SundaeSwap Fundamentals
Pi Lanningham
with input from the SundaeSwap team
June 1, 2021
Abstract
SundaeSwap is a decentralized exchange built for the Cardano
blockchain. It allows participants of the blockchain to provide liq-
uidity and create a market for others to exchange their native tokens.
In return, swappers pay a small fee and liquidity providers earn a re-
turn on their deposit. This and future whitepapers will outline an
initial product modeled after the protocol popularized by Uniswap,
with several innovative adaptations for the Cardano blockchain.
1 Background
The term “Decentralized Exchange” (DEX) refers to an application accessible
through a series of smart contracts running on a suitable blockchain, which
enables financial services traditionally facilitated by a central entity. Instead,
trustless parties can participate in a financial market, relying on the behavior
of the smart contracts to secure the transactions.
In addition to decentralizing access to financial services, a DEX also
typically decentralizes profits from those services. Participants who provide
the liquidity to create the market collect a small fee, creating a vehicle for
passive income at returns usually reserved for large institutions and unheard
of for the individual.
Several successful DEXs have been built on existing blockchains: Uniswap
and Curve on Ethereum, and PancakeSwap on the Binance Smart Chain, to
list a few notable examples. SundaeSwap is a DEX being built for a new
blockchain, the Cardano blockchain.
1
The Cardano blockchain is a new third generation blockchain focused
on, among other things, proof of stake for throughput and energy efficiency.
Some reports suggest that the entire Cardano network is 1.6 million times
more efficient than Bitcoin, for example [3].
As this new ecosystem opens up, the users and businesses that choose to
operate on the Cardano blockchain will have a great and pressing need for
the financial services described above.
There’s one difference between Cardano and other blockchains that is
of particular note for this paper. The accounting model and virtual ma-
chine are dramatically different from those on other smart-contract enabled
blockchains. Tokens are tracked as bundles of unspent outputs from previ-
ous transactions, and can be locked with a validation script that determines
under what conditions they can be spent. We’ll go into this in more detail
in Section 3.
This first paper provides the necessary background information that fu-
ture whitepapers will build upon for the SundaeSwap protocol. There are a
number of models we are evaluating, and seek to remain flexible, so details
may change as we approach launch date.
Outline The remainder of this paper is organized as follows:
• Section 2 gives an introduction to the principals of an “Automated
Market Maker.”
• We describe a naive implementation of this scheme on the Cardano
blockchain in Section 3.
• We detail a plan for enabling long term protocol upgrades in Section 4.
• We discuss some promising opportunities for additional improvements
in Section 5.
• Finally, Section 6 summarizes the work.
2 AMMs and Constant Product Liquidity Pools
In classical finance, an exchange acts as a central authority, maintaining an
order book and matching buyers with sellers to facilitate the exchange. Be-
2
cause of the incredible potential this entity has for manipulating the market
for further profit, there is an immense amount of existing regulation in place.
By contrast, one of the innovations brought by the decentralized finance
space is the notion of an “automated market maker” (AMM). In such a
model, the pricing and distribution of assets is satisfied by a mathematical
formula or algorithm.
Initially, SundaeSwap will provide an automated market maker that is
an adaptation of the model popularized by Uniswap. This section, then, will
take some time to describe the Uniswap model for background. [1]
One useful analogy from classical finance to wrap your head around the
notion of a liquidity pool is to think of it as an automatic ETF: A collection
of securities traded and balanced against the market, of which you have a
small ownership of.
In this model, liquidity providers deposit equal values of two assets in
a smart contract, and receive tracking tokens representing their portion of
the pool of assets. For example, suppose a BTC/ETH liquidity pool has
9 BTC and 90 ETH and has issued 9 total tracking tokens to previous liquidity
providers.
Figure 1: An example user and liquidity pool
If a user deposits 1 BTC and 10 ETH, since those represent 10% of the
total value in the pool, the smart contract issues 1 new tracking token, for a
total of 10. That 1 token out of 10 entitles the liquidity provider to 10% of
the pool’s total assets, which equals 1 BTC and 10 ETH, as expected.
The pool also allows “swaps” to happen: someone deposits one asset, and
3
Figure 2: Depositing liquidity and receiving tracking tokens
receives the other, according to the exchange rate of the pool. In the above
example, if I deposit 0.1 BTC, I might expect to withdraw 1 ETH.
Figure 3: An example swap of 0.1 BTC for 1 ETH
However, treating the pricing function as totally linear in this way opens
the market up for abuse. As more and more of one asset is deposited, the
ratio between them changes. Allowing a trade to execute at the current price
puts larger trades at an advantage over smaller ones. Indeed, by depositing
10 BTC and withdrawing 100 ETH, a moderately sized trader could drive the
price to infinity.
Instead, liquidity pools rely on a market maker function. In a “constant
4
product pool,” for example, the maximum amount of the other asset with-
drawn is chosen such that the product of the two assets remains constant.
10 BTC × 100 ETH = 1000 BTC × ETH
Figure 4: A constant product price curve and current price
So depositing 0.1 BTC would bring the total balance of BTC to 10.1, and
we need to withdraw some ETH to bring the product back to the curve.
The amount of ETH left in the pool to maintain this constant must be:
1000 BTC × ETH
= 99.0099 ETH
10.1 BTC
Meaning the user can withdraw up to
100 ETH − 99.0099 ETH = 0.9900 ETH
This introduces a measure of efficiency, known as “price slippage”: 0.9900 ETH
is 99% of the value you would expect to get at the current price. If, instead,
someone swapped 5 BTC, they would receive:
5
Figure 5: Constant product price curve after a swap
1000 BTC × ETH
100 ETH − = 33.33 ETH
15 BTC
33.33 ETH
= 22.22%
15 BTC × 10 ETH BTC−1
This represents a staggering 22% efficiency.
For a given trade size, a smaller pool will have higher slippage, while a
larger pool will have a lower slippage. Therefore it is beneficial to aggregate
as much liquidity as possible to give participants the most efficient trades
possible.
In order to incentivize liquidity into the smart contract, a further fee is
taken off and left in the pool.
Because the tracking tokens held by the liquidity providers represent a
percentage of the pool, when those tokens are burned, the liquidity provider
withdraws an equivalent percentage of assets at the new price, plus a per-
centage of the fees that have been collected along the way.
6
There are further innovations on the Constant Product Pool model that
we will discuss in Section 5.
3 Implementation on the Cardano Blockchain
The above model translates fairly naturally into Solidity and the Ethereum
Virtual Machine: a small collection of functions and storage space that can
be executed to deposit, withdraw, and swap liquidity, executed many times
in sequence within a given block.
Cardano, however, uses a novel accounting and execution model, known
as “Extended Unspent Transaction Outputs” (eUTXOs) [2], which makes
translation of the above model not as straightforward as you might expect.
The eUTXO model implements smart contracts more “passively” than an
explicit function call, and heavily discourages the use of global state.
eUTXOs build on a simpler UTXO model. In the UTXO model, money
is tracked as a chain of custody, by pointing at a specific (unspent) output
from a previous transaction, which then gets consumed as the input to a new
transaction. Very simple scripts can be attached and evaluated to determine
if the output is allowed to be spent.
The model used by Cardano extends this in the following ways:
• The UTXO is equipped with an arbitrary datum
• The script locking the funds has access to input data, known as the
“redeemer,” as well as the entire transaction
We first present one natural way to translate the system described above
to the eUTXO model, as well as a discussion of the trade-offs made along
the way.
In this implementation, a global “SundaeSwap Pool Factory” unique to-
ken exists, locked via a script that allows the creation of specific “Asset Pair
Liquidity Pool” unique tokens. The global token is used in conjunction with
a minting policy to ensure that asset pairs stay unique, rather than diluting
the available liquidity and suffering from poor slippage, as discussed above.
Then, these unique tokens are always locked in a eUTXO alongside the
liquidity stored in the pool, using a validator script that enforces the con-
straints of the pool:
7
Figure 6: A transaction to create a new liquidity pool
• A minting policy allows tracking tokens to be minted so long as the
appropriate liquidity is deposited
• The same minting policy allows tracking tokens to be burned so long
as the appropriate liquidity is withdrawn
• The validator script allows swaps to occur, so long as they respect the
pricing function and fee structure
8
Figure 7: Example swap and withdrawal transactions
9
This model, however, has a fatal flaw. Because any given eUTXO can only
be spent once, as part of one transaction, it appears as if only one swap can
happen per block. On the Cardano blockchain, there is roughly one block
every 20 seconds. This would be abysmal throughput for a decentralized
exchange.
We will discuss the SundaeSwap scaling solution in a future whitepaper.
4 Planning for Upgrades
Given that the space is evolving quickly, it would be peak hubris to assume
that the first model we implemented would be the best. Additionally, re-
leasing a new protocol with improvements is at a disadvantage in that it has
no initial liquidity, as liquidity is locked up in the previous version of the
contract. It is important, therefore, to plan early for that upgrade path.
At SundaeSwap, we plan on enabling a seamless upgrade. To achieve
this, the validator contracts holding liquidity allow it to be spent by a future
version of the protocol. This list of future versions will initially be null, but
through a vote of SUNDAE token holders, this value can be updated to point
to a new contract. Once that value is updated, this locked liquidity can be
spent directly into the new version of the protocol without user interaction.
Figure 8: An example upgrade transaction
This allows future versions of the protocol to be bootstrapped from day
10
one with large amounts of liquidity, accelerating adoption of the improved
protocol once community consensus has been achieved.
Obviously this is a sensitive attack surface, and a huge portion of our
efforts will go towards securing this upgrade mechanism.
5 Future Work
There are a number of extensions to the protocol above that we are finaliz-
ing the details on, and which will be discussed in future whitepapers. For
example:
• The role of the SUNDAE token
• A mechanism to increase throughput dramatically
• A mechanism to provide concentrated liquidity for more efficient market
leverage
• A mechanism to provide secondary derivative markets
• A mechanism to further decentralize the role of the liquidity pool
6 Conclusion
The Cardano blockchain offers dramatic and exciting improvements in terms
of thoughput, fees, energy efficiency. With the launch of Smart Contracts
later this year, it will be poised for a huge surge in economic activity and
utility. As more and more native tokens are created to track and satisfy real
world value, there will be an incredible need for markets to trade and acquire
these tokens.
The above model provides a simple, scalable solution to meet these early
needs, well suited for the Cardano blockchain, and positions SundaeSwap to
expand into more efficient and more sophisticated protocols and instruments
in the future.
11
References
[1] Hayden Adams, Noah Zinsmeister, and Dan Ribinson. Uniswap v2 Core.
Technical report, Uniswap, 03 2020.
[2] Manuel M.T. Chakravarty, James Chapman, Kenneth MacKenzie,
Orestis Melkonian, Michael Peyton Jones, and Philip Wadler. The Ex-
tended UTXO Model. Technical report, IOHK and University of Edin-
burgh, 01 2020.
[3] Steven Ehrlich and Charles Hoskinson. Cardano and ethereum founder
analyzes the newest evolutions in crypto and blockchain technology, Apr
2021.
12
