Inscriptions, a form of digital collectible created by writing data into the calldata or witness fields of a blockchain transaction, have caused degraded performance or even crashes on several blockchain networks, including Arbitrum, zkSync, Avalanche and others. But Branden Farmer, co-founder of Polygon, claims that parallelized Ethereum Virtual Machines (EVMs) may solve the problem once and for all.
Inscriptions were first created on the Bitcoin network. After the Taproot upgrade in 2021, Bitcoin users discovered that the upgrade allowed them to embed data into the “witness” field of a transaction, allowing images, tokens and other types of digital collectibles to be minted on the Bitcoin blockchain. Some users saw the new collectibles as a benefit to the network, while others regarded them as “spam.”
But the craze didn’t stop with Bitcoin. Soon, inscription producers started minting them on Ethereum sidechains and layer-2s such as Arbitrum, Avalanche and Polygon. Instead of being written into the witness field of a Bitcoin transaction, this new type of inscription was written into the “calldata” field of an EVM-based network. Since calldata doesn’t get stored in a smart contract’s state, this allowed producers to create their collectibles at a fraction of the cost of minting a traditional nonfungible token (NFT).
However, this cheap method of minting eventually caused high fees and instability for several blockchain networks as users increasingly flooded the networks with inscription transactions. On Dec. 15, Ethereum layer-2 Abitrum went offline for over 70 minutes from inscription spam. During December, half a dozen other blockchains also experienced degraded performance or extremely high fees as a result of these mints.

In a conversation with Cointelegraph, Farmer claimed that parallelized EVMs may help to resolve this problem. In his view, parallelism allows unrelated transactions to be processed faster, increasing the throughput of blockchains and potentially allowing them to handle the spam.
“In general, the EVM is executed sequentially, so every transaction in a block must be executed in order,” Farmer stated. “It doesn’t matter if it’s an NFT transaction or a Uniswap transaction, two things that have no dependency on each other, we process those transactions sequentially.”
However, with a parallelized EVM, transactions that are not related can be processed simultaneously instead of sequentially. “At the same time, we could be processing your Uniswap transactions and your inscriptions and your NFT mints,” Farmer claimed.
This ability to simultaneously process two or more transactions can allow the network to “localize gas fees to areas that contend with each other.”
For example, if many different people are trying to use Uniswap at the same time, they may have to pay higher fees, since their transactions are related to each other and must be processed sequentially. However, other users who are trying to mint NFTs would be mostly unaffected by this surge in Uniswap usage. Conversely, if the network experiences a surge of inscription mints, this may result in a rise in fees for inscribers, leaving other users mostly unaffected.
Farmer stated that this “localized gas fee” feature has not yet been implemented on Polygon, but it is ultimately one of the goals.
Another benefit of parallelism is an overall increase in throughput, Farmer stated. The Polygon team has already implemented “Block-STM,” which is a “first step” toward parallelization. And even with this small step, they saw “a 1.6x improvement in performance” and “either [have] achieved or will soon achieve a 2x improvement.” This means allowing the nodes to sync to the network and process blocks two times faster, leading to a greater ability for the network to handle spam.According to Farmer, parallelization was originally proposed by Solana developers, but the idea is now being implemented on multiple blockchains to improve performance, including not only Polygon but also Aptos, Monad and others. However, the Ethereum ecosystem is taking a new approach to this concept by combining parallelization with the increase of block space through layer-2 ecosystems, including Polygon 2.0, he claimed.