Transaction fee levels over time

What are the cultural norms that MobileCoin hopes to achieve over the long haul for transaction fees?

If a cryptocurrency network is able to process 10,000 transactions per second at 1¢ USD each, then it has the capability to process merchant transactions for physical goods. That would put it on par with the Visa payment card network today.

If a cryptocurrency network can do an order of magnitude better on both speed and cost, then the use cases expand. If the network can handle 100,000 transactions per second for ⅒¢ USD ($0.001) then use cases like streaming payments become possible. A user could pay small amounts of money continuously until they decide to stop a service. The level of ⅒¢ USD also enables micro transactions for digital goods in video games.

Josh said:

The fee on the testnet will start out as a static number and will have a fee controller that tunes it over time to manage network congestion. Basically, if the network is uncongested we want the fees to be as low as possible. If the network is congested, we want to clear the maximum number of transactions per second until congestion subsides.

Eventually all fees on the network will be determined algorithmically by demand. For now it’s just a number we picked.

So I’m curious about what the team would like to achieve for average daily transaction fee. Would the higher surge pricing happen all the time, like daily during peak demand, or only rarely, like once per month during a holiday celebration?

By thinking about these levels of transaction processing capability, we can ensure the network will be accessible to a wide audience.

In the long run, MobileCoin’s current protocol has a maximum block size (maybe hundreds of transactions), because the system has a throughput-limit (60tx/s I think? or is it 100tx/s?). In that case (maxed-out tx volume) a fee market will probably be unavoidable.

Design of a dynamic fee controller has not started, so I can’t speak to design requirements/expectations there.

1 Like

Thanks, koe. Since the goal is to have 1 billion people using the network, these sorts of questions are good to think through.

I found a mention about scaling issues over on Hacker News.

Josh wrote:

We’ve tested MobileCoin at ~100 transactions per second right now using low core count boxes. We suspect we can scale to 10,000/second on the existing tech stack by throwing bigger boxes at it AND doing some performance tuning. SCP has been shown to hit very large tx/s numbers so it’s just a matter of tweaking it until we get those numbers out. 100 tx/s is more than adequate for quite some time for our use case.

1 Like

I found another mention about fees on Reddit.

Josh wrote:

RE: Cost, the foundation suggests fees to the network and is working to recalibrate the fee down to ~$.04 USD/transaction (hopefully very soon). There’s a vote coming shortly to make that a reality. The foundation hopes to constantly keep fees very low (the price of the coin jumped faster than we were able to modulate the fees down).

That seems to indicate that fees will not be low enough for streaming payments or video game assets. However, that is much lower than BTC and ETH presently.

I understand that the focus of the project is on mobile fast & private transfers (not streaming or video games) that’s why. I don’t think that MOB is trying to become a universal solution. I would personally use LND for those usecases.

Cool, thank you for the suggestion. I tried to locate the network you’re talking about, but LND is Did you by chance mean a different one?

@wrt how low do you think the fee needs to be for video game assets?


@Josh One way to think about video game assets is to value them at 10¢ (US$.10). For transferring small assets like that, you’d want the transaction fee to not exceed 1% of the value of the asset. That means the transaction fee would need to be ⅒¢ (US$.001). While those are approximate values, they provide a framework to get our heads around the use case.

Another consideration is whether memo or computational data will be charged extra. In both Bitcoin and Ethereum, when you make a transaction that includes extra data, you pay more. Video game assets by their nature will include memo and computational data. So the transaction fee also needs to be low enough that it can allow for the data to be included in the transaction and still not be cost-prohibitive for the use case.

The upside, if you can get the transaction fee low enough, is that you unlock more use cases.

Low transaction fees are like the currency equivalent of the Ikea business model. Ikea sells furniture for a very cheap price. But because they sell so much of the furniture, their company makes over $10 billion in profit each year. When you can drive the cost down, while still providing a good service, then you’ll be rewarded with growth.

Steam has a fee of ~15% on video game assets sold on their community market.

For example, the CS: GO items go from 0,03€ to over 1500€. Anyway, that is a centralised marketplace and Steam would have no reason to lower their fees. :smiley: