High frequency traders are making hundreds of billions of orders per day. And there are many bigger and smaller players.

Don't forget the logs produced by the logging infrastructure.

And what if that infra goes down? Who's watching that?

But of what? What has Binance done 181 trillion times?

Obviously they have. I don’t think they’re throwing away money for logs they don’t generate or need. I just can’t imagine the scope of it.

That is, I know this is a failing of my imagination, not their engineering decisions. I’d love to fill in my knowledge gaps.

If it's 181 trillion each year, it's only 6 million per second. There's a thousand milliseconds in each second so Binance would need only several thousand high frequency traders creating, and adjusting orders, through their API, to end up with those logs.

Binance has hundreds of trading pairs available so a handful on each pair average would add up.

They are application logs, so probably nearly every click on their website.

Yeah I don't get it either, something seems deeply wrong here.

These are impressive numbers but I wonder about something... Binance is a centralized cryptocurrencies exchange. But AIUI "defi" is a thing: instead of using a CEX (Centralized EXchange), people can use a DEX (Decentralized EXchange).

And apparently there's a gigantic number of trades happening in the defi world. And it's all happening on public ledgers right? (I'm asking, I don't know: are the "level 2" chains public?).

And the sum of all the public ledgers / blockchains transactions do not represent anywhere near 1.6 PB a day.

And yet DEXes do work (and often now, seen that volume picked up, have liquidity and fees cheaper than centralized exchanges).

Someone who know this stuff better than I do could comment but from a quick googling here's what I found:

    - a full Ethereum node today is 1.2 TB (not PB, but TB)
    - a full Bitcoin node today is 585 GB

There are other blockchains but these two are the two most successful ones?

So let's take Ethereum... For 1.2 TB you have the history of all the transactions that ever happened on Ethereum, since 2015 or something. And that's not just Ethereum but also all the "tokens" and "NFTs" on ethereum, including stablecoins like Circle/Coinbase's USDC.

How do we go from a decentralized "1.2 TB for the entire Ethereum history" to "1.6 PB per day for Binance transactions"?

That's 1500x the size of the freaking entire Ethereum blockchain generated by logs, in a day.

And Ethereum is, basically, a public ledger. So it is... Logs?

Or let's compared Binance's numbers to, say, the US equities options market. That feed is a bit less than 40 Gb/s I think, so 140 TB for a day of actual equities options trading datafeed.

I understand these aren't "logs" but, same thing...

How do we go from a daily 140 TB datafeed from the CBOE (where the big guys are and the real stuff if happening) to 10x that amount in daily Binance logs.

Something doesn't sound right.

You can say it's apples vs oranges but I don't it's that much of apples vs oranges.

I mean: if these 1.6 PB of Binance logs per day are justified, it makes me think it's time to sell everything and go all-in in cryptocurrencies, because there may be way more interest and activity than people think. (yeah, I'm kidding)

EDIT: 1.2 TB for Ethereum seems to be a full but not an "archive" node. An "archive" node is 6 TB. Still a far cry from 1.6 PB a day.

Archive node can go as low as ~2TB depending on the client, but regardless you do want to compare it to a full node, as the only difference is an "archive" node keeps the state at any block (e.g you want to replay a transaction that occured in the past, or see a state at a certain block). The full nodes still contain all events/txs (DEXes will emit events, akin to logs, for transfers) - so it's fair to compare it to a full node.

On CEXes you see a lot more trades, market makers/high frequency trades just doing a ton of volume/creating orders and so on, since it's a lot cheaper. CEXes without a doubt have more trades than DEX by orders of magnitude.

Additionally, they probably log multiple stuff for every request, including API request, which they most definitely get a ton of.

Albeit 1.6PB/day may be somewhat exaggerated, comparing Binance to Ethereum needs much more consideration including:

- Ethereum's transaction throughput is normally 12~20tx/sec so there cannot be a "high-frequency trades" on Ethereum smart contracts with naive contract interaction(it will cost enormous fees). There are scaling concept like "layer-2" or "layer-3", but they still cannot beat highly optimized centralized server applications. Decentralized exchanges have different schemes to centralized ones to reduce txs to discover the price(keyword: AMM, "automated market maker")

- The transactions per sec metrics are just recording "confirmed" txs by the blockchain, and many "retail-squeezing" trading txs (called MEV, maximal value extraction) are competing behind the blockchain and only one tx is chosen by the blockchain, which will rebate most profits to the blockchain validator(which is analogous to the HFTs on the centralized exchanges).

- The blog post's argument would count all logs of intermediate hops, like L7/L4 proxy and matching engine and so on, and Ethereum's full node storage is only a single component which is almost like a non-parallelized matching engine. Maybe we should also count logs of public RPC nodes of Ethereum? (Also many txs are not gossiped to the public mempool so these are hard to count)

Defi is a fraction of whats happening on cefi exchanges. I have a customer placing hundreds of billions of orders per day on binance. And they are by far not the biggest players there.