DATUM: The Mining Upgrade that Pays for Itself

Hi, guys. I'm Jason Hughes from OCEAN, the VP of development and engineering over at OCEAN. I'm here to talk to you about DATUM. It's a decentralized mining protocol designed to give power back to the miners over block space and pretty much every aspect of mining, so that we're cutting out middlemen. We're basically doing everything we can to give that control back.

​

I wrote the source code for DATUM. I ran the Eligius mining pool from 2012 onward to 2017 or so, so just a little bit of background there.

​

One of the things that people don't realize is, you think about Bitcoin mining and everybody just kind of gets into Bitcoin mining to make money, but they don't realize what kind of money they're leaving on the table by going the centralized mining route. If you look at a bunch of factors, I just did two here to get your attention. As a 1 EH miner, you could be leaving a Bitcoin and a half a year on the table by going the centralized mining route, just from technical inefficiencies in the centralized mining process.

​

So we're going to talk about some of that. I want to talk about some of the numbers, some of the tech, and then the bigger picture with decentralization in Bitcoin mining.

​

Like I said, the DATUM Gateway is the software that I designed for OCEAN for decentralized mining. We're going to put you back in control of mining and make you an actual miner.

​

There are a bunch of inefficiencies in mining where you can lose revenue, where you lose shares as you mine. If you're familiar with Bitcoin mining, you'll see your miners submit shares to the pool. They get accepted or rejected. One of the main reasons they get rejected is because they're stale. The pool sends you work, and by the time that work goes over the wire on the internet, you end up getting it, and you have already submitted shares to the pool. Those get crossed, the pool gets it too late, and it's not worth anything.

​

That ends up being something like 0.1% to 0.5% in most cases on centralized Stratum V1 mining pools. That adds up. It sounds like a little bit, but it adds up. For one miner, as you can see on the slide there, I would go through the calculations if I had more time, but we have anywhere between 0.18 and 0.83 Bitcoin a year per exahash left on the table just because of the latency between your miners and the pool.

​

Then you have the fact that the Bitcoin network has a new block every ten minutes. Every ten minutes your miners need to know what to work on, and that information generally comes from your centralized pool. So you have a bandwidth spike. If you have a really good connection, you have fiber, and you have 5,000 ASICs at your site, that'll be fine. You'll still lose about 0.06 Bitcoin per year with something like a gigabit fiber connection.

​

If you're trying to do an exahash worth of equipment on something like Starlink or something slower, you're going to lose a lot more. As much as three quarters of a Bitcoin or more. The three quarters of a Bitcoin number that I had here was just for a 50 megabit connection on average, which is pretty typical of what you'll see on Starlink.

​

All combined, as a 1 EH miner, you're leaving anywhere from a quarter of a Bitcoin to 1.6 Bitcoin a year on the table just by going the easy route of sending your hash rate to one of the normal centralized mining pools. So we need to fix that.

​

What is the DATUM Gateway? The DATUM Gateway is a work generator that sits at your site. You can't get any faster than your miners talking to a piece of equipment that is right there with them. No latency, effectively unlimited bandwidth because you're not going over the internet. Your ASICs get their work directly from your own site, using work from your own node.

​

As a Bitcoin miner, you should be running your own node, because if you're not, you're just a hasher. You're not actually a miner. You're giving that control to the pool, and you shouldn't be. We're all about Bitcoin. We're all about decentralization. We want to keep as much sovereignty as we possibly can, and DATUM gives that back to you.

​

You have your ASICs. They connect to the DATUM Gateway. The DATUM Gateway gets information from your Bitcoin node about what work it's going to do, and then the Bitcoin node gets that information from the Bitcoin peer-to-peer network, just like your pool would. Your pool normally would get the information from the Bitcoin network, process that, and then have to blast that to every one of your ASICs over whatever internet connection you have. That takes some amount of time. It's a finite amount of time, and there's a delay there. Delay costs you money. It adds up. It's not a lot, but it adds up, and like I said, it can be pretty high.

​

With the DATUM Gateway, it is a pooled mining protocol, so you can coordinate with a pool like OCEAN or any DATUM-compatible pool. Because DATUM is open source, you can coordinate reward splits. When any miner using the pool finds a block, you get paid directly from the block. There is no middleman. The pool doesn't take the money first. The Bitcoin goes from the block to your wallet as a miner. Again, cutting out the middleman.

​

As a bonus, since you are making the block, you get your name on the block. This shows up on sites like Mempool Space and other sites that have adopted that. So there are a lot of benefits, and you're making more money mining. It's kind of a win-win overall. There's just no way to go wrong with this.

​

How does that eliminate these costs? With stale shares, now that your work is being generated on site and you have a normal, big LAN-based tunnel to your miners, there's no delay. As soon as your node knows about the new work, your miners know about it as well. Over a LAN, like one millisecond, virtually nothing. So all that stale share latency is effectively eliminated. You can drop that to near zero. Your miners will never get a rejected share, because they're right there. The work is being generated at your site.

​

Then the bandwidth spikes are the same problem. You have a LAN now. Over the internet, the only thing that has to happen is you have to hear about the new work from the P2P network and just get that work to your DATUM Gateway. Then the DATUM Gateway does all the heavy lifting of processing that work and getting it to your miners, again all on site and pretty efficiently.

​

So you eliminate both of those bottlenecks almost to zero, to the point where you can recoup that money you've left on the table. At the same time, you're helping decentralize the Bitcoin network. We don't want to have ten pools controlling all of the hash rate in the world. You are controlling the block template. You are doing all of that work.

​

If we break it down in the best-case and worst-case scenarios for what you would be doing with centralized mining and what you're leaving on the table, this scales linearly. If you're a 5 EH miner, you're probably leaving anywhere from 1 to 8 Bitcoin per year on the table. Bitcoin is Bitcoin. We're here to make money.

​

Beyond the money-saving aspects and the efficiency boost of doing the right thing for Bitcoin with decentralized mining, you're put back in control. You have your block templates. You're in control of what transactions and what signaling is happening. If there are any kind of network changes happening, you can signal or choose not to signal. You can decide to prioritize transactions. You can put your own transactions in your own block at zero fee. During high-fee spike times, you can decide to mine your own transactions and still get your normal mining rewards for that.

​

It's designed around decentralization. DATUM was designed from the ground up as a tool for decentralization, and not to pull more centralized mining back into the works.

​

Because when you find the block, your block directly pays the miners. These are non-custodial payouts. This is permissionless. This is what Bitcoin is supposed to be, and that's what DATUM does. It gives you that control back so that you can see everything that's happening with your mining. You can see everything that's going on before you do the work, not like with a standard centralized pool where you only see what happens when they find their block and just hope it's what you wanted them to do.

​

That kind of wraps it up. I guess I didn't mention that DATUM is completely open source. The gateway is open source and MIT licensed. You don't have to change anything on your ASICs or your firmware. You don't need any special firmware. This works with everything you already have. One gateway is so efficient it can handle tens of thousands of miners with no problem connected on your LAN, because this is all locally generated and it's pretty lightweight.

​

Famously, when we were testing this originally, we sent multiple exahash to a Raspberry Pi 3 running a Bitcoin node and DATUM, and had virtually zero stale shares and zero lost work. That's the efficiency of the code base. It's written in C. It's open source. It works with virtually any Bitcoin node that supports the getblocktemplate function. You can pair it with Bitcoin Knots, Bitcoin Core, or anything that supports GBT.

​

I only touched on the stale shares and the bandwidth spike savings. Beyond that, there are other ways that you're leaving money on the table with centralized mining pools and Stratum V1 going over the internet. I don't have time to touch on all of that today, unfortunately, but definitely check in with the OCEAN guys. The QR code there sends you to the GitHub for the DATUM Gateway.

​

I'll be around if you have any questions. You can definitely flag me down. We can chat about DATUM. Decentralizing Bitcoin mining is what we're trying to do, and that's what we definitely need to get back to. Hopefully I've informed you a little bit about where you can make more money mining, because who doesn't want to do that? I appreciate you guys having me, and thank you very much. Enjoy the conference.