Episode 71
PoW vs PoS
About this episode
This week on Untangling Web3, we dive into a critical aspect of blockchain technology—consensus algorithms, focusing specifically on Proof-of-Work (PoW) and Proof-of-Stake (PoS).
These mechanisms are fundamental to ensuring decentralized networks agree on transactions and state without a central authority, and they lie at the heart of many discussions on blockchain scalability, security, and sustainability.
Key highlights
- Proof-of-Work (PoW) Overview. the original consensus algorithm used by Bitcoin. PoW relies on computational power to validate transactions, where miners solve complex mathematical puzzles. While it offers strong security through its energy-intensive nature, the environmental impact and low scalability remain major drawbacks. Despite these, Bitcoin’s PoW model has been battle-tested for over a decade, proving its resilience and decentralization.
- Proof-of-Stake (PoS) in Detail. PoS is highlighted as a greener alternative, used by Ethereum after its transition in 2022. Instead of energy consumption, PoS relies on validators who lock up tokens as a "stake" to participate in block validation. This system is more energy-efficient and scalable, but concerns remain about potential centralization due to the rich validators controlling the network. Ethereum’s PoS model has shown significant promise in terms of security and transaction speed.
- Comparing PoW and PoS. there are many advantages and disadvantages of PoW and PoS. PoW excels in security and decentralization but struggles with scalability and high energy use. PoS, on the other hand, offers improved efficiency and transaction throughput but faces criticism regarding potential centralization. The episode also touches on other variations, such as delegated PoS, and newer models like Proof-of-Authority and Proof-of-History, expanding the conversation on blockchain consensus mechanisms.
Both PoW and PoS have unique strengths and trade-offs, with PoW being the more secure but resource-intensive model, while PoS provides a more scalable, eco-friendly solution at the potential cost of decentralization.
As blockchain technology continues to evolve, these consensus mechanisms will play crucial roles in shaping the future of decentralized networks.
Transcript
Auto-generated from the episode audio.
welcome to another episode of the untangling web 3 podcast today we're exploring one of the most important aspects of blockchain technology or cryptocurrencies which is the consensus algorithms that underpin them so these mechanisms effectively enable decentralized networks to agree on state or effectively what's going on in The Ledges that basically describe them so this is how you add transactions how you kind of agree on things all without essential Authority which is kind of like the novel bit of all this stuff so they're ital in differentiating between various blockchain platforms and at the heart of discussions about scalability security sustainability all these types of things so it should be an interesting episode today how are you doing jack yeah I'm good I've been excited to talk about this one for quite a long time I'm surprised it's taking us this long actually to to do an episode on it because it is obviously one of the not most controversial topics but one of the most kind of debated topics in in all of blockchain and web
3 right this is so fundamental to how blockchains work so yeah I think it's it's good to finally you know lay the the groundwork for at least a few of them yeah I mean just to like kick this off I think when we think about you know what blockchains are the simplest way to think of them in my mind is they're effectively Excel sheets where we have you know all a list of transactions on the entire history of say Bitcoin or ethereum or whatever it is but the beauty of this in really simple terms is that many people hold copies of the same Excel sheet but the tricky thing is there if Jack you know myself and my mom all have the same copy of an Excel sheet and then someone else adds uh a new transaction to that Excel sheet how do we agree on how to add that transaction who adds it when do we add it when do we all confirm it sounds like really simple you know just save and it'll
upload and all that kind of stuff but when you're talking about you know big transactions high volume things moving quite quickly then it actually does get quite complicated and that is what consensus is right yeah exactly it's what's the point in a blockchain that stores all our you know our digital money all these assets nfts that may be used for all sorts of other applications you know data type applications if you know people are disputing well what's the actual historical record we talk about blockchain as almost like a universal source of Truth and I'd say truth in quotation marks because you know consistency doesn't mean truth but for the purposes of this conversation you know what is the ground truth the state as you said of the blockchain it's crucial because if you have you know we're talking about distributed networks if if different nodes think different things about the state of the chain then the network thinks Al has the money that I should have right in my transaction so yeah it's it's it's
a crucial aspect of blockchain so you know why don't we talk about the fundamental problem that people talk about often when we say consensus algorithms and that's the Byzantine generals problem right so this is really the starting point for Designing consensus algorithms I know that you're always talking about the Byzantine generals problem so I think this is your like you know your party topic so I you can kick off yeah and obviously we talked about this with Kurt as well recently so he's if you know if you haven't heard that episode he probably has done a better job than I'm about to do but essentially the Byzantine generals problem is is framed around this hypothetical group of Byzantine generals who are trying to decide whether or not to attack a nearby city or town or something like this and they are distributed so you can see the link already with blockchain they are distributed in nature so they you know one's attacking from the West one's attacking from the east north south Etc they're all
in their little encampments now they need a way to decide on whether or not they should perform this attack you know maybe it's a bad day for it maybe the bad there bad weather maybe it's more fortified than expected they need to work out how to how to communicate with one another such that they're either all going to attack or none of them are and that's basically key to their success so how do we ensure this consistent communication between these parties and you know you have potential issues you might have Bad actors you might have you know generals that don't want to attack maybe they think this is doomed to failure I'm going to leave my Garrison back and I'm going to let my people live uh but the others can go ahead but the problem is if they don't all attack they'll fail right so there's kind of two requirements to solve this problem so one of them is consistency and that's all the loyal generals so all the good actors should agree on
the same decision and if all the loyal generals again all the good actors receive the same information then they should all come to the same agreement basically so they should all reach the decision they shouldn't have you know if they all get the same information they should all all kind of receive they should all get the same opinion on whether to attack or not that's the basic framing of this problem exactly and you know how we like to make things relatable to people at home all the Byzantine generals out there who are listening you've got a solution now to solve your your method of attacks no I know this kind of is the way that is described because people have been talking about this for hundreds and not thousands of years right I think that one of the things to make it a bit more relatable to people today is in that example of us all having you know the individual Excel sheet and we're all trying to agree on what transactions Who's got what
money has Jack paid me have I paid Jack whatever is there's this big problem we always talk about with currency and sending payments around double spending if I have you know this 20 pound note physically say digitally and then I give it to Jack I say Jack here's 20 pound note for the good and then I also try and give it to my mom here has 20 pound note and I give it at the exact same time how do we prove that you know Jack has the 20 pound note or my mom has a 20 p no because there's this kind of I guess if it takes a long time to start processing things there'll be this time I'll receive two sets of goods and both people will think they've got the 20 pound before actual confirmation say you have got the 20 pound so that's one way to make it relatable there double spam problem and how to avoid fraudulent transactions yeah exactly the equivalent of attacking the city is the state of all
the accounts in the Excel sheet right because you know if they're all wrong we're going to have a very bad time all our money is going to be all over the shop so it's kind of equivalent to to to failure scenario in the Byzantine General's example if you are just tuning in yet this is now a history podcast um no just kidding I think like fundamentally what this comes down to is trust right you said truth earlier and I kind of think of it as trust everyone talks about blockchains as being like trustless systems and we kind of defined it what am I trying to say here no I'm not going to go down that Railway thing I talked before too far too confusing is that these These are more trustworthy systems and the kind of the trust changes where the trust means that things are more visible and when we're talking about people trying to interact with a central Authority it's actually really really tricky to manage this trust manage everything like in a
kind of decentralized way so yeah that's the point of these consensus algorithms they maintain trust in a kind of a trustless environment and I guess as of 2023 there are over 10,000 active cryptocurrencies and many I think there's hundreds of different consensus mechanisms so trying to cve through these to work out which ones you can trust and which ones you can't is an absolute nightmare but we're going to go through some of the leading ones today and try and outline what they look like what the pros are what the on how they compare yeah exactly and it's worth noting as well that these consensus algorithms are the subject of like really heated academic discussion not just kind of engineering technical stuff from the people building blockchains but you know there are some serious academics who have proposed these different consensus algorithms over the years and that also feeds into it but yeah as you said there are so many different ones out there typically they'll follow the naming convention proof of something and the reason
for that is the kind of the OG the original one as we always say goes back to bitcoin which came out with its own algorithm that was really touted as the first practical solution to the Byzantine General's problem in the context of kind of digital ledgers digital Excel sheets and that was proof of work right so this was the one of the key innovations that spawned Bitcoin from satosi yeah I mean that happened back in 2008 right and you kind of said that that there were things going around before that but this was one of the Practical examples where people started to get buying and this has been going now for what over 14 years 16 years now and it's probably the most battl tested well-known consensus algorithm going I don't think there's been any downtime or any like major malicious attacks that have been successful in that time and Bitcoin obviously manages trillions of dollars and people are now institutionally buying into this factor and starting to trust this system so should we try
and outline how proof of work actually works let's go for it I mean this is this is one where we might have a bit of debate on how to describe it I know again we did with Kurt previously so essentially with proof of work the whole idea is how do we ensure this consistent communication this consistent signal between our nodes on the network and satoshi's idea was to attach a costly signal or or a computational cost to this message to the idea of adding new blocks so if you want to add transactions to the blockchain to your spreadsheet then you need to have this additional uh computational cost attached to that message to prove that some work has been done to sync some work into doing this right so sometimes this is called a mathematical puzzle I don't love to call it that but it is illustrative of what it is so essentially it's like a guess and check system so you're going to repeat this hash this computation many many times so you're going
to take this new set of transactions that you want to add to the Ledger you're going to perform this hash function over it where you're going to basically compute this function this very simple dumb function essentially that spits out a random number and you're going to check if that random number meets a requirement and we can get into what the actual requirement is but you're going to do that over and over and over and over and over again until you finally hit one that meets the requirement and then if it does that's the condition then you can say great I can go and communicate out to the network to my generals of Bitcoin nodes and say here I've got a solution I can add a block to the network and then they can verify this computation that you've done very easily as well yeah I think you described it very nicely there is you basically I mean a lot of them are relyant on this hash function the stupid function that you've talked about where
you can throw in data and you get a unique unpredictable representation of that input data so the data that's being inputed is like we talked about the transaction data some other data basically you get a unique representation of that transaction data so what all the transactions that are in that block the updates to The Ledger we talked about in the Excel version of the file and then something called Annon which is effectively a random number that you try and get the unique representation post has of all that data below a certain threshold and we don't really need to talk too much about you know what the threshold is but it's just you're trying to get a unique representation that's below a certain number effectively and that's kind of how the work is computed and like Jack said the kind of the big Paradigm Shift here is you know previously when we had Central authorities like Banks or something like that managing the transactions and they would do everything you're trusted in the bank you trust
that their single version of the Excel file is the kind of the Holy Grail and everyone goes to that single version whereas now with Bitcoin we're trusting in the hash power the people that are Computing this mass mass amount of hashes to do these kind of transactions um you're trusting in their hash rather than their physical identities but I mean you could say that you also know who they are as well and there is some identity based stuff in there yeah and I think there's a couple of things to point out about this dumb hash function as well right so as you said this unique representation sometimes called a digital fingerprint so it's like this uniquely identifies the input data I.E the block you're trying to produce um it's also effectively random as well so how do you repeat this over and over again as you said you take the the the data for the block that's kind of static and then you add this nons in that's the bit that differentiates this hash and
then you can you know try it once doesn't get the right answer then we change that nons that number only used once and you just repeat that process over and over and over again so it effectively produces a random new output that may or may not meet that criteria every single time and the other kind of CR cral part then is that okay what does that really solve well this needs to happen in conjunction with the difficulty adjustment algorithm right so this is the part that makes it competitive in nature right that means that you know everyone is competing to try and find solution and the difficulty adjustment algorithm is constantly making sure that there's a sufficient amount of energy or computation that has to happen to send one of these signals because if it was really easy then it wouldn't really solve anything right it would just be you know we can all do it instantly and we'd all come to this same problem of well everyone's got a solution they're all different we're
not in agreement anymore right exactly so the point is the entire network all these many miners they get say similar transaction history once they've got their transactions they start putting them into a block and that's effectively trying to get compute this value of the hash with with this random number to get it below a certain threshold as soon as they've done it like their computation with the random number is below that threshold they say bingo I've done it I've solved it first and they send it out to the other minor saying look I've got a a hash value now with the the random number and the transaction data there is below the threshold I'm the first to solve it give me the money and the one of the the kind of the key points here is it's very difficult to do but it's very easy to verify so very quickly other all the other miners who are still trying to compute this block and say oh yeah verification done they have done it they get
the rewards we're now going to build on top of their block and start propagating the network further or you can say actually no you've done it wrong and I can verify that very quickly as well so I'm going to keep trying to mine that block so this kind of brings in the incentive part as well where whoever does it first they get Bitcoins we've kind of done previous episode about how many Bitcoins they get and the number going down every few years um but yeah that's how the reward mechanism comes into this as well yeah number go down which is not not the best thing in crypto normally but actually it's the thing that uh underpins it all but yeah that is really crucial what you said there you sometimes call these trapdoor functions and the hash function that you're Computing is one of these trap doors so it's hard to compute this problem is hard to solve to find a valid solution that satisfies the criteria but yeah once you found it you have
to all you have to do is tell everyone what inputs you use and they can verify instantly they only have to compute the hash once whereas you may have done it you know billions of times uh in order to get your solution so that's kind of the key there and then yeah this block reward is the other part so we say all the time how Bitcoin combined you know computer science and economics and Game Theory into one thing and this is you know in the proof of work consensus algorithm that's where it's most obvious I think because you have the computational thing you have the trapo function that you're you're hashing over and over again but then you have the game theory the incentive side of it where everyone's competing to do it first and then the final little piece of the jigsaw puzzle is the fact that you have the difficulty adjustment algorithm that's making it sufficiently hard all the time so you know it's not just we're going to be running into problems
all the time yeah and we're talking about like the amount of work you have to put in this difficulty algorithm that's basically forcing miners to do lots of high energy task to produce these blocks that says this is you know the new state of the network these are the transactions have been validated and that's one of the main parts of the security mechanism is this is very high energy and there's High equipment costs that would determ malous actors to do this um and another big component of this is you know once I've built my block I when I'm building the the next block I actually build on top of that block and what we kind of I don't want to go into too much detail on this but you're effectively having a hash the unique hash of the previous block built into your current blocks hash as well and that in this way creates this Block Chain this chain of blocks and the point here is if I'm doing lots of work in Block one
to create that and then I do lots of work in Block two to create that and then lots of work in block three obviously when I'm like linking these together effectively if I want to change a transaction in Block one and we're now at block four I have to recompute do all that high intensity high energy stuff for all those subsequent blocks as well for them to be valid like we said that's very to do very easy to verify so when we kind of talk about Bitcoin we have this previous idea around you know your transaction's not safe until you've got you know 10 Bo blocks that are buil built on top of this that used to be under this kind of assumption that it's so high energy intensive the further back you go that it's almost in physically well i' say very impractical to do it don't know if that's true anymore because actually the amount of energy that goes into even the previous block is so high these days that it'd be very
difficult to do that anyway but yeah that's one of the assumptions that's part of this yeah that's a really good point right because is going back to the original problem that we're solving when you talk about Bitcoin being secure using proof of work it's we talk about it in terms of uh 51% secure or it's it's secure against 51% attack so it basically says if you have more than half of the nodes on the network honest so more than half of my generals are good generals that are acting in in good faith and honestly then see the Napoleon bonapart part of you there are they good generals are they acting honestly yeah not let's not get into all that because that's going to be a huge rabbit hole but yeah the condition being the network will be secure if you have at least half of the half of the field of of your generals and your participants are honest and as you said it's not just a single block you're building up the work throughout
successive blocks so you can actually measure um how much work is behind a chain instead of just a single block um and and the deeper that chain the more work that has to be done to overcome it right so the idea of the 51 % security being so long as more than half of the the the generals are good then they can outpace any attackers so if any attacker tries to you know build an alternative chain well so long as they don't have half of that hashing power or more then the honest people will always together be able to build on on on on the main chain and you can actually they talk about this in terms of following the heaviest chain so the heaviest chain being the one that has the most work attached to it the most compute uh that's been done but it's it's kind of interesting because when you look at you know how much hash power has grown over the whole lifespan of Bitcoin those early blocks are nothing right
you could actually speedrun the first you know few hundred thousand blocks of Bitcoin with today's equipment relatively quickly but it's that last few thousand or last 10,000 100 thousand blocks that where it becomes really really heavy because each of those blocks has more work associated with it than you know many many of the early blocks of Bitcoin so that's kind of interesting as well to think of you could try and Forge an old part of the chain but you couldn't get very far with it yeah like you said there let some the obvious Pro that you just identified is the proof and security Bitcoin and proof work has been going there for 16 years and there's been you know that's over a decade of secure operation like what is it called almost zero downtime no malicious hacks no kind of sustained periods of people trying to actually you know produce these bad malicious blocks so that's a big Pro like kind of Kurt described as completely bomb proof or I've heard a few people describe
as completely bombproof if there's a nuclear disaster and everything goes wrong Bitcoin will still keep going and no one will be able hack into Bitcoin after that or something like that and we actually talked a little bit about the quantum episode um actually no we're not going to go to that rabbit hole but if you are interested into knowing how some ways in which Bitcoin could be vulnerable maybe go back to our Quantum episode because that's a whole rabbit hole that we're not going to go down right now yeah rabbit holes on rabbit holes we're kind of going into some other Wormhole or something now but um yeah and and proof of work there have actually been I mentioned the academic credentials of proof of work and other consensus algorithms there have been some academic attacks on proofer work saying well in theory uh it could be vulnerable to this kind of attack there's a kind of selfish mining attack that's mentioned as well but in practice as you said there's never really been any
issue of this nature there's never and there's never and there's such an incentive to attack the network that you also have to remember so that that huge uptime is you know proves it's very secure but coupled with the fact that of how many assets and the value of those assets that are secured by Bitcoin the incentive to attack it is so so high that also kind of proves in practice how secure it is um so yeah we haven't really seen anything else like it the other kind of I guess benefits of this are I always think of it as like decentralization of the network is more an emergent property of this consensus algorithm than it is a design goal necessarily so because it's competitive it means you're always going to have multiple miners always competing right there's so much money on the table to collect with this subsidy the 6.25 currently bitco Coins Plus the fees that you're always going to have people trying to compete um and that's kind of made it you know
secure in practice right that's why you're never going to have it all cessing into one because then the whole value of the system will be slightly undermined right so that's another big Pro in in the proof of work box in my in my book yeah one of the things I've read as well is that because the margins are so tight between these people these miners competing to M the next block is they're always trying to improve their equipment always trying to reduce their latency always trying to improve their connections in the network I mean that's in theory I guess to try and mine the the blocks as quickly as possible and get the rewards but then in practice when you look Al at the performance criteria of Bitcoin outside of just the security stuff just about the not just the energy required to make these blocks in terms like you know throughput scalability blah blah blah all the things that we kind of you know actually compare blockchains with numbers like it's not doing that
well right it's still this kind of slow lagard that does the job very well super secure but you know isn't kind of hitting the throughput that we're seeing on some of the other chains so that's that's one of the you know pro/con the kind of the obvious con that is always thrown in the media is obviously the amount of energy consumption that goes into this and this is kind of it's been compared to the energy consumption of Argentina the energy consumption of Sweden blah blah blah all these kind of massive numbers that are always out there and like we always say on this energy consumption in itself isn't a bad thing but it's energy for the utility that it produces and there's always questions around you know what utility does Bitcoin actually provide for society right now and that's still a question mark in my mind it's not the pure idea of consumption that's the issue it's the what do you get as a result right and for me it's also a bit of an
open question as to how much energy do you need to make it secure it's because it's this Dynamic thing the difficulty adjustment algorithm is constantly changing how much uh hash power how much compute power how much energy by proxy is actually needed you know there's not necessarily anything to say that every everyone could collectively reduce their energy and it still be secure but it's then if some other malicious bad General actor pops up with loads of energy then it might look um insecure for a short while right so it's the security is a very Dynamic thing in Bitcoin and that's why I think it's hard to to P pin the finger on exactly you know how much energy do we need should we be consuming it but you're right like you're completely right it's if if if it's proving itself via utility no one would be asking the question right if this was being used daily as a monetary system if you were making payments in it day in day out people would not be
questioning it because you know it's on it's on a a reasonable order of magnitude of energy to other Global Systems except right now it isn't it isn't achieving that so yeah it's that's probably the the biggest sticking point it's got just on the point because we kind of have to get both sides right like on the energy debate like it is using lots of energy the one thing I would just want to say is that it's something like over 40% I think it's actually over 50% and I've seen some stats around over 60% of Bitcoin mining does come from renewable energy sources and like you know the hardcore environmentalist will say well that's renewable energy that could should be going to something more useful and the rort to that is typically well you know the point of Bitcoin is we can shove one of these mining rigs in the middle of the Arctic where you know this energy is kind of completely isolated and there's no real practical use for outside of these Bitcoin mining
rigs um but that's just to give both sides of the argument not sure how fall on that but yeah a lot of this energy being used is renewable based yeah I 100% And I I for the record I think it should all be I have always maintained the position that I think Bitcoin is the first technology that is truly self- incentivizing people to use green energy right because it has an Associated upfront cost to build you know say your own renewable energy infrastructure it has a Bitcoin mind and because again we're talking about big big institutions now like lots of public companies or or companies of that kind of magnitude that are these mining operations these mining pools so it's not in feasible they can go and build their own uh electricity production things we've seen in Al Salvador them you know looking into collocating mining rigs with geothermal energy right which makes perfect sense and because once you've made that initial investment you're you're being paid for all that energy you produce via the
Bitcoin rewards system then it's kind of a direct incentive to use it and as you said it could be used for energy that's otherwise going to that can't be transported from from different sites but yeah it's it's always going to be a big it's always going to be a big talking point when you got people like Green Peace threatening to to to kind of Sue people and uh and lobbying lobbying for change right because that happened I think last year or the year before they were lobbying to change the consensus algorithm so yeah it's there are some big enemies of it right out there yeah so in that you just talked about mining pools as well that's one of the big kind of cons I would say around Bitcoin is because to compete you do need these really high SP bits of equipment like you can't just you know spin out your laptop and try and like start mining Bitcoin they are very specialist large bits of kit that are usually based in like say
random areas that are kind of low energy cost screen related hopefully like Alaska or something like that um you kind of have these pulling operators the point is that people will contribute their energy to one mining pool and then effectively like distribute the hashing power so rather than one PC doing everything they'll kind of like Shard it or distribute it or parallelize it so you can have a million different people with their tiny little laptops all contributing to one mining pool that will then try and get the difficulty of the block being mined under a certain value and then they share the rewards afterwards the downside of that is like on paper you could say I've got a million people all actually mining this network but in reality you've got one orchestrated there that's kind of a centralization point and we are seeing that even though online you might see the Bitcoin core has like you know thousands of different mining nodes really I think it's like actually run by say up to 10 kind
of actual mining pools and it's like yeah there's a centralization potentially there yeah and the only thing I'd add to that because I agree it's you know whether or not this that level of centralization is a problem is another question I think but for those that care deeply about a high deegree decentralization it's clearly an issue and it's partly because those mining pool operators they control the contents of the block so the hashing operators they just get to contribute energy they don't really control what goes into the block right so you have kind of single decision maker um my maybe my my final point in defense of proof of work which you know I think is is fair enough is that it's just it's just very elegant isn't it compared to some of the other Solutions out there right which we'll get on to I think it's it's so simple and it's it's it's much easier to understand and you can see very clearly how it solved this problem and and that's borne out by
the popularity of Bitcoin how much it's grown and the fact that you know 15 years on it still has this kind of robustness but you know with that in mind let's move on then to the other proof of X right proof of what uh what's the next one we should do I think I think I know you which one you're going to want to cover Al I think I think I know um it's proof of stake right when you look at like say the top 50 top 100 blockchains the vast majority of them are now proof of stake which when I checked before this episode I was kind of Blown Away by because almost all of them are proof of work like 10 years ago and so proof of stake obviously you know this is a big Landmark event that ethereum's been talking about for I don't know like six years how long did they were they proposing the merge between like the kind of their their test um test net version of the proof
of stake ethereum and then the main net version of proof of work ethereum like this has been going on for absolutely ages but the point is now ethereum is proof of stake based um ethereum as you know is kind of like the second biggest by market cap blockchain hosts thousands of decentralized applications has probably the the biggest developer Community by far so ethereum is like the top dog in a lot of ways when we talk about the growth utility all these kind of words um ethereum has what around 14,000 nodes and they're running one million validators so how the hell is that possible yeah it's it's it's an interesting one and you know in the run up to the the merge which is in 2022 I think around September there were there was this long period of people basically adding themselves as eligible to become a validator in this new model right so that you had to lock up I can't remember how many it was it was something like is it like 16 eth
or something I'd have to double check but it was something it was something not insignificant as an amount of money to lock up to become a validator for this this new um this new consensus algorithm for the network when it shifted and you know one of the most notable things for me is that the fact that Pro ethereum was able to migrate from proof the work to proof the state like that in of itself was huge technical achievement the fact that didn't really go wrong for all intensive purposes I think is amazing but yeah we should definitely get into Pro State actually works so what's the current state of of you know how things are validate in ethereum so I'm going to I'm going to pass this one to you Alec because I did the first one so now you can have a chot you did the very elegant one that you fancy I think um so yeah Jack talked about you know the point of decentralized networks we're trying to agree on consensus what
the state is when you have a centralized Point obviously they can make all the decisions but in Bitcoin we kind of trust miners ability to kind of hash which is effectively the amount of energy they consume whereas when we look at ethereum and proof of stake what you're trusting is the amount of stake they put into the networks the amount of money so Jack said I think the actual cap currently is to be a validator on ethereum you have to have 32 eth and that's what 32 eth it's about three grand right now so around 100K let's say so you will have 100K of the network tokens that are locked up and that is your say basically your input to the network to say I am trusted I am part of this if you know if the ship goes down I go down with it so that's their way of saying okay you are securely part of this network and you wouldn't want to be a malicious actor um the protocol then goes around these
you know one million validators and will randomly select every few seconds the ethereum will effectively randomly select one of those 1 million validators with a 32 e locked up and ask them to validate the next block now rather than doing this high-intensity energy process that we talked about proof of work the validation here is effectively I will just validate the transactions I will look through them kind of in like a centralized way at that point of time and say is this right do they make sense I'll mine them effectively um and then everyone else will check my work and you know the more in the kind of this system the more tokens you stake obviously the higher chance that you are selected because it's this random selection based on all the stakes um and this you know this randomization the minimum staking requirements of 32 e is meant to prevent any single entity kind of dominating the network yeah that's a lovely lovely sucin description Al like it was just not easy to do because
you know the first thing I would note about it is that prove a stake you you know surprise surprise I don't think it's quite as elegant in a way because it has a few more moving parts right it has has your block proposer and then your your validators which are kind of both being selected in proportion to that stake and you know that's the equivalent of in proof of work where we had your probability of mining a block is proportional to how much of this compute power you have how much this Hardware you you've amassed whether it's on your own or in a pool in ethereum it's how much money you've got logged in right so the you don't have to have a massive uh data center operation be involved you just have to be rich right you just have to have a lot of tokens either you bought ethereum when it was really cheap or you have enough of the the Fiat currencies of the world to buy enough to stake now and then
yeah as you said there are there are lots of other elements involved with securing proof of state right so you know there are some issues associated with it maybe we'll get into you know what makes it secure um in terms of the Byzantine General's problem it's kind of analogous to proof of work though it's it's doing a similar job by saying well so long as we have a a majority of the money being honest a majority of the stake locked up being honest then it should be secure but there are some issues because you don't have that same immutability type property with Bitcoin because there is no energetic cost now to faking a chain you can create an alternative chain without any energy cost right you just have to sign a different version of things which you could do quickly if you had enough money yeah I think um one of the big points here is kind of related to the malicious actor versus say if one of the five Bitcoin nodes drops off the
network you still have the other four constantly competing and if say I drop off it doesn't matter because four other people are going to mine that block almost as quickly um whereas in ethereum obviously because we have this random selection I can lock up my my stake my ethereum stake and if I suddenly drop off the network and I'm randomly chosen as the person who's proposing the next block well you know the block doesn't get mined effectively and they obviously have this kind of backup program if you don't respond in a couple of seconds or whatever it is then you it'll go to someone else and someone else will be proposing they'll mind the block but they also have this idea now because you have this stake locked up if you don't say hit the minimum requirements of I will be online all the time I will mine the block within this time then they start to slash your stake they'll say okay well you're not a very good validator you haven't hit at teas
and C's so now I'm going to slash your stake by 50% all the way up to 100% And this also kind of fits in with the malicious actor thing you're talking about if say someone accidentally or purposefully doesn't really matter validates a wrong transaction somehow or tries to mine a block that isn't valid then that stake will be invalidated as well and this is kind of done in the separate stage you talked about we have deposer then we have a an attest that's randomly selected then we have a like a finality thing which is lots of people all testing that the block is is done so it is much more complicated than Bitcoin um but that's how they kind of have this redundancy in checks but this is enabled basically with the biggest cryptographic developer community in the world so it's incredibly complex but it also means that they really do have like you know hundreds of thousands if not millions of different validators all properly contributing to a network it is much harder to
describe because there's so many different layers but they are advancing so quickly and is quite impressive what they're doing yeah 100% there there's very good reason for it being widely adopted you know apart from the fact that that that lack of energy requirement means that it's much more attractive for you know big business and governments to start building on right because you don't have that question or worry over how much energy it's consuming um there are then as you said questions over the security model you have to introduce slashing to stop bad behavior or uh maybe not even malicious behavior but but just being idle right and not actually being uh available when you need to be to validate propose Etc on the network it's it's a kind of much more akin to just putting your money in the bank and getting a return on it right because we should mention with Prov stake typically you're putting this token in and you're getting rewarded in that same token for doing so it's almost like earning
interest on the state money that you have in the state tokens um and then when you add in these different layers you know typically when we talk about security and computer science the more complex it is the bigger the attack surface the more unknown unknowns you might be introducing to to have to worry about in the security model and we also know that as with all blockchains you know proof of State blockchains like ethereum and others are also heavily investing in these Layer Two scaling Solutions which are adding more more complexity more layers to that problem now and even though there hasn't been kind of I would say major breaches of proof of stake that we know of I think I've seen more kind of academic theories about how they could be breached um so yeah I think it's it's worth noting again that that complexity is potentially a worry but we haven't seen it in practice too much yet yeah I I would agree with that I think I say bitcoin's been around for
16 years now but I think like the number of chains that are now deploying proof of State you kind of have like them in parallel all kind of testing it out in different ways you are right it's incredibly complex but it does mean that they have a a kind of um well it's not just because of say the consensus algorithm it's also because of all the things they're doing on top of that consensus algorithm that it is incredibly complex like the ethereum virtual machine that basically allows programmability like there's lots of more complex things that also feed into why it's so complex right it can't just be done in the way that bigcoin is but it does have faster block times it does have higher TPS I would say there are lower bar to entry which is good you are right it is effectively money in Money out but you know this the world runs on money fortunately or unfortunately so it kind of is the thing that we're all used to it does make
sense and let you say the big thing when people think Bitcoin is high energy slow but very secure and ethereum is a bit more versatile but yeah bit quicker as well maybe yeah that's why we kind of people refer to bitcoin is this like super secure settlement layer and ethereum is more like a smart contract platform right it's more active in its processing um I think it's worth also noting the differences between proof of work and proof of stake in terms of that centralization risk you mentioned because where in Bitcoin and proof of work you tend to have maybe moving towards a small number of these pools or companies that act as the decision makers effectively right maybe it's around 10 in Bitcoin they're much more visible I think than in proof of stake system because again in either Bitcoin or ethereum there's no identity system natively that you don't need to identify yourself to participate in mining you never have done in in in mining proof of work or being a validated and proof
of stake but if you're a proof of work Miner it's much harder to hide right you have to have one of these huge data centers you have to you have a pretty big electricity bill right it's harder to hide and they're much more easily identifiable with ethereum all you need is to associate this much this amount of ethereum with your wallet address and then your validator right how do we know that say the top 10 wallets in ethereum aren't controlled by the same entity under Underneath It All right I think that's one of the bigger concerns I have with it in general yeah I would I would agree with that point I think that maybe pooling kind of undermines that argument a little bit because like you say if a million people are all contributing their laptops all around the world yeah you still need a data center to process transactions and manage all that but it's much smaller than say the assumption that it's one Giant Center with all the as6 in the world
basically mining this stuff I would say that you know in the same way that Banks also don't have high computation pair but they have lots of people that are servicing the the moving of assets I imagine that pooling Solutions as well also probably have that you know degree of trust I'm not just going to throw my 32 eth to anyone I'm probably going to prefer you know someone that's got a bit of credibility behind them yeah I do take your point like the hardware does make it easier to track people than just money but I say the the world runs on money but the big thing here is like you know we're kind of questioning the utility of Bitcoin and everyone says hold the value blah blah blah no one can question the utility of ethereum like it is the place to build apps it is the place where you know there 15,000 you know active developers constantly on the project like it is the place where everything's happening everyone who's anyone I'm not gonna
say not anyone everyone is anyone Bigg a statement but you the vast majority of people do build an evm they do build on solidity know they know about the programmability that's what's being taught in universities like there's just so much happening there it's like it's such a you know good liess check yeah it's kind of a weird situation isn't it where proof of work underpins the biggest most well-known blockchain but proof of stake is like the standard for the industry right so you have this this very strange situation where you have the one outlier with all the history that has this proof of work really but then everything else has kind of almost moved on and gone to the Next Generation so yeah it's interesting to see how that will play out in the future um you know I think we've covered nicely the main properties then of both proof of work proof of stake and and compared them a little bit shall we have a kind of quick honorable mentions of some other uh
versions of these or other consensus algorithms just to round things off yeah let's do it another one which is very very closely linked to proof of stake is the idea of delegated proof of stake so the only minor adjustment here is you know token holders rather than individually validating themselves can kind of offset that or delegate that kind of um that power to one person so the point here is you kind of rather than having a million validators which is like we said extremely complicated to manage all those validators and all their inputs and randomly select them whatnot you could each of you one million people all you vote for say one person out of a 100 you decrease the validators the number of validators which makes it much easier to actually manage the network manage the blocks manage all the interactions you still have the same amount of stake that's actually being staked in there but much easier to manage so that's that's one you know slight minor modification on proof stake is the
idea of delegated proof stake exactly so that is yeah one of the major variants that we're seeing more and more and it kind of gives a bit more uh f exibility generally to to to prove your stake and I think that one's you know gained a lot of popularity in the last couple of years even though ethereum's only just moved to proof of stake other things have started natively with with Depot the main reason there is because it when you're talking about proof of stake networks or stake based networks is the amount of stake you have that's actually being you know staked as part of the validation is effectively the security profile like the more stake you have on there you you assume the more secure the protocol so some big malicious actor from North Kore career can't just come in and say you know I've got enough money now to overthrow and 50 over 51% so like it's important to have lots of stake but lots of stake when you have like a low
barrier to entry like ethereum means you have to coordinate so many different people so many different validators so the main point of delegated proof stake is you still have the same amount of stake but you're kind of minimizing the number of people and they all vote for say one in 100 people to actually kind of manage the transactions which makes it much easier from an operational perspective and you know they tend to have you know higher transaction through but um but maybe there's a risk of centralization there and I think this is very analogous to the mining pools in proof of work right that we said so it's like people pulling resources together except in that case it's computational resources here it's kind of financial token resources um and it's much more liquid again there's variance of Depot now right but kind of liquid proof of stake as well where you can much more easily change who your State against right so the person you've delegated your state to can you can switch up much
more easily if something goes wrong or if you don't want to a bit like the slashing we said before so yeah that's that's one of the big major Alternatives another one would be proof of authority which we've spoken about a little bit on the show before you know things like vchain Tomo chain these all use proof of authority and that's kind of almost bridging that Gap or that that issue I mentioned with you know who is actually controlling this stake under the hood you have much more kind of a a sense of known identity and verification in the valid dat set proof of authority right so these are more likely to be kind of reputable known entities that have to be kind of permissioned in to the network as I understand it um and it's quite Limited in number so it's almost like a hybrid of of all those models we've mentioned but it has kind of some similar properties right you have these assigned turns between validators you have very quick consensus typically and
a lot of the you know part of the security me me part of the security mechanism is actually also based on like reputation risk because those validators that come into the network they have a reputation at stake right as well so in addition to you know any other at stake just a position there let's not get confused um yeah you are right so like um the proof of authority is kind of like in my mind a a web 2.5 kind of approach where you still have these somewhat known entities that are kind of doing the validation or the mining or whatever you want to call it but they're also doing in a kind of more transparent way so they're all managing their own versions of it but you know who those parties are so I imagine like it's much more Enterprise friendly because you know who's the you know the that you said there the ethereum validator down the road I don't know who they are but I know who a PWC is right or
I know who a you know big company Walmart is or something like that so it's somewhere it's somewhere in between my understanding is obviously vchain as well they also have a staking mechanism so you are like signed off individual so there's the proof of authority where your your reputation's at risk everyone knows you you have to get signed off but you also have a stake in there as well so you wouldn't you know sudden get signed off and then just start burning whatever the vchain tokens are for that um so yeah there's um you say probably bit more friendly towards Enterprises higher throughput Less on the decentralization risk because you have to sign off and you have these authorities that are in between but and there's maybe regulatory risks around that yeah exactly I think you you kind of touched on the that that good old Chestnut the blockchain scalability tra there as well right like we always talk about well do you sacrifice security or decentralization or scalability and each of these consensus algorithms
as we've kind of mentioned throughout they they tend to have a slightly different profile of each of those things you also mentioned burning there right proof of burn is another consensus algorithm that's kind of weaved its way into to to the Consciousness now as well where you're kind of doing something similar to what we described but you're actually burning tokens when you when you validate blocks and there's actually an element of that in ethereum now as well right because fees get burned when you uh when you mind block so that's an interesting one another one is proof of History which you know salana has risen to become one of the major changes over the last few years and salana is interesting because it does take quite a different approach like its consensus mechanism proof of History has these things called verifiable delay functions so where Bitcoin had this this computational puzzle let's call it again um there's a there's a slightly different cryptographic puzzle or technique used where it's actually got a concept of time
so solving for the next block in in in in the uh that creates the blockchain will be a kind of standard amount of time away and this proof of History uh builds up using these special verifi viable delay functions where you know if there's a solution there's been x amount of time added on so it's very interesting but again one I probably know a bit less about as as a purist of proof of work yeah and I think everyone Sal everyone knows about salana they're achieving like absolutely crazy transaction through like 65,000 transactions per second which is obscene numbers actually starting to compete with like traditional kind of Fiat systems which is a bit Bonkers when you think about how low the numbers are for bit Bitcoin and ethereum but you know like you said it's very complex I think salana is also known for a lot of network outages isn't good um and I said it was something that we're not going to dive into in any depth in this episode because it's a
bit too complex for us in this episode we've talked about you know consensus algorithms why they're needed why they're so important and some of the main differences between the big ones proof of work proof of stake proof of authority proof of history and delegated proof of stake the various versions of proof of stake um and like I say there's a lot of pros and cons with each of them but we've got a remember as well that these are experiments at the end of the day like some of these have only been around for less than 10 years so we really are working out you know what the security profiles are what the scalability profiles are how they affect decentralization but you know the golden kind of thing we have to keep in mind is the idea of the blockchain trilemma that Jack has talked about multiple times now that you can only choose two of the three of scalability security and decentralization so we are starting to see these blockchains actually position themselves in certain
areas where some are known for scalability some are known for security and someone known for decentralization but yeah there's lots to do like I say there's lots of things we could talk about like Advanced zkps and layer twos and all the things that are building on consensus algorithms but maybe that's for another episode so I just want to thank you all for listening and thank you wherever you may be and join us next time as we untangle a little more of web 3 thanks for tuning in to another episode of untangling web 3 don't forget to send us your thoughts questions and comments on social media and be sure to follow us on your favorite podcast provider to catch the next episode see you next time to untangle a little bit more of web 3 the views we express here are our own and do not reflect the views of our
Topics: Blockchain & Crypto