Ethereum: Why Constantinople upgrade delayed until 2019

The roll out of Constantinople on the Ropsten testnet (which mimics the real Ethereum network and is used for trialing upgrades before they go live) has encountered unforeseen troubles including communication issues, technical problems, and differences between two of the client implementations. As a result Ethereum developers have opted to postpone the upgrade until 2019.

When the block intended to trigger the fork was hit — number 4,230,000 — the network stalled, and ultimately failed two hours later as it was revealed there were no miners working on the new chain.

Although this problem was eventually fixed by volunteers upgrading their software to the new Constantinople client, the episode brought to light several problems that are now holding the live launch back.

A three-way fork

These problems were discussed in a post-mortem developers meeting last Friday, where the Ethereum Foundation revealed that recently added hashpower had caused the hard fork to happen two days earlier than expected, and as Parity had only released their updated client one day before, not all of the miners had been able to upgrade in time.

This added hashpower also meant that the fork was attempted during the weekend — a terrible time for testnet upgrades, according to Parity release manager Afri Schoedon, as there were very few people monitoring the network.

Better communication, and improved coordination between clients and miners, were highlighted as priorities moving forward. But the key technicality holding the network back was described as a "consensus issue", in a tweet immediately following the event from Infura:

This was later expanded on in a volley of tweets from Parity developer Afri Schoeden, who described the "consensus issue" as a "three-way fork between Geth and Parity", suggesting that the issue lay between the two clients working towards the upgrade — which both offer different ways of interacting with the Ethereum network.

The meeting — which was live streamed as usual on Youtube — concluded that to allow time for these issues to work themselves out, the Constantinople upgrade should be postponed to early 2019. "I keep getting the feeling that we’re trying to rush this and I would second that we should breathe and see what happens."_ _said Parity developer Afri Schoeden.

ProgPow in Constantinople

A crucial stepping stone towards the anticipated Casper upgrade, the Constantinople hard fork consists of five improvement proposals that will increase the network’s efficiency, lay the foundation for new scaling solutions, and set the stage for a seamless transition to Proof-of-Stake.

Since the changes won’t be backwards-compatible, the new network is to be forked, an event which some are now suggesting provides the perfect opportunity for another improvement to be implemented. Martin Holste Swende, security lead at the Ethereum Foundation, has suggested that there could now be enough time to add code for another proposal, dubbed "ProgPow," into Constantinople.

Programmatic Proof-of-Work (ProgPoW), is an algorithm proposed to minimize the benefit of specialized ASIC mining equipment, thus making mining more accessible to amateurs, and keeping the dangers of centralization at bay. This proposal was initially considered for inclusion in Istanbul, the hard fork set to follow Constantinople.

Although the switch to Proof-of-Stake will ultimately make concerns over centralization through mining irrelevant, timeframes for the Casper transition are still not set in stone. In the meantime, centralization still has the potential to undermine the trustless nature of the Ethereum network, and even influence future development decisions.

At the meeting, Martin Holste Swende suggested that the delay could allow for the early implementation of this additional safety measure into the Constantinople fork. "I’m going to go out on a limb here and say that if we do decide that Constantinople isn’t until January or February, then I would probably try to push for including ProgPoW into Constantinople."

Despite the delay, there is no indication of any deeper issues with the Constantinople code, and the Ropsten testnet has since completely recovered. Speaking at the Ethereum hackathon in San Francisco earlier this month, founder Vitalik Buterin said he was feeling optimistic, claiming that over the last four months, the research side of the project has "stabilized", leaving no "significant unsolved theoretical problem[s]" standing in the way of Ethereum.