“The Ethereum blockchain is almost full,” said Vitalik Buterin in an interview with Canadian newspaper The Star earlier last month. He told no lie; Ethereum’s network utilization has been ominously bubbling north of 90 percent over the past couple months, according to tracker Etherscan.io.
That’s a big problem—if there’s no more space on the Ethereum blockchain, the network could become too slow for people to build on it, or force people to rely on solutions that compromise on security or decentralization by, for instance, delegating certain processes to trusted third parties. At that point, why even develop on blockchain at all?
One solution to Ethereum’s scaling problem is “plasma”, essentially Ethereum's version of Bitcoin's Lightning network. The general idea, proposed by Ethereum co-founder Vitalik Buterin and Joseph Poon back in 2017, is to create a layer of super-efficient “plasma chains” on the Ethereum network that only communicate with the bottom layer of the Ethereum blockchain when absolutely necessary. Moreover, these plasma chains are to be non-custodial, meaning that if a plasma chain is attacked, then the hacker won’t be able to damage the rest of the network.
Developers have rushed to the fray to support the good ship Ethereum. But implementations of Plasma are so varied, so niche, and so full of boing-boing marketing spiel, that it’s difficult for the Ethereum community to keep track of what’s really going on. So, what has happened to plasma, and can it save Ethereum?
It turns out that Plasma Group, a community of researchers funded by backers including the Ethereum Foundation and Consensys (which funds an editorially independent Decrypt), has been hard at work. Back in July, the Plasma Group released its grand project, the Optimistic Virtual Machine (OVM), which brings all of the many different layer 2 solutions together.
“We realized throughout doing all this plasma research that most layer 2 constructions actually have a lot of overlapping architecture,” Jinglan Wang, executive director at Plasma Group, told Decrypt. “So rather than reinventing the wheel each time, we thought it made sense to create an abstraction, formalize it in logic, and provide really clean shared infrastructure and security models for everyone to use,” she said.
“One of the reasons I think [the OVM is] dope from a usability standpoint is that wallets wouldn't have to integrate five thousand different scaling solutions. They would just have to integrate the OVM, and through that be able to access any layer 2 construction that's been written in the OVM.”
Over the past nine months, the company has developed proof-of-concepts for three different scaling solutions on the OVM: plasma, state channels, and optimistic rollup. Teams like Cryptoeconomics Lab, Matic, and OmiseGo have already begun implementing systems using the technical blueprints drawn up by Plasma Group.
Today at Devcon, the Ethereum Foundation’s annual conference, Plasma Group launched for the first time a working demonstration for one of those proof-of-concepts, Optimistic Rollup, an Ethereum scaling solution that runs autonomous smart contracts on layer 2.
The demo is a partnership with the decentralized exchange protocol Uniswap. It lets you swap “UNI” tokens into “PIGI” tokens, Crucially, these transactions are instant and don't require gas, the currency that powers the Ethereum network, meaning applications using Optimistic Rollup won’t clog up the Ethereum network. "Ethereum can scale today,” reads a message on the demo’s website.
“It feels just as fast as any other normal centralized application that you’re accustomed to using. Which is actually very sexy to a scaling company like us, and a DeFi app like Uniswap,” said Wang.
Hayden Adams, CEO of Uniswap, told Decrypt that the achievement here is that Optimistic Rollup manages to accomplish all of this without using sidechains, which are more centralized, and without sacrificing the ability to create complex applications. That’s important for Adams’ Uniswap, which requires a lot of different parties to interact with each other.
Make no mistake—Optimistic Rollup is not Plasma. “Plasma and Optimistic Rollup are similar in that both of their scalability comes from the fact that everything is executed “optimistically”. The meme is that you don’t go to court (Ethereum) every time someone parks their car - only if something goes wrong,” said Adams.
So what’s the difference? If an invalid transaction is made on a smart contract built on a plasma chain, the data you need to dispute invalid transactions is held by the plasma operator. The plasma operator, if they so wished, could withhold that data; “For applications like Uniswap this makes dispute resolution extremely difficult,” said Adams.
But with Optimistic Rollup, all data needed to resolve disputes has to be posted to the Ethereum main chain. It’s very cheap to post the dispute data, and data can’t be withheld by operators, preventing forced mass exits.
The tradeoff? Optimistic Rollup doesn’t completely solve Ethereum’s scaling problem: the Devcon demo of Optimistic Rollup allows for 250 TPS, and, when optimised, it will reach 2000 transactions per second. That may be a hell of a lot faster than Ethereum’s current rate of 15 transactions per second, but it’s no plasma. Plasma, said Adams, offers near-infinite scaling opportunities.
Yet the team argue this is an acceptable compromise, considering Optimistic Rollup’s benefits. “Optimistic Rollup is the middle ground,” said Adams. “It doesn't have as high of a theoretical throughput as other layer two solutions but allows us to do anything you can do on Ethereum while still achieving a pretty high throughput.” He said that although a single two-party state channel could technically do thousands of transactions per second, those transactions wouldn’t be able to be very complicated, such as “simple transfers between those two parties.”
“The decentralization and security vulnerabilities in sidechains are well documented, but for apps where decentralization is not a priority, sidechains these days are wonderfully easy to build apps on,” said Plasma Group’s Wang. But, “If an app wants to be censorship-resistant and truly decentralized, then the security model matters. So gaming companies might be okay with sidechains, but financial apps should use Optimistic Rollup.”
Wang said that in any case, the importance of TPS is overstated. “We're testing the hypothesis that people are actually ok with lower TPS, so long as they can have scalable computation,” she said. "2000 TPS is obviously not as sexy as 20,000 TPS, but honestly, who needs 20,000 TPS right now?" said Wang from Plasma. “Optimistic Rollup on the OVM scales Ethereum more than enough for the needs of dapps today.”
And though Optimistic Rollup doesn’t offer the transactions per second that plasma does, “it's a promotion for the developer experience,” said Wang. “It's pretty difficult to build an application on plasma—we know because we pioneered it with Generalized Plasma Predicates.”
So, what, no more plasma? “We're not abandoning plasma, but we do think that it's an early over-optimization for the apps we see in use today,” said Wang. “Even though we're called Plasma Group, our primary desire is to scale Ethereum securely—there are many ways to do that.”