I’m happy to share Settle, a personal research project I’ve been working on over the past 8 months.
Settle’s goal is to explore a new financial trust primitive, and doing so, construct a decentralized trust graph enabling (totally) free exchange of value without relying on a blockchain.
Value-exchange for humans and machines
There are still barriers to enabling at scale and fluid transactions between machines (and humans).
These transactions would have to come at very little or no fee which rules out fiat currencies (also why would a machine trust a nation-state backing?) but also most cryptocurrencies backed by a blockchain, as maintaining the blockchain generally costs quite a lot on a per transaction basis (if only to prevent spam).
Lightning networks resolve that issue at the transaction level, but users still have to get their hands on the underlying currency to set up channels, which creates a real barrier to entry as well for machines.
Also, and this is more likely a hunch, machine to machine transactions at scale won’t happen on “globally shared currencies”. They’ll transact in KWH of energy, KB of bandwidth and storage, FLOPSH of computing power, APICALL, or KM of drone delivery directly, in a decentralized way, without jumping through a shared, centrally trusted and therefore expensive currency (expensive in transaction fee, or expensive to setup, or expensive in regulation burden).
Currencies operate on a centralized trust graph.
This sentence is almost tautological but let’s consider both fiat and cryptocurrencies:
Fiat currencies value is by definition backed by the government that issued it. Hence, by construction, the trust graph powering exchanges based on fiat currencies is centralized (it’s no surprise that we generally call the issuer of a fiat currency a central bank). Two individuals transacting in USD must both trust the US government backing of dollars.
Conversely, cryptocurrencies such as Bitcoin or Ethereum have decentralized the process of issuing and managing a currency. But while the operations of such currencies, based on blockchains, have been fully decentralized, the trust graph of these cryptocurrencies have remained entirely centralized. Everyone need to trust Bitcoin to transact in Bitcoin, and everyone needs to trust Ethereum to transact in Ethereum or assets issued on the Ethereum blockchain. The centralized nature of the trust involved in these cryptocurrencies being actually at the core of how these currencies operate, as the only viable way to properly incentivize a proof-of-work system.
Explore decentralized trust based value exchange
Settle was motivated by the exploration of whether it would be possible to create a decentralized trust graph between user-issued IOUs (“I owe you”) and leverage it to safely exchange value without a central authority or shared blockchain to maintain, getting in exchange free transactions and free onboarding, specifically important for machine to machine transactions.
Rephrased in researchy terms, Settle demonstrates that it is possible to operate a safe credit network (a currency system without double-spend) without requiring global consensus (a shared global state machine, or blockchain) if you accept the following constraints:
- nodes have to be online
- trust between nodes has to be expressed explicitly
- when trusting a malicious node, users can lose up to the amount of trust they placed in it.
It’s still a work in progress, but more information can be found on these claims in the documentation.
Even more importantly, the engineering contribution of Settle, is a simple HTTPS API to operate such network and an even simpler command-line to interact with it.
The Settle network is operated by mints.
Since there is no shared blockchain, the nodes of the network have to be online. So, very much like email, users register on “mints”, a server of their choice (possibly their own), that manage the assets they issue as well as the trust they express towards other assets in the network.
You can freely register with the
settle command-line on the mint I opened
m.settle.network (no need to be human, you just need to be able
to receive email):
And you can also directly log in the mint of your choice with:
More precisely, mints are in charge of maintaining the following for their users:
- the list of assets issued by its users
- balances associated with these assets
- an authoritative list of offers put on the network by its users
- an indicative list of offers involving the assets it is authoritative for (order book)
One nice property of this, is that users don’t have to manage a public/private key pair, they just have credentials with their mint that can be rolled or retrieved easily. But they do have to trust their mint.
Value is exchanged by crossing offers across mints
Users of Settle issue assets (or IOUs) of the form email@example.com[USD.2]; an IOU from Kurt on a mint operated by Princetown, for USD dollars expressed as cents. Each users can freely issue their own assets. As a user’s mint is authoritative for the assets they issue, it maintains and manages balances (of other users) for these assets. Hence, issuing or transferring IOUs to others simply involves posting instruction to one’s own mint. As an example firstname.lastname@example.org can simply issue a certain amount of email@example.com[USD.2] to firstname.lastname@example.org by posting an instruction to the mint at princetown.edu. Since the mint is authoritative for that asset, no synchronization is required with the mint at npl.co.uk.
Let’s assume that both email@example.com[USD.2] and firstname.lastname@example.org[USD.2] were activated by their respective users on their respective mints.
settle mint USD.2 # activates USD.2
At the mint level, users express trust in the network by posting offers on their mints to issue and exchange assets they control in exchange for assets controlled by other users they trust. Maybe Alan trusts Kurt for up to $200. Alan will represent that trust by posting on his mint an offer on pair email@example.com[USD.2]/firstname.lastname@example.org[USD.2] for 20000 at price 1⁄1.
settle trust email@example.com USD.2 20000 # alan runs this
Similarly, let’s assume that Kurt trusts John up to $100 and has an outstanding offer on pair firstname.lastname@example.org[USD.2]/email@example.com[USD.2] for 10000 at price 1⁄1. These two offers will allow John to transact with Alan without requiring Alan to trust John directly. If Alan sells a machine part for $10, by crossing the two offers, John can buy that part from Alan. He first issues and exchanges firstname.lastname@example.org[USD.2] 1000 for email@example.com[USD.2] 1000 using the offer from Kurt, and then exchange the acquired firstname.lastname@example.org[USD.2] 1000 for email@example.com[USD.2] 1000 using Alan’s offer that he can finally credit back to Alan in exchange for the part.
settle pay firstname.lastname@example.org USD.2 1000 # john runs this
After this operation is complete, Alan holds an IOU from Kurt for $10, i.e. he will have a balance of email@example.com[USD.2] 1000 on the mint at princetown.edu, and Kurt holds an IOU from John for $10, i.e. he will have a balance of firstname.lastname@example.org[USD.2] 1000. Alan’s offer will be valid for a remaining 19000 and Kurt’s offer for a remaining 9000.
The mint protocol provides a mechanism called transactions (backing the
command above) to cross a chain of offers atomically and safely. A
transaction involves reserving funds along that chain and committing the
balance operations on each mint that participate once the offers are secured.
Transactions, if successful, are instantaneously confirmed and fee-less.
Of course users can also specify trust with a discount (because of a possible risk) or an exchange rate, as an example:
settle trust email@example.com GBP.2 1000 with USD.2 at 122/100
/The m.settle.network mint/
Available open-source is a Go implementation for mints, as well as a
settle, that you can install locally with:
curl -L https://settle.network/install | sh && export PATH=$PATH:~/.settle
I’m also maintaining a mint at
m.settle.network to let people issue and
exchange assets as well as a guide to setup your own mint:
At the end of the day, Settle’s motivation is to make currency issuance and exchange easy and available to everyone (including machines). I think that Settle can be particularly useful in situations where currency issuance is implicit and therefore often lacks liquidity.
One perfect (though rather mundane) example where Settle could potentially inject liquidity in an implicit network of IOUs, are the balances we maintain with our friends and family. Settle could (at least with your technophile friends until we have more than a command-line) let you use the balance you have with one friend to pay another one, or even buy stuff, depending on the topology of the trust graph. Even if mundane, powered by a mobile app, this could be a pretty exciting application of Settle.
While I’m also excited by potential use-cases related to self-driving car networks or online commerce, I’m convinced that a any successful use-case for Settle, is likely to come from somewhere unexpected.
With that in mind, if you wish to get involved here are a few things you can do today:
- First and foremost, install
settle, register, mint a few assets, ask your friends, colleagues, drones, self-driving cars to do so, trust them and ask (instruct?) them to trust you back and attempt to run a few transactions. That’s the best way to test the system and discover interesting use-cases.
- If you want to learn more about the underlying protocol, you can play with
setting up your own mint: that’s definitely beneficial for the comunity and if
you do so, let’s add it to the
- If you think of, work in, or have heard of implicit networks of IOUs that are lacking liquidity, let’s try to apply Settle to it and see if it can solve a pain point experienced in that market. (Another hunch: I’m quite convinced, despite not being an expert at all, that online games artifacts could be a good fit?)
- Finally, there are quite a few large-scale engineering projects that can be built on top of Settle. Most obvious ones I can think of are: a trust path discovery service, a mobile app (and probably service as well) that makes it easy to use Settle for non-technical users, or a variety of gateways.
In any case, I hope that you’ll enjoy learning more about Settle and the model it proposes. Please don’t hesitate to reach out directly or on the public mailing list if you have any question!
 That’s what’s great about personal projects, it’s OK to invest time and work on a hunch!
 See the Mint documentation. In particular, the protocol is safe in the sense that there is no double-spend (but users can lose up to the amount of trust they place in malicious users).
 Skipping a few steps and notions, but for reference, the exact commitment (ensuring safety) that a mint is doing on behalf of its user when accepting to participate in a transaction is the following:
Mint at hop h (position along the offer path), commits to: - settle (irrevocably issue/credit the funds) if: - it has not yet canceled the transaction. - it is presented with the lock secret. - node at h-1 has made the same commitment. - cancel the transaction if: - it has not yet settled the transaction. - node at h+1 has canceled.
 The main idea here would be to pay rides on the Tesla Network in KWH.2 instead of any specific currency. Teslas could be configured to trust the Tesla Network solely modelx_AX7GD@tesla.com[KWH.2]/firstname.lastname@example.org[KWH.2] 1⁄1 with the Tesla Network trusting the Superchargers Network with email@example.com[KWH.2]/firstname.lastname@example.org[KWH.2] 11⁄10 such that Tesla would get a cut of each transactions (10% here). Reciprocally the Superchargers Network could trust the Tesla Network at a flat rate email@example.com[KWH.2]/firstname.lastname@example.org[KWH.2] 1⁄1 to let Teslas pay for their recharge as they earn KWH.2. Car-owners would also trust email@example.com[KWH.2] using it to “pay” at Supercharger stations when they ride their own Tesla which would in turn enable Teslas to pay for charges at home stations. From there Tesla could set prices for their Supercharger in each currency by trusting local nodes at various rates, or even local power plants for KWH.2 directly.
 Online merchants could use Settle to issue store credit that is easily usable at other places. This could potentially solve the problem of unlinked refunds in e-commerce while ensuring merchants that these credits would, by construction, be eventually used at their store.
 Lets users deposit currencies (fiat or cryptocurrencies) in exchange for trust on the network, and get that currency deposited back to their account by paying the gateway on the network: by depositing $20 at acme.com (using a card payment or transfer), I would receive a trustline from firstname.lastname@example.org[USD.2] to email@example.com[USD.2]. Conversely, when paying firstname.lastname@example.org USD.2 100, it would automatically deposit back the associated funds to my bank account.