- Fastest Cryptocurrencies: Blockchain Speed 101
CRYPTO 101 GUIDE
Transaction per second speed is getting one of the most important parameters among cryptocurrencies as same as fast-growing internet speeds and technologies.
The ultimate end game of cryptocurrency is to replace the corrupt and inefficient financial systems we have today. So far, we’ve seen amazing progress towards this goal. Decentralized payments, saving, lending, borrowing, and even stock investing can be done directly peer-to-peer without a middleman, and without KYC documentation.
The decentralized applications that provide these services collectively hold over 100 billion dollars in cryptocurrency, and their TVL (Total Locked Value) continues to grow by the day.
There’s just one thing that’s holding cryptocurrency back from true mainstream adoption, and that is speed. Most cryptocurrencies are just too slow to support the billions of transactions made by billions of people every day.
Even though lots of cryptocurrency projects have made a habit of bragging about how many transactions per second they can handle, the numbers they provide rarely stand up to scrutiny. That’s why today, I’m going to reveal which cryptocurrencies are really the fastest and why.
Disclaimer: All the content converted from Coin Bureau’s “Fastest Cryptocurrencies!! Blockchain Speed 101!” video after getting the whole permissions
Unless you’re completely new to cryptocurrency, you’ve probably heard of the blockchain trilemma. The blockchain trilemma was conceptualized by Ethereum founder Vitalik Buterim in 2016 and it states that a cryptocurrency can only be two of the following three things, scalable, decentralized, and/or secure.
Scalability refers to a cryptocurrency's ability to maintain its performance, even with millions of users and transactions. For the purposes of this post, let’s just say that scalability is synonymous with speed.
Decentralization is where you have hundreds or even millions of unrelated computers around the world verifying and processing cryptocurrency transactions. If you’re wondering why decentralization is important, I’ll just show you this headline from 2018, when the 21 validators on EOS froze 7 wallets on its blockchain.
Security is pretty self-explanatory. This includes stuff like a cryptocurrency consensus mechanism and the encryption of its code. As I understand why you can’t have all three, imagine that you have a cryptocurrency that can handle a thousand transactions per second. This checks the box for scalability.
Let’s say this cryptocurrency also has 1,000 independently run computers connected to its network processing transactions, this checks the box for decentralization, and finally, let’s assume this cryptocurrency as a proof of stake consensus mechanism that super secure because it requires more than half of all computers on the network to sign off on each transaction.
On paper therefore you’ve overcome the blockchain trilemma, and this is what many cryptocurrencies have claimed to have done. However, things change very quickly when the rubber hits the road.
Recall that our cryptocurrency has 1,000 computers connected to its network spread around the world. 501 of them must confirm a transaction for it to be valid as per the consensus mechanism.
When the number of users and transactions is low, this isn’t a problem. But as the number of users and transactions starts to grow, it’s no longer possible to efficiently get 501 computers to check each transaction.
For our fictional cryptocurrency network to remain operational, we have to either reduce the number of transactions per second the system can handle. Reduce the number of computers connected to the network, or change the consensus mechanism, so it requires fewer computers to confirm each transaction.
If we choose the first option, we sacrifice scalability. If we choose the second we sacrifice decentralization, and if we choose the third, then we sacrifice security.
In the case of Bitcoin creator Satoshi Nakamoto consciously decided to sacrifice scalability to ensure security and decentralization. Vitalik and the other Ethereum founders made the same trade-off at the outset and Ethereum 2.0 will lightly sacrifice decentralization to get the speed it needs to support its ecosystem. You’ll see why later.
Layer-2 solutions for Bitcoin like the Lightning Network are designed to enhance Bitcoin scalability while leveraging the Bitcoin blockchain for security and decentralization.
Layer-2 solutions, for Ethereum like Polygon (MATIC), do the same with the Ethereum blockchain.
Most high-performance cryptocurrencies have opted to sacrifice decentralization to become scalable and this arguably applies to the layer-2s I just mentioned as well.
To use two quick examples, Polkadot (DOT) currently has about 300 validators. This sounds decentralized at first until you realize that each parachain running Polkadot smart contracts will only use 10 validators.
Similarly, Polygons technically has 100 validators at the plasma chain, which actually runs the decentralized applications on Polygon only has 7 -10 nodes.
What Is Speed In Cryptocurrency?
Now, this brings me to an important point that is almost never raised when examining cryptocurrency transaction speeds and that’s not all transactions are created equal.
Transferring a coin or token from one wallet to another is not the same as executing a complex smart contract in a decentralized application like AAVE flash loan, for example.
There are very few smart contract compatible cryptocurrencies that make a distinction between their smart contract transactions, and regular transactions when advertising their transaction speeds.
When you go digging through their documentation, you’ll often find that smart contract transactions are significantly slower than regular transactions assuming they list those figures at all.
Moreover, many so-called Ethereum killers actually leverage a version of the Ethereum Virtual Machine (EVM), which is inherently limited to 300 transactions per second, even when running on a single computer.
These are a few of the reasons why two of Cardano’s top software engineers did a presentation last year titled quote Lies, Damned Lies, and TPS Benchmarks.
In it, they explained the cryptocurrency transactions are more like packets of data than blips in a digital system. Cryptocurrency blockchains are basically like databases that process and store transaction data.
For example, each block on the Bitcoin blockchain can hold one megabyte of data, which is 1 million bytes. The size of the average Bitcoin transaction is just under 400 bytes, when you do the maths this means each Bitcoin block can fit between 2,500 and 2700 transactions. Now because 1 Bitcoin block is made every 10 minutes, this works out to about five transactions per second.
The reason why Bitcoin’s block size is so small and its block times so long has to do with security and decentralization. Every time a new block is created this adds one megabyte of data to the Bitcoin blockchain regardless of how many transactions are in the block. Don’t ask me why? That’s just how blockchains work.
This means that over time the size of the Bitcoin blockchain grows. Today, the Bitcoin blockchain is around three 337 gigabytes in size.
Every computer that joins the Bitcoin network needs to download this transaction history, so it can know what’s going on and process transactions correctly. While increasing the coin's block size would mean each block could fit more transactions. It would also make Bitcoin’s transaction history grow much faster.
To use an extreme example, if each Bitcoin block were one gigabyte in size, it would be able to handle 5,000 transactions per second. However, the blockchain would grow so fast that only massive data centers would be able to store its transaction history. Meaning only a few computers would be able to join the Blockchain blockchain.
Not only that but every single one of these data centers would need to have the fastest fiber internet on the planet just to keep up with each new block being created. Similarly, if you accelerated Bitcoin’s block time to six seconds, this would increase Bitcoins transactions per second to 500.
The problem here is that six seconds isn’t nearly enough time for enough computers to check whether the transactions in that block are valid or not. This would result in Bitcoins blockchain splitting into multiple different blockchains with conflicting transaction histories which would effectively corrupt the network.
Calculating Cryptocurrency Speeds
Understanding that cryptocurrency speed is just its ability to process and store data makes it possible to calculate the real transaction speed of any cryptocurrency using a few metrics.
All you need to know is;
How big each block is?
How frequently a block is made?
How large the average transaction on that cryptocurrency’s blockchain tends to be?
This calculation is sometimes easier said than done because some cryptocurrencies will actually adapt their block size to accommodate transaction demand.
Ethereum has been gradually increasing its block size over the years for the same reason, Ethereum’s TPS is 15 to 17 and that’s because each blog on Ethereum is 51,000 bytes, a new block is generated every 13 seconds, and the average Ethereum transaction as of late is around 2,500 bytes.
Cardano’s blockchain can handle roughly 200 transactions per second and that could be mapped out from Cardano’s block size of two megabytes, its block time of 20 seconds, and Cardano's average transaction size of 500 bytes.
Obviously, the reason why transactions on Ethereum are so much larger than those on Cardano is that Ethereum has smart contracts whereas Cardano doesn't, yet.
Luckily for Cardano, smart contracts shouldn’t clog up its blockchain because of its dual chain architecture, which separates smart contract data and transaction data.
Unfortunately for Cardano a block size of 2 megabytes with a 20-second speed means that the size of its blockchain is growing fast, and that could become an issue for decentralization in a few year's time. Anyhow, the coolest part about this calculation is that it means the theoretical limit for how fast a cryptocurrency can be is fundamentally set by the speed of the internet around the world.
According to speedtest.com, the average download speed is 100 megabytes per second and the average upload speed is 50 megabytes per second. When you compare this to Ethereum’s two kilobytes per second and Cardano’s 10 kilobytes per second, it’s clear that there’s lots of legroom for cryptocurrencies to increase their speed. So this begs the question, how fast is fast enough?
Fastest Crypto: Transactions Per Second
The de facto benchmark for speed in the crypto space is Visa. This is because, there are an estimated four billion Visa cardholders, which makes Visa's payment network a good template for future crypto adoption.
While Visa’s self-reported transactions per second are 65,000 Visa only processes around 2,000 transactions per second on any given day.
The only existing cryptocurrency that comes close to this level of throughput is currently Solana (SOL), which can handle between 50 to 65,000 transactions per second, even for smart contract transactions.
Now, although I haven’t personally stress-tested Solana’s scalability, FTX and Alameda Research founder Sam Bankman-Fried has.
Sam chose Solana to host FTX’s project Serum DEX, specifically because it was the fastest smart contract crypto on the market.
As you can imagine, Solana has also had its fair share of network issues. This is because it had to compromise on security to have such fast speeds without running off a single computer.
When you consider the fact that Solana transactions can be as large as 500 bytes, this also means the Solana blockchain is generating an insane amount of transaction history data every day.
Although, the Solana blockchain has only been live for a little more than a year that blockchain is already more than 2 terabytes large.
To put things into context, this is more than twice the size of the blockchains of the top 10 cryptocurrencies combined.
As I mentioned earlier, this seriously threatens its decentralization. Now, Solana's solution to this exponential growth is to outsource its data storage to a decentralized storage cryptocurrency project called Arweave (AR).
Solana validators only keep the last two days of transaction data. Now, if you want to learn more about Solana, you can read my most recent post below.
Solana: More Update For SOL?
SOL is going insane and stay tuned for updates new ATH.
Fastest Crypto: Transaction Finality
If you think the crypto speed case is closed with Solana, I’m afraid you’re mistaken. That’s because transaction execution and transaction confirmation are two very different things.
You’ve probably noticed that when you send your cryptocurrency to exchange, there are a certain number of confirmations you need to wait for before the coins or tokens are credited to your account.
Confirmations are basically the number of blocks that have been created since you made the transaction. The first block your transaction is included in is usually the first confirmation and of course, every cryptocurrency has a different block time.
For most exchanges one confirmation is enough for them to credit your crypto unless you’re sending Ethereum Classic (ETC) to Kraken in which case you’ll be waiting for 40,000 confirmations, which is about a week.
This is because Ethereum Classic has been subject to multiple 51% attacks in the past, which is essentially when the network is manipulated by miners to credit cryptocurrency to their accounts.
Cryptocurrency exchanges always want to make sure the transaction is final before they credit your account, and the likelihood that a transaction is final typically increases with each confirmation.
For Bitcoin, the likelihood that a transaction is final is over 99.9% after six confirmations. Naturally. This needs to be much longer for less secure blockchains like Ethereum Classic.
What’s interesting, is that even though Solana has a block time of 400 milliseconds, its transaction finality time is actually somewhat longer. This has to do with Solana’s elaborate consensus mechanism.
By contrast, there are multiple cryptocurrencies that offer instant transaction finality. These include Algorand (ALGO), Ripple (XRP), Stellar (XLM), Cosmos (ATOM), and any blockchains that use the tendermint consensus mechanism.
Even though all these cryptocurrencies are slower than Solana in terms of TPS, exchanges can know for sure that transactions coming from them are final the moment they generate a block.
That’s why the deposit times for all of them on Kraken are noted as near-instant, whereas a Solana deposit can take as long as two minutes on Binance, and not surprisingly this is possible because all the cryptocurrencies are just mentioned are extremely centralized. They only have a dozen or so validator nodes processing transactions.
While these finality thresholds aren’t always needed for the wallet to a wallet or dApp transactions, it’s likely that the crypto merchants of the future will use finality benchmarks similar to today’s cryptocurrency exchanges.
When Speed Doesn’t Matter
There’s only one category left in this crypto speed competition and that is whether this transaction speeds actually make any difference to the end-user. Whether a cryptocurrency is packing a 10K TPS engine or a 65K TPS engine under the hood, is kind of irrelevant and there’s a science to back up this claim.
Decades of research in human factors engineering shows that any feedback a person gets from interacting with a computer within 0.1 seconds is perceived as instant. If the feedback comes within 1 second of the interaction, it’s noticeable but not interpreted as a delay.
Anything longer than 10 seconds is long enough for a person to get impatient and start doing other things while they wait. Attention spans these days.
Anyways, you’ve probably noticed that when you make a payment using a debit or credit card, it normally takes between 5 and 10 seconds to complete, even when it’s an online payment.
Many cryptocurrencies have block times they fall within this range, which would put their transaction feeds on a par with traditional payment systems for the average user finality notwithstanding.
You’ve probably also found that most wallet-to-wallet cryptocurrency transactions, take roughly the same amount of time assuming it’s not Bitcoin, Ethereum, or an ERC-20 token.
This seems like a good thing at first glance but it reveals just how many cryptocurrency projects are centralized and likely lacking security. Bitcoin and Ethereum are slow but do they are secure and decentralized because of that.
I reckon it’s why people are willing to pay such a premium in gas fees to move their funds on those networks. While Ethereum 2.0 will apparently increase its TPS score to 100,000, the things that we’ve explored in this post suggest that decentralization and security will need to be sacrificed to do that.
To quickly recap, Ethereum 2.0 will feature sharding, which is when transaction histories and balances are split between multiple blockchains called shards. To preserve security, these shard chains connect to a central blockchain called the relay chain.
As far as I understand, each shard on Ethereum 2.0 will only have around 128 validators at any given time. These will be pseudo-randomly selected based on how much ETH they are staking.
That’s pretty decentralized compared to other high-performance blockchains, but it’s nowhere near as decentralized as Ethereum’s current form.
What’s more is that communication and transactions between these shards will likely be limited, if not non-existent at the outset.
We would end up in a scenario where Ethereum 2.0 is essentially made up of 64 isolated blockchains with suboptimal security decentralization. This is the same issue that Kusama (KSM) and Polkadot (DOT) are currently facing with their parachain slot auctions, and you can learn more about those by clicking the below post.
Kusama (KSM): Parachain Auctions Are a Goldmine!
Polkadot and Kusama have been preparing their parachain slot auctions to add new projects to their own ecosystem. But…
The massive increase in institutional adoption of cryptocurrency we’ve seen over the last year has put scalability in the spotlight, but it’s left decentralization and security in the shade.
I am aware that there are people much smarter than me who argued that the blockchain trilemma is just an imaginary problem, but try as I might I don’t see a solution to it. Whenever I find a cryptocurrency that scalable, it’s almost never decentralized and sell them secure. The cryptocurrencies seemed to be all three never provide the same performance for their smart contract transactions.
Now that said, I really think Cardano has the right approach when it comes to addressing scalability. Embracing the notion that cryptocurrencies are just distributed computing systems akin to the internet helps identify what the scalability problem really is, and this makes solutions easier to find.
Solana Founder Anatoly Yakovenko seems to be hyper-aware of this and Solana’s decision to delegate its data storage is evidence to my observation. Solana’s only real bottleneck seems to be data storage and that’s something that can be easily addressed as data storage becomes easier and more cost-efficient inside and outside of cryptocurrency.
When you combine access to data with faster internet speeds it suddenly becomes possible to have blockchains that are in fact scalable, decentralized, and secure. This seems to be the kind of future we’re heading towards, but every so often I see something that makes me wonder if we’re headed in the opposite direction.
Many of these centralized high-performance blockchains have become prime candidates for central bank digital currencies (CBDC) namely Algorand (ALGO) and Stellar (XLM). It seems that governments have the same end-user approach in their development of CBDC’s health passports and other borderline dystopian technologies.
The technologies probably won’t be very secure and they definitely won’t be decentralized, but because they’re scalable the end-user probably isn’t going to care so long as it works for them. That is a scary thought but the bright side is that most cryptocurrencies are already capable of throughput that’s comparable to existing systems for the average person.
For me, personally, I found that Polkadot transactions are insanely fast regardless if I’m sending DOT to from, or between exchanges. I could swear I’ve seen summer I’ve within 5 Seconds which is obviously impossible given the Polkadot’s block time is six seconds.
All that’s really missing is awareness education and user interfaces for applications that don’t look like they come from the inside of a SpaceX shuttle. This trifecta poses even bigger problems than the scalability trilemma at this point and that is why I’m here.
- Date of publication:
- Mon, 06/07/2021 - 08:50
Click on the link - it will be copied to clipboard