencies this is fundamental because the the the whole point of ethereum is to give freedom to users to uh be transparent to be trustless and uh to do this we need to know what's happening there uh when I run the ethereum software uh I need to be aware that it's doing exactly what it's supposed to that it aders to the to the um specification and uh to what I agree are the rules of eum network um then in in '90s there is the open source movement coming now you might notice that till now I only talk about free software but open source is a different different part of it uh the open source was created by some people uh from the free software Community who didn't like the branding let's say because free software kind of sounds like free beer it means Libra it means like uh you know free man free speech um but but it's not really good word if you want to do business with it to you know give away free software so the open source was kind of a rebranding with also its fundamentals said it's its uh its values written out but um the the open source movement talks more about the pragmatic side of it more about uh you know it's good to be open because you can be more audited have more security you can bring more contributors and so on so it's more more like this business pragmatic approach which is okay but it om it forgets the importance of the freedom there the whole point of this thing was to give the freedom to users so we say free and open source software or false that's where you see the uh This Acronym so for ethereum I believe all of these parts are important to be free and open source software to enable collaboration and give freedom to users okay I've been I should I should go on it's a a few minutes there um yeah uh next up important important invention in 1976 important part uh uh the new directions in cryptography this paper by D and Helman uh created a new paradigm in the modern cryptography uh basically uh if yeah it's it's kind of basics for most of the people here I guess but uh this is the creation of the asymmetri cryptography or what we also call public key cryptography or key pairs because before that uh before before this paper uh the encryption was done only in a one way symmetric way so uh since um uh since the um uh I don't know Cesar since the since the since uh thousands of years ago people uh encrypted messages in a way that I take a message and I somehow scramble the the words I change the change the letters According to some key According to some rules uh and if I pass this to the other party uh I need to give him uh the cipher text but also the rules the key how to decrypt it uh so we have a single key and single encrypted text but in in with as de cryptography we have we have a very strong tool which enables us to use one key we call it public key to to encrypt something and its corresponding private key which is only held uh by the by its owner uh to decrypt it only the private key can can can use a decryption but anybody can use the public key to encrypt a message this means that I can have this public key on my business card in my in my website and anybody can send me an encryp me email this was the dawn of encrypted emails uh mostly of the implementation of pgp um and of course if uh this Rings any Bell uh this is what all the cryptocurrencies use if you ever used ethereum if you ever used Bitcoin you used public key or some derivation of a public public key uh as an uh as an U address and the private key is What let you spend because of another unique feature of this key pair cryptography uh which is digital signatures uh digital signatures are mathematical proof uh which work in a way that uh uh with the private key we sign data we can sign any piece of data it can be terabyte file it can be a simple message uh we sign this data and the signature is some uh mathematical proof which can be verified uh against the public key so anybody who is seeing this signature and has your public key can verify that only the owner of the private key could produce this signature this this is the way as the transactions are done in ethereum when I send a transaction I create a proof that I'm owner of this address and I'm signing the message that I want this amount of funds to be transferred here so um to another publicy to another address right um and uh yeah this today we we take this for granted it's everywhere you are connected using TLS the htps in your browser here that's end to end encryption using using exactly this this technique uh when you are using signal or Whatsapp it's using uh using some form of this cryptography so it's really everywhere today but that wasn't really the true uh in the 70s and 80s when with the dis dawn of encrypted email with pgp the public pretty good privacy fe uh tooling uh This was um this was considered dangerous was considered a weapon um so uh the thing is that uh before uh before the uh the computers and internet spreading to uh to uh to houses you know to uh uh just civilians all over the place uh it was mostly used really just by military and now anybody can encrypt anybody can uh achieve their digital privacy and governments didn't I don't want to say they didn't like it they didn't have framework they just didn't understand it it was something completely new and they considered cryptography a military Arsenal a weapon which cannot be uh used by civilians has to be regulated cannot be exported sorry many of these early cryptographers get into trouble because their open source free software cryptography being being uh um uh exported so for example uh Phil simmerman the creator of bgp uh uh was under threat uh that he's going to be uh that he's going to go on trial for treason or something because he created this Email encryption which is just a you know mathematics programmed in a in a software that's it and um uh what he did famously he he he created uh a book published by by MIT press which uh contained the whole source code including make files of the pgp version like 2.7 or something so and then they could export this book and the the US doesn't have any regulation against exporting books so now like he protects himself but we also can see that it's sort of a free speech to just you know talk about the rules of mathematics and write them in a language of a software or like literally the programming language just a form of language right so so there uh there was this community of people who've been doing these things building the software advocating based on it um this this very popular uh movement called cyer punks here on this picture looking kind of scared with the mask but we can unmask them um the cyer punks were many different people from uh different uh Academia from business uh computer scientists programmers or just researchers um uh who've been uh building these s they started cypherpunk mailing list which is a famous place where many of these tools were announced like SSH was announced there I believe Bitcoin was announced there um on the right here uh on the on the very right is um uh uh Gilmore um who' been U uh who've been uh the founder of electonic Frontier Foundation uh eff very important organization till today uh in the middle uh Hughes Eric Hughes the author of cypherpunk Manifesto he actually um WR wrote down uh these principles that The Cypher punks believed then in a Manifesto uh which uh um except of it is here on the slide uh where he talks about that part that we buil uh some distributed open free software which cannot be which cannot be um somehow forbidden forgotten because it's really uh um uh uh uh because it's free and it's it's uh just form of a form of a speech form of a um uh explanation of mathematics for example the also famously they produced uh t-shirts with the cryptography code on it so like you know I have illegal words on my t-shirt this is illegal mathematical notation here um so yeah uh I really recommend getting inspired by the Cy Bank Manifesto uh on the left here is Timothy s who um was I think Chief scientist at Intel he has some uh interesting inventions there but uh in this context he uh as a as a one of the founders of The Cypher Bank mailing Le of The Cypher movement he um he takes this idea a bit further he creates Manifesto of crypto Anarchy where he uh talks about uh about these tools not just as like we are able to create them and and and have privacy and protect ourselves but we can actually create a digital realm this digital world where we can uh be completely free protected we can uh have uh digital money Anonymous uh uh Anonymous credentials uh our own reputation uh uh uh systems our own markets uh Etc so like uh these were the people already envisioning uh digital currency is based on cryptography um in and well early '90s and in late 80s we had we had this idea already David cha famously created the Dig cach one of the one of The Cypher punks as well um so we see these attempts and we see uh these ideas pointing to like this like some uh uncensorable free digital platforms uh yeah and with all of this the question that arises like how ethereum fits in all of this uh this is of course like my somewhat subjective point of view and I don't want to I want to influence your opinion and I want to ask you guys like uh I I kind of described in this view on History how I see ethereum where I see it's coming from and what are its values and uh in the last part kind of what it is but I want to ask you what do you think the ethereum is can you describe ethereum in one word or one sentence write it to the the Discord chat I'm going to check it uh in the in the thread World computer cool cool yeah John Gilmore superent employee yeah yeah uh good uh that's his his credits nonstable lovely computer oh I love that yeah we have many we have many many different opinion decentralized VB decentralized computer trustless trust uh World computer as well and trustless state machine oh I love that cool cool uh well uh we have many different opinions but I got to say all of you are correct uh of course I I think there is no single way we could Define ethereum uh because everybody sees a different part of it oh new internet it's also cool one I believe all of this is right right uh I believe that ethereum is this digital realm that uh anybody can find their place in anybody can anybody can see it it's uh it from different perspective and uh you know depends on what's your motivation where you're coming from what you're interested in what you're working on you have a certain tunnel vision of like what the ethereum is for you how do you envision it so this uh this little elephant here is what we are talking about and uh in this talk of course I need to focus on the whole elephant on the on the technical part of the protocol itself um I uh need to talk more about like handson what is what is ethereum um and uh yeah uh so because what would we have all in come on even if we have like different views on it uh we want to achieve these goals we want to achieve these visions and that means that we need to um use and improve ethereum to uh to be able to get there and to answer like what is ethereum uh yeah in there there are many simple terms uh another description of what is ethereum of course you could Point people to white paper if somebody asks what ethereum it's a description of what ethereum is uh from 10 years ago now uh and here in vitalic revealing it on bitcoin Miami Lineup of that conference changed quite a lot um yeah so so that's that's one way uh the white paper is a more of a general high level description it's more of an ideas of how this network should work but we can take it further to yellow paper which is actual formal specific ation of ethereum it uh uh it uh it's written by Kevin Wood one of the co-founders of etherium uh Al the creator of the original C++ implementation who actually uh sit down to to write down how it how it supposed to work not just in theory but practically uh using using mathematics using um using formal Expressions how it's How ethereum uh is supposed to be implemented um according to the specification so that that's one way and you see here that it's Berlin version the current version is Paris so it's already uh out of date but uh it's out of date because ethereum is changing this is just like the original this was just like some original Vision uh but it's not the only way where only place where theorum is specified it's not just metalic or Gavin or luin or these Founders who Define what ethereum is but it's the community so using this process of ethereum Improvement proposals anybody can propose a a change to ethereum Improvement or or um uh change the belief is uh is um important to it um the EAP process is inspired by the bip process from Bitcoin Improvement proposals started by Amir Taki uh which is inspired by the similar process in Python and it's generally a community way of uh working uh on a on uh this big distributed project um but having a you know yellow paper plus eips and you are supposed to implement something based on it it's kind of a mess right like all of these different documents so we actually have a proper specification in these two repositories uh that's the execution specs and consensus specs uh with also the apis here um that's what really defines all of the ethereum and the way it's supposed to work uh it's written in Python and uh for for some people people like me it's hard for me to read the mathematical notations in yellow paper and in the python it actually is even more readable and of course you can generate tests based on this you can test your implementation against the specification so so it's very practical and very very uh useful way that we specified ethereum here uh and of course if you are not progr or if you just want to learn about the way it works uh it's pretty heavy to read a code I would recommend reading annotated specification which comes with explanation uh how it works why it works this way um and gives you a context so for example the one from benedon here um and the E2 book um I'm sorry so um there are versions of the specification different versions of ethereum we also call it hard fors uh the ethereum upgrades which have been happening well now for 10 years and um starting with the frontier Homestead many years ago uh uh now um you you can you can see all of the history all of the upgrades in uh in the ethereum orc uh where you see what eips what upgrades have been coupled in each of these each of these upgrades um however um even though ethereum changes a lot is developed uh heavily over these well 10 years now um it's uh it's always aderes to certain principles there are basic tenants that uh ethereum design must follow uh it was outlined in this now non-existing website uh archived there it's really nice explain there I recommend diving into that uh but the point is that these developers need to have certain values uh need to need to follow um some rules some some design decisions uh and uh we cannot take tradeoffs which would uh which would disprove this so uh yeah simple City universality modularity non-discrimination agility um these are the basic uh basic principles uh in a way that ethereum needs to be as simple as possible in a way that needs to be possible to um somewhat easily uh implemented it needs to be you know re implementable if you compare it to like I don't know Linux kernel like I cannot imagine somebody just like writing another version of that um universality so uh ethereum also like the generalization and non no no features in the in the in the later part um the universality and non generalization means that ethereum is not um uh yeah not discriminating it's uh it's uh it's not biased towards certain features uh ethereum is not supposed to do a single thing but rather be a generalized platform for whatever people want to use uh this means that the core protocol does not cannot include um features which would which would uh Define some some beh some user Behavior or something uh more high level like for example uh there cannot be like token implemented token standard implemented uh in the core protocol it's with the ERC people created on top of ethereum but um it's not uh uh it's not within the the co the protocol itself um the modularity I already talked about and you will see it as I show you the protocol later but this what we follow here in the theorum is that the uh also with the encapsulated complexity mentioned here uh we we encapsulated different parts of ethereum different uh uh uh parts of the network uh and they talk to each other within some relatively simple API so we are able to uh debug it we are able uh to have more security if only one of these parts fail and uh yeah uh the non one thing I want to elaborate is the non-risk version that uh as the ethereum changes uh changes are not uh are are conditioned by uh What uh benefit they bring so the developers are willing to take certain risk for a certain benefit right uh if there is a small change which bring us a huge benefit it definitely worth doing it so that's uh that's the principle and I'm sorry guys it's too early for me um yeah now I want to ask when I talked about the different modules different parts of ethereum I always mentioned like parts of the protocol and whatnot what would you consider is the heart of ethereum the the the most important part we should start from let's build ethereum together and uh let's uh let's see what would you consider is the is the uh brain the the heart of ethereum yeah uh right right uh right into uh the group okay see evm consensus people oh that's nice um H so what else uh feel free to feel free to write Phill the FAS okay I'm talking more about the it's really nice to say the people or philosophy I mean the technically the modular the part okay that data structures uh yeah now you do don't see anything because I'm I'm waiting for you to reveal it um the peer-to-peer Network okay Miracle tree okay guys uh I again I believe that I believe that you are right uh I believe that all of you are right because these are all important part of ethereum uh I wanted to start from um from um from somewhere so when I ask about the heart of ethereum I wanted to start with evm uh that's like uh often often mentioned important part of ethereum and it's it's pretty much uh one of the core um very etherum specific features so what is evm um it's um well it's a basically a virtualized CPU uh now if you are using your computer here you either have x86 or arm uh based CPU with like instruction set specific to the architecture okay that's a hardware which implements some instruction sets some BAS basic assembly some basic instructions which are able to do logic arithmetic and work with data to uh to well do anything like your computer can do anything everything um uh to be a general purpose machine so that's that's on the hardware level but in in software in in programming in computer science what we often do is virtualize these instruction sets uh we create virtual machines for example if you are Developer again maybe you work worked with uh jvm the Java virtual machine uh these machines are created so that if you write you know if you write a certain code it might do it might behave somewhat differently on different machines or different manufacturers which can understand some instruction slightly differently let's say uh famously said like it works in my machine so now if you want to I replicate this experience it works in my machine I create a virtual environment which is always the same on every machine because if I execute something in this virtual environment uh it will always end up in the same result that's what evm is uh it's a virtual machine which is which has its own instruction set with like 30ish something of codes um uh and these these can these are T in complete they can do anything um the uh the the the evm is is the actual execution engine the environment where uh the ethereum programs are are are run and now if all of the computers every computer running ethereum has the same virtual machine or the same imple not just the same implementation but it's implemented the the same specification can be implemented now in different languages and different ways and if it's uh still uh follows the specification the the result of the execution the result of the run of the program within this virtual machine will always be the same it means that all of these ethereum clients can agree on something um I would also mention here the uh swich complexity that I have in previous slide uh because that's that's how ethereum works so we have you might probably you heard about solidity or Viper or some other these high level languages which are readable and auditable and then we have like very complex compiler which is like a huge part and it's it's a beast uh very complex and on the bottom we have the evm bite code the instructions for this virtual machine which again there is not so many of them and it's somehow readable somehow uh formally verifiable uh what the actual virtual machine does so when we have like these two non-complex pieces with the complexity in the middle we create a complexity sandwich with the evm on the bottom but for evm to work it needs some data it needs to input and output some data so for that we have um uh the state um so the the data in ethereum the data structure database which holds all of the uh all of the addresses its balances its current state uh the current state of the network of all of the contracts all of the uh all um the the the the code and the data of all of the contracts right uh this is what we call the state I know many of you are anarchists you don't like state in ethereum we have state because ethereum is a state machine uh State machine is a principle from uh from a computer science where we have again a machine which can result in different states and different inputs um so so evm processes uh outputs some data updates the state uh creates we call it State transition based on some input which is transactions so transactions uh are these instructions on how do the state transition is supposed to happen and if it's correct if it if it works within the rules of the of the evm of the network it uh creates a successful State transition we update the state of ethereum um if you are familiar with Bitcoin design if you uh or similar utxo based uh cryptocurrencies this is where the M major differen is ethereum has a single Global State basically one database which has the current state the current value of everything uh in Bitcoin or R currencies we have like utxos we have uh uh just unspent outputs of each transaction and uh not a single single uh single state so so this is this is uh what what is one of the design decisions in ethereum one way it works the state is a big merco tree merco Patricia Tre and uh the transactions are um are stored in a different tree but update the state uh they are uh the transactions of course come from me pool which is um the pool of all the transactions come into the network and those are uh those are uh gossiped over peer-to-peer Network um so the peer-to-peer is um the communication interface with other clients uh with uh other nodes in the network which can uh broadcast a transaction so your uh uh your transactions uh go through um this Pew topw Network and each client which receives it verifies whether it's correct and ex and then if it's correct uh again broadcast it to all of these peers so um even if the network is big and decentral is that this uh the transaction spreads to all of the network and uh this all of this much more but let's say this uh is the execution layer this is what we call the execution part of ethereum um if you are using some uh ethereum wallet or if you are using some dab some uh you know dist application um you are talking to this to the execution layer of etherum using some standardized uh API uh it's RPC Json RPC API uh which can read the data if you want to know you know how much eat is on this account you need to read the state so you read the data from the execution layer or if you want to write if you want to send a transaction your wallet signs it and then sends it to this client to broadcast it to the network over the the client receives it verifies it and sends to the p-to-p layer so this is the the execution part um now another API that uh the client implements it's engine API uh which talks to uh the consensus part of it so uh how does this execution know what is correct what is the head of the chain what all of the other uh all of the other notes agree on uh in the I put a link to the uh bft the uh presting fa tolerance problem in um in the in the docs and this is this is where it comes like uh we need to agree on what is the current state of the whole ethereum not just as I see as one single piece of the network and uh this is uh dictated to the execution client or this execution layer by a fork Choice which uh uh is of course uh uh managed by the consensus algorithm of ethereum the lmd ghost or the latest message driven uh greediest heaviest observed sub three uh which is very jargon heavy way to explain uh this uh this for Trace Rule and uh and this uh uh this cons mechanism um So based on which which based on uh number of uh the most number of attestations the the fork or the the the threee chain with the with the most votes the most stations is the one which is considered canonical and um uh then is dictated by Fork choice to the to the client uh and this together with its own peer-to-peer net work and some other parts with the current hardw work we have the blops we have the r which dictates the randomness uh all of this this is the consensus layer so um ethereum right now is these two layers connected to each other and um over the engine API and it's uh uh and uh yeah and the consensus also has its own API you can also communicate with this client again to learn now about state but not the E theorem itself not about about like accounts and transactions but about the consensus State how many validators out there how many uh what validators do what attestations what is the finalized uh head of the chain Etc we we we ask uh from the beacon API of the consensus layer uh if you are an validator if you are participating in the consensus uh your beon node or this this cons layer verifies uh the latest block sends it to the execution layer to verify uh uh the Integrity of the block and then uh if it's considered valid the validator uses its keys to to sign an attestation okay I believe that this network uh this this uh block is correct I attest to it and sends this data the signature uh to the consensus layer which where it's broadcasted over its own peer-to-peer Network so the consensus has its own peerto peer it's a different from from execution layer this is so this whole thing is like simplified overview of different parts of the protocol again from what I said before you can really like each of these boxes are huge rabbit holes you can dive into each of these and uh and become an expert of uh on literally every single box here um and to make it bit more high level bit more abstract so these are implemented in clients in the execution engine the execution client or the execution layer client and the beacon note or the cons client client different words for those each of those have the own peer-to-peer Network on the right C the execution clients are only talking to the execution clients and the beon notes are only talking to beon notes they are they have they have their own communication protocols uh it's Dev peer-to-peer for the execution and a lip peer-to-peer for the for the beacon and they have their own apis which uh provide uh different uh different data and they to be connected over the engine API so so yeah this is this is um uh this is like high level overview of of uh what constitutes an ethereum node or ethereum client these two parts connected uh but of course this is somewhat new uh uh this is uh this is something that that happened after the merge um so to show you the the peer-to-peer network uh this is how ethereum looks like right just to give it's again something basic but just to give you an idea of on the left what is centralized if you're watching this stream right now or if you're watching the recording your computers are uh connecting to a single server which is sending data to all of you guys um but in ethereum it's similar to more something on the right uh maybe in the chart everybody is connected with everybody which is not really true uh you are connected to some notes those are connected to some Etc ET you Cann be connected to everyone but uh each node can be connected to tens or hundreds of different peers we call them peers because they are on the same level this like this network doesn't have any any Master uh no Masters no single point of failures as we say so uh it has uh it has all all notes on the on the same level everybody's peer and uh can talk to each other and can connect so so your so the client discovers the notes around him connects to them and exchanges the gossip data over the over the particular layer and this way so uh yeah and to show you how the block looks like with uh this consensus and execution client connected uh this wasn't the true before the merge on the left here you see the old ethereum before the merge the proof of work ethereum uh where the beacon chain and the ethereum were basically separate things they were separate networks the beacon chain was uh looking into the ethereum for the deposit contract for the validated data there but uh the ethereum itself didn't learn didn't use uh the big engine in any way in the middle this is the moment of the merge where the uh the the block from the execution layer becomes part of the block of the beacon chain so in the execution layer we have the uh we have the transaction data you see all of the transactions from 0 to n and they're they're in a meracle tree and their root is in the in the in the header um and the hash of this header is the hash of the whole block right and uh the beacon chain uh is now you see they they it holds all the all the attestations attestation uh zero to n so uh the uh signatures which are confirming that this execution layer block this this transaction data this this block was verified uh the then this deposits and exits uh from the from the validators um and yeah this is this is how the on the right uh yeah the the arrows are pointing to the left now I realize might look bit confusing but it's of course this is the blockchain right the in the blockchain we hash the previous we include a hash of the previous block so this block references the previous one uh so on the right here you see how the how the blocks look like today with the execution data with transaction data within the within the beacon chain and now yeah so this this this little thing this Beacon and execution not like all of these parts implemented in these clients are actually written in some languages as actual programs that you can run right that's the whole point so this is where it comes together this is how you use ethereum so if you want to become an ethereum user which means use ethereum privately trustless securely you run some of these clients which verify the the the data in the network and let you interact with it let you connect to um to the actual ethereum Network to send transactions to read its data without any third parties this is this is the actual ethereum so the implementation of the execution layer of the the what was the bottom part in my chart uh is um go ethereum in go nethermind in net C Bess in Java Aron in go uh was originally stored as a turbo Gat as a fork of go ethereum but later developed its like very unique architecture uh but still keeps some part of the go cod code base uh or like still in go while red uh is written in Rust and it's basically using the Argon architecture silk worm is uh uh rest is and Al red like these four on the top are production ready those are running the network the network consists of instanes of the the computers running the software in the first row in the second row the red is uh in Alpha there are some people running it but it's like you know not recommended for production use silk worm is still in development it's also implementation of the Aron architecture in more modular way and then there is ethereum JS which is in JavaScript typescript and uh this one is more of an experimental client for um used for testing us used for verifying the specs Etc and then on the consensus layer side the upper part which drives the uh the the consensus the validators uh these are all production ready clients which are again running the network um the prism in go lious and rest teu and Java from uh nimus and Nim and L typescript uh there are other clients which are not production which are in development and there's one proprietary client BR that I don't mentioned and there is a client in alexir developed by L class which is also like in production ready but uh there are others in development as well but these are those that are there are ready you can actually go to their repositories learn about them because as I said before all of this is free and open source so you can actually look into the client code bases to make sure that you know they uh follow the specs you can contribute there uh and to ensure the security stability of the network uh this all of these clients need to be tested need to be heavily tested because well the network is out there running like huge amounts of money and people's funds and whatnot so we cannot just you know uh take it recklessly and um uh we heavily test these clients the the client teams the developers and uh other test teams which are de dedicated just to testing and security are always especially before these hard for before these upgrades are heavily testing the clients uh so first of all there is uh the ethereum tests which is just uh basic station tests and like basic sets of tests uh that the client should pass um the this coming from the uh from the origin etherum design with only one layer not the consensus layer so tests as the execution layer uh similar to execution spec tests which are uh which are based on the uh based on the uh the specs uh very interesting very important uh is Hive uh the one in the middle lower middle Hive is um is a end to endend harness tool it's like a huge tool which lets you do everything you can spin Network works with different clients in it and run them together through different tests uh there are many simulations that you can run using it and uh it creates really like complex and uh strong uh testing Tool uh crosis is uh Al lets you like create a a local test net with different clients and then there is fuzzing which is like uh heavily testing different edge cases uh Mario is doing the the fuzzing of different parts of the clients and then other kind of testing like you know benchmarking stress testing making sure that like what if we deploy huge amount of validators or huge amount of this or that like does it go Belly Up does it survive what if we create a huge block like can the client process it uh so like testing the actual performance the features of the client uh Shadow Forks are very important testing Parts these are like kind of devops kind of like a uh more High l testing uh in a way that like we take uh part of the network we take like 20 clients uh which are synced to mainnet and uh then we put them through a hard Fork uh as they will go in the in production but uh they just they just create like their own uh Fork Network and we see uh what happened uh we see in this case on the left you see that like two or four of these clients or client pairs uh went went went down uh during the during the the shadow Fork um and yeah client uh of course like clients are testing their their uh their clients themselves and uh with like continuous integration uh they uh they can uh test every PR every comment to the code so uh to reiterate on what I what I described here how ethereum was created and how maybe everything that uh was outside of the ethereum but how these processes in etherum work is that we start with some ideas some general description which then needs to be researched uh uh for its feasibility for the for the effectivity we figure out how to make this idea actually real in some doable practical way uh based on that it's specified uh the specific apption are implemented in the software the software gets tested and then it gets rejected because nobody likes it uh it's really a thing that um uh some of course like many of these things are adopted running on mainnet but it's it's it's really a thing in if you are if you are ethereum researcher developer you can spend literally years like I know people who spend like three years or more working on a single thing single upgrade for ethereum which is really good and useful but it has some HK or something so it's still not adopted it's still being rejected and you have to come back uh you come back from well any point on this chart you might uh get rejected and go back to researching your idea you might fail at testing and you need to fix your implementation or the implementation doesn't follow the specs properly and you need to you need to um you need to fix that so yeah it's like a very uh iterative cycle like it's always always one uh two step forward one step back I would say uh and it's not it's not easy so now you might ask like okay how does whole process with all of these different developers all of these different teams uh Works uh so I showed you like here the different clients right there are five there are uh I'm sorry uh seven clients on this page five clients here I mentioned some others which are being developed so we have these testing teams client teams uh research teams all together it's basically like 20 different teams around 20 different teams and like maybe 200 plus people from these different teams and independently working on ethereum so the question is how they all coordinate how it's possible that we ship how it's possible that we agree on what ethereum is supposed to look like so uh first there there are many parts to it I me start from the bottom e research it's a forum where people post their ideas people post uh um something coming in these first two categories um so when I want to you know when I when I see some problem or some missing feature in ethereum or something that should be changed tweaked uh I post uh I can post on the research where I show outcome of my research my ideas and the way uh that I I tested it the way uh that I propose it and it gets heavily discussed there um these Community uh proposals are then formalized as eips that I mentioned before uh the eips have uh their discussion uh under uh eat magicians forums another Forum which is not not discussing like general research ideas like e research but specific eips and uh the way they are uh they are specified um and then like EAP AP is a process about the eaps themselves so we also have like community you know editing the eips and making sure that you know they make sense they they are formally correct this is done mostly by c herit community uh they also have this uh peep at EIP series with videos uh uh giving a space to like uh EIP uh Champions to like uh talk about their proposals um so these are the Forum discussions the public discussions a lot of discussion a lot of uh of coordination is happening in r& Discord uh the r& Discord is a server similar to our EPS one uh with uh where people from many different teams participate in discussion and and updates and so on here uh uh on the R&D Discord is also where uh you can see a release of the def calls so now there are different calls of developers which are regular and maybe uh the most important part of this to follow if you want to learn about like what's happening now in ethereum so there are what we call ACD calls the all core devs and these are focused on either execution layer or consensus layer always on Thursday like today um so later today they will be called with like maybe I don't know 50 people representatives of different teams uh discussing the current changes the uh actions before uh before the hard Fork so right now uh the biggest topic topic is planning of the upcoming 10ap upgrade and uh I believe today was a deadline for the client releases so uh if somebody didn't uh so so so on on these on these calls they um they uh agreed like okay this is a reasonable deadline for us of each each of the these teams to ship this feature so they so they talk and they are able to coordinate using this and then they are like not all coord but like specific uh feature or like calls dedicated to implementers of specific feature like we saw 4844 or when we t uh test net I'm missing T there uh when uh there was gly deprecated and and uh we were deploying in test Nets or EF was a feature which was discussed like this uh there's also wallet deps coordination call for like standards for wallets which is a bit outside of the protocol but worth mentioning uh so yeah uh these are these are the regular calls it's it's word following and it's it's a it's it's here where the coordination actually happens where people like some ideas and it's not that simple you know like people uh I want I don't want to say argue because like it's civilized discussion but still it can be a passionate discussion which uh takes sometimes takes weeks and months before some some something is agreed on and uh that's very important that people come from different perspectives different backgrounds with different teams voicing their opinions and this is all public you can uh it's streamed on YouTube and you can even participate by joining the R&D Discord so anybody can voice their opinion about the ethereum core uh yeah this is coordinated uh in the PM repo so you can see all of these calls uh they schedule there the topics to be discussed there you can propose a topic and so on uh you can see the recordings the uh uh the uh record of all of the previous uh meetings in uh the uh ethereum PM project management repository this is how it looks like um one Live Hack is that if you are uh uh if you are pretending you are working you know somewhere you can just put this on your screen the YouTube call and act like it's uh uh you in a call um and yeah in addition to this the people also meet in person uh because sometimes it's you know better to talk face to face and and just figure out like uh things hands on uh and this is a famous picture where uh people are clapping to the first uh merch like this was the first time that the merch was actually tested um actually tried imple uh implemented in in Greece uh was 20 2020 I think uh no no 2021 and then this was this was last year a year ago in Austria where uh people were finishing uh withdrawals working on 4844 again like 100 people meeting uh in different sessions uh uh focusing on on uh shipping the the next thing um so yeah that's that's how the developers coordinated but the question is like okay what these developers are actually developing what they are working on where is it coming from well there are researchers and uh people proposing these these new upgrades new features uh and uh there is a some lose road map of what is happening in theum right now when I see a lose road map you maybe saw this uh this is a like the road map uh from uh vitalic who makes this beautiful beautiful chart however uh this is not like you know the definitive ethereum road map this is not like this has to happen this will happen this is just uh enumeration of the current things being worked on the current idea as uh the the landscape of the R&D right now because it changes and uh maybe we'll figure out that some of these boxes actually don't fit the ethereum vision or come with some trade outs and so on as we go along um so to give you an idea what is happening right now again like what I said about the boxes before uh in in my my chart of in my my visualization of the ethereum there again you can deep dive and become expert in any of these boxes basically uh and uh the first big part is the the merge um which was successfully finished we merged the ethereum with the proof of stake however there are some upgrades to the uh to the proof of stake uh like single slot finality making faster finalization so we can still improve the the consus protocol um The Surge uh in this case is focused on scal scalability so to to uh able through of more transaction uh that includes implementing P implementing data Ava availability sharting and the 4 for for blobs which are coming in the upcoming upgrade uh Scourge is focused on economics that's M part uh so the the maximum minor extractable value uh which uh which has some issues can can cause some issues in the network like uh uh like centralization or um or problem problems building blocks so there there are proposals like uh protocal Builder separation and Shrine in protocol um H there's some sensorship resistance and so on uh with the also also there is mentioned staking economics it's also like uh proof of stake improvements of uh exploring how we can uh use the uh the the crypto economics there more effectively uh uh the verge is the veral tri so as I mentioned before the current state of ethereum is the Merle Patricia uh three which is uh like one way we have been doing it for 10 years but now there is this proposal in testing in development for uh Vector commitment Merle trees which are using vectors uh to create a tree which enables like uh uh Witnesses and more simple proofs like a proofs which can be uh which can enable stateless for uh much lighter clients so more scalability um Purge uh means getting rid of the technical depth and uh old data so like there is bunch of State bunch of old data which is never going to be used but the clients the nodes are required to hold it um so we can we can actually have uh some improvements to to uh have the old data expire uh Splurge is the uh the rest of it like evm improvements uh vdfs Etc yeah so this this like a short overview to give you an idea what's happening in different tracks different urges of ethereum um ultimately all of this is trying to fight this monster of a triangle maybe you saw this before and this is a problem of uh the blockchain itself because well the whole point of ethereum is to be this this distribute the decentralized network which uh which is secure and uh to enable to decentralization what I said before to run one of these clients to actually use ethereum in a trustless and secure way you need to run one of these clients and that can take uh to verify the uh so it means that you need to actually verify the transa the data in the network so if I want to use ethereum I use my own client which verified all of the history and the current state um however that takes hundreds of gigabytes of data and tens or hundreds of hours of the CPU time so it's pretty heavy to run on your laptop and uh if we put more data which needs to be verified of course it becomes more and more heavier and if we keep doing this becomes exponentially heavy and ultimately there there would be only few uh few single single uh institutions running these nodes and that would be centralizing that would uh give us the single point of failure so ethereum needs to be decentralized ethereum is limiting itself it's limiting its onchain uh uh activity to the point where you can run this one of these clients on Raspberry Pi on like very cheap low powered 100 200 bucks device plus SSD uh so so yeah uh this this is the point and now what we what these researchers are fighting is of course different problems but this is a big part of it to be able to run ethereum trustless uh and at the same time allow more data being processed which is kind of oxymoron kind of kind of um uh dilemma here but uh there are ways uh outlined here that that can improve it The Verge the search um yeah The Purge uh which are focused on this and yeah so there are there are there are many different improvements to it here many different things that we can we can fix and uh ultimately uh you might think that well it's kind of an infinite fight it's an infinite Garden right uh it's uh never ending always growing always to be trimmed and maintained garden and uh that's somewhat true like I believe that ethereum core protocol is a garden which always needs some gardeners some people to take care of it uh and right now it needs to expand and fix many things and you know uh fix uh uh fix bunch of its uh uh issues to get to some better State however at one point there will be beautiful garden which only needs maintenance at that moment that would be OIC uh that's something that I hope I believe we are aiming for to like some more of a final design with uh ethereum uh not allowing any more major changes only doing minor maintenance up updates so that's that's the ideal end game here um however still it's an infant Garden which uh never really ends uh but it would be beautiful if we get to this point of the aifi not really changing ethereum uh that would be it would be a beautiful world isn't it and italic would be happy we need to we need to make him happy guys so yeah if you are excited about this if you're excited about ethereum about all of the this progress happening uh the different clients development uh the the the research uh you are in the right place you are in the EPF study group where you can learn all about it uh but yeah what now what you can uh what you can do when you when you got here so I believe I provided a lot of resources uh right now you can uh see the the some resources from this presentation I believe all of them are in the uh uh in the Viki there uh in the week one Wiki and when I say resources I also mean just you know the pictures the charts that I show you and are uh explaining different parts of the clients you can start uh diving into these you can uh learn more about it find more relevant related connected documents and share them come to share them on Discord and uh update update them in the Viki so you can uh uh you can um uh find what makes you more interested what's your Niche what's your expertise uh and and contribute uh like yeah something that I didn't mention properly before so when you when you see this this monster of a machine all of these different uh uh roadm points or if you see the ethereum uh uh with so many parts here and the testing and so on there is not a single person person who would be expert in all of these there is not a single person who would uh who would be a developer of the whole ethereum but it's always been an expert in some of these parts or like in execution or consensus so uh that's what the modularity allows us to like focus on what do you like the more what do you find the most important exciting and you are maybe you have some PR some talents for so yeah uh focus on that find your Niche and contribute find these resources write more documentation and upload it to the Vicki open a PR learn how to use git learn how to actually uh contribute to the open source projects by uh contributing to our free open Vicki uh there are already some PRS open if you are here guys who opened it thank you so much uh it will be you know slowly reviewed you can see the PRS which has been opened and contribute there maybe give your add some content there or add your idea like okay maybe this should be written different ly Etc so yeah come to contribute um and yeah you can uh at this point we did like four EPF uh cohorts and uh you can you can uh find projects from all the previous cohorts uh in the same uh in the same gith up organization that we use for the for the Vicki so uh in the E protocol follows scord four core three uh in the SL project the project directory you see all of the projects that has been worked on uh by the previous follows and now what I what I mentioned in the beginning that this uh program is inspired by EPF is because these people spend the first weeks first months of the fellowship program learning about the protocol and about their Niche about their domain and if you read their projects if you read they write weekly updates you can see bi-weekly or weekly uh documents written by uh many talented people who've been learning about ethereum or different parts of it you can you can see their learning curve you can see what they've been working on and uh maybe find the resources that they are they are mentioning there bring them to the VII learn from them so we can follow follow their paths because that's why they created it that's why we kept it open public so you can all study it you can uh learn about their projects and uh their their Journey yeah and uh yeah I believe that's That's all folks uh there is uh that's that's the week one presentation uh right now there are already resources up for the week two uh next week we will be diving into the execution layer with FL CL but before that like please check out more of the resources from the week one and some pre-reading for the week two part and yeah it's been 90 minutes I used all of my time oh my God yeah uh so sorry to keep you up so long guys so let's take maybe five minutes for some questions if there are any and let's rrap it up uh yeah do we have any questions in the oh thank you so much Josh for sending all of the uh all of the all of the um uh links so here the the screenshot that this this monster of a markdown table that he uh just shared there uh it's uh um uh it's it's um uh all these weekly updates so you can see what a person in the SL project you can see what project the person has been working on and if you're interested like I don't know uh contribution to Lighthouse client uh creation of stealth addresses uh creation of new Beacon API so if you are interested in like different projects there you can see what's been the Journey of the person in this weekly updates yeah ah yeah awesome uh if there are any any questions any more comments feel free to go on I I I think you still can have some time and uh looks like there's a few people typing right now but not seeing any questions come through at the moment Al history of false it's the F missing before the O is there uh just just what I was was I was going on on about no I don't I don't I'm just I'm I'm I was just texting stowman like yesterday he's a friend of mine and I just uh I'm biased here I I uh push for the F part there um anyway um if there are no questions right now um the camera doesn't work um no it's horrible um anyway if there are no questions right now like please just feel free to you know ping me anytime in the Discord I'm I'm here for you guys um uh it's warning for me so the rest of the day I'm I'm I guess I'm uh ready to answer your questions feel free to put them in this thread or in a corresponding channel on the left you see the protocol VI Part with different parts of it so feel free to uh ask there if you have anything yeah yeah thank you so much for your time uh really appreciate that so many of you showed up uh I had to wake up for this so I really appreciate it I would like to wake up so or at least like have a public talk for so many people this early um awesome thanks so much Mario was a great talk once again and uh we definitely got it recorded uh this time so uh yeah for those of you who didn't catch it either of these two times you can see it uh on the recording and we will uh be back next Monday for uh the second week's content with L client on the execution layer yeah thank you guys see you on Monday 4 pm UTC I believe yeah uh awesome yeah yeah thanks thanks for staying up see you around of course see you [Music]