The ‘Unique Opportunity’ to Upgrade Ethereum’s Virtual Stack

With Eth 2.0’s first hard fork spec mapped out, attention has turned to the planned merge of Eth 1.x and Ethereum 2.0.

And, not wanting to lose momentum around the merge, Vitalik Buterin has proposed making some additional changes to the network, given most people don’t see Ethereum changing much afterward (minus some cleanup, more shards and, of course, our new favorite Ethereum word, rollups).

In two blog posts and on Friday’s All Core Developers call, Buterin made the case for stripping less useful – or maybe even harmful – functions in Ethereum’s codebank sometime before or during the merge. Buterin mainly focused on opcodes used in Ethereum’s Virtual Machine (EVM).

“We have a unique opportunity to make some backwards-incompatible changes to the EVM that could be valuable for Ethereum in the long term,” Buterin said on GitHub Feb. 18. “The portion of applications that would need to be rewritten as a result of these changes is quite small, but it is nevertheless nonzero.”

Chief on that list is the SELFDESTRUCT function which rewards anyone who destroys a contract sitting idly on the Ethereum state. The intended purpose of the opcode was to incentivize Ethereum developers to practice “good hygiene” and destroy contracts when they weren’t necessary anymore. That would help reduce Ethereum’s long-term state size.

However, it hasn’t really panned out like that. Right now the function stands in the way of scaling Ethereum by making it “difficult to move to a different state storage format in the future,” among other reasons, Buterin said.

In fact, many people use the function as a discount of sorts in case Ethereum’s fees rise. Called gas tokens, these tokens can be bought when gas is cheap and spent later when gas is expensive to help lower the cost of a transaction. Ethereum developers have considered removing the opcode from the EVM a few times, most recently in September.

Making changes to the EVM or any other technical descriptions in the Ethereum Yellow Paper has not made everyone happy. Some decentralized application (dapp) creators expressed frustration that functions their projects rely on may be removed, such as the gas that enables dapps to check in on how much gwei is left in a contract execution.

It’s unclear how much support the EVM cleanup pitch will receive. Moreover, any changes to the EVM will come with ample warnings beforehand, Buterin said.

“The overwhelming majority of applications are not dependent on anything that is expected to break here,” Buterin said. “It’s a very small percentage.”

The CoinDesk Ethereum 2.0 validator, officially dubbed “Zelda” by Director of Engineering Spencer Beggs, was activated on Feb. 17. Over the past six days or so, Zelda has earned 0.04 ETH, which is worth roughly $61.80 at time of writing. At this rate, the annual percentage return (APR) of our validator operations is expected to be around 7%. 

If you’re new to Valid Points and the topic of Ethereum 2.0 in general, be sure to check out our 101 explainer on Eth 2.0 metrics to get up to speed about jargon and terminology used throughout this newsletter. 

In the first couple of hours after Zelda was activated on Ethereum 2.0, our validator operations lost roughly $3.45 worth of ether. This was due to a file permissions issue that prevented Zelda from signing off on attestations, which is the most common responsibility required of an Eth 2.0 validator node. (The other less-common responsibility is proposing blocks.)

Updating file permissions and rebooting Zelda was a simple fix that got our validator operations back in the green within 24 hours. 

The first lesson learned from this minor mishap was this: Remember to stay awake for the activation of your validator node to ensure all operations are running smoothly from the get-go.  

Most validators after they have deposited their 32 ETH to the Eth 2.0 deposit contract will be put in a pending queue before they’re activated on the network and able to earn rewards. The amount of time needed for validators to wait in the queue before activation can range from a few days to a couple weeks. 

Rough estimates of the exact day and time a validator will exit the queue, based on how many other validators are also waiting in the line for activation, can be found on block explorers BeaconScan and Beaconcha.in. 

Unfortunately, Zelda’s activation took place at roughly 4:00 (ET) in the morning, which is why most of the CoinDesk staff, including myself, were asleep. Had any one of us been awake for the activation of the node, any irregularities in our operations could have been noticed in advance and resolved more quickly. 

Another important thing to remember is to keep validator operations as simple as possible. About 132 validators have been slashed since the network launched on Dec. 1, 2020. Being slashed on Eth 2.0 carries more consequences than missing out on a few attestations. Slashing occurs when there’s evidence of malicious behavior by a validator. The network can correctly or mistakenly view the actions of a validator as a potential attack or attempt to rewrite blockchain history and data. This results in the validator being forced to exit the network, meaning it is no longer eligible to earn rewards on Eth 2.0. 

Slashing happens commonly when Eth 2.0 validator operators are trying to maximize rewards by setting up two computers to run one validator. When one of the computers goes offline, the other automatically boots up and takes over validator operations. While this sounds like a perfect idea to maximize APR by having your validator running virtually without any downtime, it can lead to mistakes where both computers are running the same validator at the same time. 

As soon as the network detects instances where a single validator is proposing different blocks or signing off on attestations more than once, operations could get slashed. 

“The risk is not worth it,” said the co-lead developer of Prysmatic Labs, Raul Jordan, in an interview with CoinDesk. 

While it might be tempting to try and maximize rewards by complicating the node setup so that there is never any downtime, it might come at the expense of losing the ability to earn any rewards on your staked ETH. 

For more information about slashing events on Eth 2.0 and more comments by Jordan, be sure to tune in tomorrow to our weekly podcast series “Mapping Out Eth 2.0.”

Feel free to reply any time and email research@coindesk.com with your thoughts, comments or queries about today’s newsletter. Between reads, chat with us on Twitter.

Valid Points incorporates information and data directly from CoinDesk’s own Eth 2.0 validator node in weekly analysis. All profits made from this staking venture will be donated to a charity of our choosing once transfers are enabled on the network. For a full overview of the project, check out our announcement post. 

You can verify the activity of the CoinDesk Eth 2.0 validator in real time through our public validator key, which is: 

0xad7fef3b2350d220de3ae360c70d7f488926b6117e5f785a8995487c46d323ddad0f574fdcc50eeefec34ed9d2039ecb. 

Search for it on any Eth 2.0 block explorer site!

Finally, Will Foxley and I will be continuing the conversation on Ethereum 2.0 in a CoinDesk podcast series called “Mapping Out Eth 2.0.”  New episodes air every Thursday. Listen and subscribe through the CoinDesk podcast feed on Apple Podcasts, Spotify, Pocketcasts, Google Podcasts, Castbox, Stitcher, RadioPublica, IHeartRadio or RSS.