Ethereum scaling solutions. What a wonderful, fiercely debated topic by the community. Since way before the DeFi craze started and long before high gas fees started eating into our profits, we've been talking about it.

Let's get one thing out of the way. Until about 2 hours ago, I wasn't a big fan of any proposed scaling solution that isn't natively integrated into Ethereum. But after reading the Hermez whitepaper, do see a glimpse of light at the end of the tunnel. Let's look behind the current CT hype and find out why.

What's Layer 2?

In short, Layer 2 is an umbrella term for all attempts at taking some of the processing off Ethereum (i.e., taking something "off-chain") to ease congestions and (quite drastically) lowering Gas fees.

If you've been around for longer than 5 minutes, you've probably heard about all the amazing solutions. And they are always just around the corner. There are even people that believe Ethereum scaling solutions are already here.

In reality, Ethereum still only processes about 15 transactions per second. This number is most often compared to Visa, which can process about 2,000 transactions per second.

Ethereum transactions per second (Source)

A variety of these Layer 2 solutions is currently under development. Some of these have even made it to the mainnet, but none of them is mature or adopted enough to make it worth using — for now.

Many of which you've probably already heard about. Plasma, Raiden, xDai and OmiseGo, to name just a few, have been around forever. But the hype, as usual, is overpromising and underdelivering. Is Hermez any different?

Ethereum vs Hermez transaction size (Source).

What's Hermez?

One of the proposed scaling technologies for Ethereum is fittingly called rollups. The idea is that you take a bunch of transactions, verify them off-chain, and then only store the proof on-chain.

Hermez aims to deliver one of these solutions.

Every 10 minutes, Hermez will use what's technically called "ZK-Rollups" to create a batch of transactions and only store the result on-chain. This significantly increases the amount of throughput as the heavy-lifting is done by Hermez instead. There are drastic implications, positive and negative though.

Assets are stored off-chain

With the current design, L2 networks cannot communicate anything on-chain. All funds are stored, and all transfers are executed, in a fenced environment.

Imagine Ethereum transactions would cost $100/transaction, but Bitcoin could do the same transfer for just $3/transaction. You could swap your $ETH for $BTC, transfer the funds and swap them back into your original $ETH.

Let's ignore for a second that you would have to pay an exchange a fee (twice) and that the BTC/ETH rate might change while you're exposed to Bitcoin. The concept of Layer 2 would then be somewhat similar. All assets you want to make "Layer 2"-capable would first have to be converted to off-chain assets.

Expect assets like $hETH, $hWBTC and others to take this spot. These would be 1:1 pegged to the assets, equivalent to $WBTC right now. In Hermez's case, the originals would be locked in smart contracts while within their Layer 2.

Not all transactions are supported

Continuing this analogy and assume you are still in Bitcoin. Some awesome smart contract just popped up and you'd like to invest/play/ape into it. Unfortunately, Bitcoin doesn't have ability to do that. You'd have to swap back first. Of course, as soon as you swapped back, you're exposed to the high transaction fees of Ethereum again. 🤯

Hermez only supports ordinary transfers. Think: sending some Ether from your personal wallet to an exchange (that doesn't use contracts, e.g. Binance). This is still huge, as per their whitepaper.

Why focusing on transfers? It turns out that more than 50% of transactions on the Ethereum network are transfers, and a large percent of these are deposits and withdrawals from exchanges. Demand could be reduced by a significant amount if exchanges started using rollups, or (in the ideal case) even agreed to meet on the same rollup. (Source)

Incompatibility across the board

Here's the main issue with approaches that are not integrated into the Ethereum core itself. Any solution will have its own protocol, it's own process for on/off-boarding assets and it's own nodes/coordinates/stakees — whatever you want to call it. As it stands, they'd all be incompatible with each other.

What that means is that, if Binance for example opts into supporting Hermez, but the wallet you use only supports xDai, you will not be able to benefit from it. This incompatibility and competition will make it hard for companies and individuals to choose and, probably, hinder progress for a while.

Transaction costs may be less transparent

This one is debatable, and it's something that I did not understand from reading the whitepaper. But after contacting the team, it became clear that fees paid on Layer 2 are not always paid in Ether.

Anything you transfer to, from or within the Layer 2 network will cost a small fee, deducted from the total amount you're sending. I can see all types of issues with this, but we will have to see how this turns out.

A bad user experience

All the above leads leads to a terrible user experience. If it isn't easy to use, widely adopted and therefore provides a meaningful benefit, nobody will use it. When I started to read the whitepaper, I prepared myself for disappointment. However, my concerns were addressed on page 20:

To achieve a robust ecosystem of rollups, it is important to enable migrations between different rollups and other L2 applications. To this end, together with the Ethereum community and other researchers from different rollup solutions, some members of the Hermez decentralized community are working on the standardization and compatibility between these technologies. (Source)

It appears that the Hermez team is aware of this and, in fact, aims to build a standard for migrating transactions between different rollup solutions. In the whitepaper, they link to this Ethereum research thread.

Hermez aims to further improve the user experience by providing a web wallet that interacts directly with MetaMask. While still requiring movement of funds between different channels (on/off-chain), it should make it easier.

De/centralized 🤔

Hermez used what they call "coordinators" to verify off-chain transactions. These coordinators can bid to be rewarded the right to verify the next batch of transactions. While the network is intended to be fully decentralized, it will initially be driven by a "boot coordinator" which the developers control.

Over time, this boot coordinator will become less relevant as the amount of fees that it can consume (and therefore amount of HEZ it can bid to stay sustainable) is reduced by 5% every month. Eventually, other coordinators could therefore outbid the boot coordinator, making the network fully centralized.

For all intends and purposes however, this means that Hermez isn't going to be centralized from the start. The whitepaper mentions that the bootstrap period is expected to last 1-2 years. Any longer, the control will be transferred to the HermezDao, powered by Aragon.

What about the token?

Hermez introduced their native $HEZ token on Uniswap, without prior private or pre-sales. Starting at $1, it quickly went to $3.16 before settling now at around $2.50 per token.

The token's primary purpose is for coordinators to bid on the right to verify the next batch of transactions. Coordinators get a bit of everything that is transferred, so there is a bit of a "spend money to make money" dynamic going on.

So far so good, but unfortunately the inflation is quite high. Quite a lot of it will go to the founders (different from the developers). Personally, this is a big red flag for me as founders might dump without hesitation.

Vesting schedules

A little confusing, but Hermez has two different vesting schedules. One for their founders and developers, and one for the token economy and partnerships. Instead of boring you with the details, here's what's going to happen.

After 6 months, very high inflation is going to kick in.

  • Founders / Developers: 36,400 HEX/day
  • Token Economy / Partnerships: 106,600 HEX/day

This comes down to 0.143% inflation per day or 52.195% per year. Quite a bit of selling pressure for such a young and unproven project. This will last 2-3 years before all 100 million tokens are released.

It is worth mentioning that the whitepaper mentions incentives for early adopters in the form of HEZ airdrops. Once the network is live, it will take snapshots on to remember who deposited DAI, USDT, ETH and HEZ.

Final thoughts

Well, this one was a little bit of a rollercoaster. I started with the assumption that Hermez will be just another Layer 2 solution in a sea of many. After going through the whitepaper, I changed my mind however. Offering a web wallet that directly connects with MetaMask as well trying to create a standard for the industry are both good starts.

The team also seems to be very knowledgable, not only in the context of scaling solutions, but also when it comes to security. For example, borrowing Groth16 from ZCash to prove/verify transactions efficiently off-chain. These proofs are then stored on-chain, which is different from some other solutions.

I still haven't quite made up my mind, but Hermez is definitely interesting enough to follow and, potentially accumulate after the hype has died down. The inflation may prove problematic though, so I'd likely exit before it starts.


What's your take on Ethereum scaling solutions? Do you follow any in particular or are you just waiting for a fully functioning Ethereum 2.0? Let me know your thoughts on Telegram.

— Robin