After years in development, the Cornell team has finally launched Avalanche, the newest blockchain which promises to address scalability as they claim it handles 4,500 transactions a second.
“Over the 45-year history of distributed systems, there have been only three approaches to the consensus problem: Classical, Nakamoto, and Avalanche.”
So says Collin Cusce, a Senior Software Engineer at Ava Labs who goes on to explain these three protocols as follows:
“Where classical protocols must reach consensus with a probability of 1 (P=1), Nakamoto reaches consensus with a probability of 1 minus some teeny tiny chance of error (P=1 — ε). In Nakamoto, this error value is made smaller and smaller over time as more blocks are produced. The more blocks, the chances of a reorganization drop exponentially…
Avalanche consensus, like Nakamoto consensus, is a probabilistic protocol. Just as Nakamoto traded off a small chance in probability for performance, Avalanche embraces probability to make the chance of error just as microscopic…
To start, let’s talk about what a validator does in Avalanche consensus. Avalanche is a voting protocol. Validators listen for transactions. When they hear a transaction, they vote on whether a transaction is to be “Accepted”, and if it is not, it is marked “Rejected”. Transactions that appear virtuous will be voted on as “Accepted” by a validator. If it is conflicting, the transaction will be voted on as “Rejected”. That’s the sum of it.
This voting process in Avalanche is what makes it so unique. Each validator is a completely independent decision-maker. There is no leader election. However, by protocol, each node uses the same process to determine whether a decision is virtuous and how likely they are in consensus with the rest of the network. Once they see a high probability of agreement across the network, the node locks in their vote and accepts a transaction as final…
At a high-level, this means a validator randomly selects other validators to ask them what they like. It does this over and over again on new randomly selected nodes until it builds up enough data to determine that its probability of being correct is so high you may as well consider it impossible that it’s wrong.
By having a validator randomly select other validators in order to ask them what their preferences are, participants in Avalanche build confidence in the correct decision shared by all nodes in the network. With enough confidence, a decision is finalized immediately.”
This insight probably arose out of the blocksize wars when people were discussing what exactly was the role of miners, of nodes, with the two teams saying miners decide, no nodes decide, with neither quite being the case.
The protocol decides, the code. Then the ultimate decision is by individuals who choose what code to run. So if that’s the case then some new Nakamoto is suggesting the code can be made a lot more efficient to both be scalable and runnable on your laptop.
Do we believe it? Well, this has been on testnet for some time. Peter Todd and Gregory Maxwell as well as other bitcoin maxis must have taken a look.
We haven’t heard anything went wrong with the testnet, so if any of these claims are not quite the case, then it’s their job to show so.
The claim being this is a Proof of Stake network that uses a consensus protocol which provides a bigger guarantee than Nakamoto’s 51% or ethereum 2.0’s 33% and does so through what we’ll call the wisdom of the crowds.
This mechanism they say combines the Classical consensus with the Nakamoto consensus to give all the desirable qualities, which amounts to this team claiming they have solved the triangle dilemma of security, scalability, and decentralization.
On the surface, generally this looks like any other blockchain. We run the node as can be seen in the screenshot above, and that seems light but this has just launched on Monday.

Their blockchain explorer is a familiar sight. The usual gibberish numbers and letters making an address, then the to and fro movements.
There’s no epochs or slots here like in eth 2.0, with it looking more like the current bitcoin or eth explorers.
There are however new things. Tokens here for example go to other chains, with three chains out there right now.

On eth, there can be many tokens which all run on eth. Here, you can have many chains which all run on AVAX, the Avalanche token used to pay the transaction fees.
The x-chain is the most used right now because that is where the AVA were claimed from with a private sale raising $12 million while a public one raised $45 million.
The c-chain has been indexed now, and this is basically eth, but plain eth with even addresses looking like the eth one, but they don’t have your balance.
A new chain will be launched and that will have not just your eth balance, but all your tokens and defi farming and MKR CDP positions and basically all and everything that is currently ethereum.
Aethereum is set to be the first fork of its kind for ethereum and to access it will be as easy as setting a custom RPC on MetaMask according to Daniel Laine, a senior software engineer at Ava Labs.
“The date for the Athereum snapshot has not yet been determined. We will let the community know when there’s more information,” Laine says.
“Athereum, when it starts, will have all of the state of the Ethereum network up until a certain block. All the blocks, contracts, etc,” he says.
So if you have YFI, you’ll get YFI on aeth with everything the same except gas will be paid in AVAX instead of eth and also it will be able to handle 4,500 transactions a second.
Months not weeks, Laine said regarding the launch. He didn’t elaborate on what is left to be done. Presumably they waiting for phase zero of eth2 to launch first.
Avalanche however is currently running, with a system overview explaining the c-chain is basically ethereum, but a blanket slate of it.
While athereum is the Avalanche consensus + the ethereum state machine, so developers don’t have to start from scratch, just custom RPC instead.
Avalanche itself seems to be a lot more than just an ethereum mirror with a lot more scalability. You can define your own blockchain, it can be permissioned or otherwise, all still running within the security of this one network.
They’ve also launched accelerators, grants, and much else to attract new devs with the team here being very competent. Emin Gun Sirer, who leads it, is the one guy you don’t want to publicly say he gone look at your code.
We do however want the Todds and the Buterins and all the Nakamotos to look at this and break it if you can, bring down its nodes or crash its chains or whatever else brings out the popcorn because better we know now if it can be broken than later.
If they can’t, then we may be looking at a candidate for third position in crypto rankings, to begin with, because they say this is hugely scalable. Something we don’t believe until we see it, but it’s the Cornell boys that are saying it.