The Real Discussion About Ethereum's Next Hard Fork Is About to Begin

More than two dozen ethereum improvement proposals (EIPs) have been submitted for review and inclusion in ethereum’s next system-wide upgrade or hard fork, dubbed Istanbul.

The list – with 28 official EIPs and at least one other set to be added – include changes to the $27 billion network that impact its mining algorithm, code execution and pricing, data storage, and much more.

About a dozen of these proposals were discussed at length by ethereum core developers during a bi-weekly call on Friday. However, the majority ended up being tabled for further debate, with only one EIP receiving a tentative approval.

“We’ll talk more on the All Core Devs Gitter channel to wrangle in some of these EIPs that are still stuck in proposed and as quickly as possible decide on which ones are being implemented for Istanbul,” said Ethereum Foundation community relations lead Hudson Jameson before ending today’s call.

As noted by Jameson, the hard deadline for all Istanbul EIP submissions passed last Friday and now developers are working to reach agreement about which proposed EIPs can be deemed officially “accepted.”

Decisions made

The one EIP to receive a tentative approval Friday was EIP 1108, which proposes a minor change to gas fees on the ethereum network. Developers emphasized that this proposal, while approved, requires benchmarking figures that will be presented at the next core developers meeting.

Alternatively, at least two other proposed EIPs look slated for delay.

Developer Rick Dudley explained that EIP 1559 – which introduces a new transaction fee model to ethereum – is “a pretty complicated change.”

Dudley further highlighted that it would most likely not be ready in time for Istanbul, which is scheduled for mainnet activation possibly as early as mid-October.

“[EIP 1559] we should assume that it’s possible that it will make it in [to Istanbul] but it seems unlikely right now,” said Dudley on the call.

The second EIP with a high potential for delay is EIP 1057. It is a proposed change to ethereum’s proof-of-work (PoW) mining algorithm, which since April of last year has been susceptible to mining by specialized computer devices called ASICs. With an estimated $655 million annual market for ethereum’s mining rewards, ASICs outperform graphics cards, or GPUs, which developers worry may lead to a more centralized mining landscape.

EIP 1057

proposes a revamped PoW algorithm known as “Progressive PoW” or ProgPoW in efforts to better leverage GPU-specific computing capabilities.

While approved twice in the last year by ethereum core developers, ProgPoW according to Jameson may face delay due to various logistical issues in organizing a third-party audit of the proposal.

“We ran into issues starting the ProgPoW audit,” explained Jameson in a Ethereum Magicians post yesterday. “We had a hardware partner who specialized in ASICs who was going to work with Least Authority to perform the hardware parts of the audit. They are no longer participating in the audit so we are looking for other auditors for the hardware portion.”

As such, Jameson proposed today that the EIP be held back from being in the approved category of EIPs until further details about the pending audit are sorted.

Looking ahead

The next official deadline for the Istanbul hard fork is merging accepted EIPs into existing versions of ethereum software called clients.

One EIP author, James Hancock, told CoinDesk that this step is akin to getting your code together so it can be fully tested.

"The suggestion is to have reference implementations in two 'major' clients," said Hancock to CoinDesk. "The definition of major is pretty loose."

Hancock also noted that he has put together an updated spreadsheet with all of the proposed Istanbul EIPs and their relative "readiness" for mainnet activation.

For now, the upcoming “soft deadline for major client implementations” is sometime in mid-July with an eventual mainnet launch slated for mid-October.

However, the envisioned timeline for Istanbul is a rather new creation that has never been replicated by previous ethereum hard forks. It was proposed by former ethereum developer Afri Schoedon and Ethereum Foundation developer Alex Beregszaszi as a way of breaking down hard fork process into “a fixed 9-month cycle.”

As such, Ethereum Foundation grant recipient Alexey Akhunov wrote in the Gitter chatroom that everyone should be thinking and iterating upon the new suggested “deadlines.”

“I myself will be questioning all the deadlines from the point of view of 'what is the purpose of this deadline?'," said Akhunov. “Because this is the first time lots of these things are introduced, we are here to make sure that what we do is done for the reason and not because "someone says so"

For now, blockchain protocol engineer at Consensys Danno Ferrin affirms that at the very least, the list of proposed Istanbul EIPs “stops growing” and will in all likelihood begin shrinking.

And down the road, the software upgrade itself must be accepted by the nodes that underpin the ethereum network itself when the hard fork event actually occurs.

Ethereum image via Shutterstock