Network congestion often hits in the worst possible time — MakerDAO 12th of March case
The world is in turmoil — markets are crashing and Covid-19 pandemic gets worse every hour. You don’t need extra problems in this environment but this is exactly the time where many systems are stress tested and flaws appear. Unfortunately, for MakerDAO, the provider of stablecoin DAI, problems just started with the market crash.
12th of March — the market crash
MakerDAO is a dominant player in the decentralised finance realm. It offers a solution that allows you to use your ETH or BAT cryptocurrency as collateral and create a stablecoin DAI out of them. You need to collateralise more than 150% worth of the underlying asset (ETH or BAT) to keep your position from liquidation.
On the 12th of March, the ETH price crashed more than 30% in a matter of hours. A lot of positions fell under the required collateralization level and liquidation process started. This is normal behaviour for MakerDAO, but the extreme market situation pushed the system to its limits.
Market crashes always cause a lot of activity t — users move their holdings to safer places, need to add more collateral to leveraged positions, or just need to re-capitalise from the market moves. This activity clogged the Ethereum network: transfers got slow and network fees went through the roof. For MakerDAO, this meant that many were not able to act in time to protect their positions and got liquidated before they could act.
On the other side, liquidators had a similar situation. Clogged network and fast-falling prices prevented Keepers (users who liquidate and buy these positions) from liquidating positions in a timely manner. In addition to that, the MakerDAO price feed was lagging behind actual market rates. For example, at one point the quoted price of 160 dollars for 1 ETH in the MakerDAO price feed was in reality 130 dollars per 1 ETH on exchanges. This meant the 3% discount provided to Keepers on liquidated positions was still considerably higher than the actual market price, preventing Keeprs from entering the internal MakerDAO liquidation market. However, even if they wanted to bid, the Ethereum network was so congested that it kept them from bidding on the available positions.
This situation was capitalised on by one Keeper who was willing to pay high fees to get into the network and made basically zero dollar offers for the locked Eth collateral assets. As incoming cash was held up in network congestion, many users lost all their collateral (instead of 13%, the expected maximum loss in normal collateral call situations on MakerDAO). As the Eth collateral had been purchased for zero dollars, this subsequently introduced about 4 million dollars worth of bad debt into the MakerDAO system, resulting in a material loss for both the people who’s collateral had been liquidated at zero value, and a loss for the holders of Maker, the loss bearing token that stands behind the collateralised contracts on MakerDAO.
The one redeeming feature of the lag in the price feed is that it arguably helped some users to win time and come in to add extra collateral to their position to prevent liquidation.
Congestion, even a possibility of it, introduces fear
Clogged network events like the 12th of March or Crypokitties situation are more than just a temporary network slowdown.
- These events prevent creating an environment that users trust
- They cost real money for network users — from higher fees to lost positions because it wasn’t possible to act in a timely manner.
- Throughput volatility and cost will deter businesses from the network. Joining a network that can kill them with its fees or slowness is an unnecessary risk.
Over two years on from Cryptokitties causing the first big network congestion event, Ethereum is still showing that it is not ready for products that need to handle big throughput during periods of peak load. This time, real money was lost.
Linear scalability is needed today
Ethereum is still in its early days but its usage is growing rapidly. Early adopters are more willing to go along with various problems, unfair situations and take risks, but it’s not suitable for everybody.
As the network and its various products mature, we need an environment that is able to act as a value layer on top of the internet. As it was the case with the internet, many use cases today were not even imagined at the beginning. The same thing is happening with decentralised value protocols. Any exponentially slowing function is not future proof — linear scalability is desperately needed to future proof the decentralised revolution currently underway.
Radix has a laser-sharp focus on building a linearly scalable, low fees network that is able to offer anyone anywhere friction-free access to the digital economy. Read more on how Radix DLT will achieve the future of scalability with its unique data model, Cerberus consensus and Radix Engine https://www.radixdlt.com/post/cerberus/
Piers Ridyard, CEO of Radix DLT