Istanbul in Trouble: Ethereum’s Latest Hard Fork Causes Controversy

In the year or so since Istanbul was announced, several different camps have formed around the upgrade, with some believing that Istanbul is the first step towards a truly scaling Ethereum, while others predicted that it would break some smart contracts—and argued that this is unacceptable.

Istanbul Is an Upgrade to Some

The network upgrade, which went live as of block 9,069,000, made several changes to the way Ethereum works. Examples include several opcodes cost adjustments, enabling Ethereum and Zcash interoperability, enhancing overall performance, and partially readying the network for the later Proof-of-Stake (PoS) consensus change. In order to benefit from these changes, Ethereum miners and node operators must update their clients. The Istanbul fork implements six Ethereum Improvement Proposals (EIPs) as follows: EIP-152: Making BLAKE2b more cost-efficient, EIP-1108: Reducing alt_bn128 gas cost, EIP-1344: Implementing the ChainID opcode, EIP-1884: Increasing the gas cost for trie-size-dependent opcodes to prevent spam attacks, EIP-2028: Reducing the cost of transaction data gas, improving the efficiency of transactions using zero-knowledge SNARKs and STARKs, EIP-2200: Implementing net gas metering for SSTORE operations. Although these upgrades seem benign enough, they have drawn significant controversy in recent months, since Istanbul is expected to essentially break 680 Aragon smart contracts, since most of these have gas prices hardcoded into their base codes. If this occurs, this will severely hamper development using Aragon, and the team behind the open-source project will need a new motto, since “build unstoppable organizations” would no longer apply. The Kyber Network side chain implementation will also be affected, making token swaps for its users more expensive than prior to the hard fork. Not only this, but since EIP-1884 isn’t backward-compatible, many DApps using these repriced opcodes would have to be reworked or scrapped entirely.

Current Network Status

As of this writing, according to Ethernodes, nearly 15% of Ethereum nodes have not yet been marked as Istanbul ready. Although this signifies that the majority of the network has accepted the Istanbul upgrade, but there are concerns across social media that Ethereum is becoming a more centralized blockchain than it was initially designed to be since only a minority of nodes need to complete the upgrade to enforce the hard fork. Not only this, but concerns were intensified when Parity—a popular Ethereum client; announced that their users needed to apply an emergency patch to their clients before upgrading to the latest Istanbul fork. This was particularly significant because Parity represents nearly 23% of the network, a patch made with maligned intentions or botched codes could have created a chain split or opened the doors to double-spending.

I find the Ethereum community's dependence on Parity concerning. Yesterday's emergency release underscores a long line of poor quality control practices at Parity.

— Austin Roberts (@austinrobertsiv) December 6, 2019

Although Parity’s patch was designed to add the missing EIP-1344-transition for mainnet users, the fact that this patch was released at the last minute reflects poorly on the developers of what claims to be the “Fastest and most Advanced Ethereum Client”. Commenting on the incident, the CEO of OpenRelay said that an “emergency release underscores a long line of poor quality control practice of Parity,” and that the dependence of the Ethereum community on Parity was “concerning.” Despite the concerns surrounding Istanbul, several exchanges have already made the upgrade, including Binance, Coinbase, and Kraken, while several others, including Poloniex, have announced plans to support the hard fork. With many of the most popular Ethereum exchanges making the change, it appears that most Ethereum users will be forced to accept the hard fork, whether they like it or not.

---

Images are courtesy of Twitter, Shutterstock.