by @CryptoIsCute

How fast is Nano? -------- tps!

In other words...

-------------- people can use it daily

using ------ GPUs of PoW.

that's ------ times Bitcoin's 5 tps,

and ------ times Ethereum's 23.10 tps.

Tps history

Recent measures   exportaccounts

So how does this work?

Each day checks the network's performance via a brief burst of activity. Blocks are broadcast all at once from a few thousand unique accounts, ensuring conditions are as favorable as can be. I've tried to make this analysis as fair and straightforward as possible.

How is tps calculated?

To avoid underestimating the tps, the timer starts the moment the entire batch has been broadcast to all the nodes. We then wait for that many confirmations to occur, which may include blocks from real users. That way other people using the network won't change the results.

Once we have our confirmations, we divide the batch size by the number of seconds that have passed. We then divide by 2 to get our final tps. This is because sending Nano from one user to another requires two blocks, a send and receive, for the money to be spendable. If we didn't divide by 2, we'd need to more than double bitcoin's tps since a single transaction usually has multiple inputs and outputs.

Is this safe for the network?

Absolutely, and due to Nano's prioritization system this traffic should be deprioritized over other people's activity. I've timed the tests to run when usage has been historically low, and the same accounts are used each day to maximize prunability.

What could raise these numbers?

Better hardware and fewer nodes. Nano's network overhead scales with the square of the number of PRs, so if you're delegating to a small or slow one, consider moving away. There is a security tradeoff in doing so but if capacity is your goal, this makes sense. In the long term software updates and moving away from ORV may help as well.

What could lower these numbers?

A few GPUs spamming, the introduction of forks, and bandwidth cuts dramatically reduce throughput. Forks take substantially longer to resolve and can be generated faster than normal block spam since every side of a fork shares the same PoW.

Doesn't Beta do better?

Beta's performance isn't representative of mainnet due to the composition of PRs. This as true for Nano as it is in other cryptocurrencies, where beta setups commonly perform thousands of times better due to their unrealistic testing conditions.

Why make this?

People ask all the time about Nano's tps, but there haven't been good resources for tracking it. Numbers in the hundreds or even thousands per second are commonly thrown around. Nano's official twitter recently suggested that it was about 1000 (the team deleted it and thanked me once I showed this wasn't true — huge props to them for doing that). Sometimes people will be linked the peak performance observed during various spam attacks, but what end users should care about is not the rare moments of bursty confirmations but rather the sustained tps with which they can interact with the network.

Who are you?

I'm Rachel, a blockchain developer who entered crypto after discovering Nano and fell in love with web3. I'm in it for the tech, and want to empower people with insight into the platforms they use. In the days before the infamous spam attack that made the network generally unusable for a few months, I used to make Nano hype videos and fun side projects that demonstrated the cool stuff you can do with Nano. I was also the #1 PR a couple of times.

Now my focus is primarily on Ethereum scaling but I'm rooting for our favorite fast and feeless payments network from the sidelines. Follow me on twitter for more crypto content and pick me as your rep here 💙