Plasma is not dead — Part 1(EN)

철학자(정순형)

Published in

Tokamak Network

12 min readMay 11, 2020

Written by Team Tokamak Network

TL;DR

Intro

Plasma:: Scalable Autonomous Smart Contracts, an Ethereum-based layer 2 scalability solution written by Vitalik Buterin and Joseph Poon in August 2017, attracted the community, raising expectations that it will be able to fully solve their long-cherished blockchain trilemma. However, research and development and discussion of the Ethereum community for about two years is being gathered in the direction that plasma is not the ultimate means of solving the trilemma. Blockchain researcher Ashwin Ramachandran wrote in an article titled “The Life and Death of Plasma,” saying, “The Ethereum community was in limbo. While new Plasma constructions continued to emerge and marginally improved on their predecessors, the Ethereum community failed to rally behind any of them.” adding, “A new solution, such as Optimistic or zk-Rollup, is expected as a better scalable solution and “It seemed that Plasma was dead.”

The emergence of new technologies in any field inflates people’s expectations for our rosy future that this technology will change. This expectation leads to huge investment and a large number of experts to develop technology and to pour out examples of application. At this time, some eloquent researchers use writings and lectures to expose the reality of technology to the public. And the public spills out disappointment, frustration, even anger and disillusionment, realizing that the bare face of the technology they unveiled is not as beautiful as expected. At this stage, many researchers leave the field without enduring public cynicism. However, there may be cases that are fully re-evaluated later, such as the artificial neural network of Jeffrey Hinton, or Wei Dai’s b-money, or at least find useful uses in certain areas.

From this point of view, we would like to argue that Plasma has not yet died and can still be usefully used.

Plasma was dead, it seemed to be..

Plasma, which was expected to be able to fully resolve the blockchain trilemma, a long-cherished desire of blockchain community, got a lot of attention from the community. However, after about two years of research and development and discussion, Plasma came to the conclusion that it was not a deus ex machina that would simply solve the trilemma, and the interest from the community has gradually decreased.

Interest in plasma is gradually declining.

The reason why the community’s interest in plasma has decreased is because the following three problems have not been resolved.

Data Unavailability
Mass Exit
Monitoring Issue

Strictly speaking, number two and three can be seen as a form of complications, caused by number one. Fundamentally, all problems are due to data unavailability. Data unavailability in Plasma is a situation in which the operator of plasma withholds the user’s assets or states data, making it impossible for users to access the data. In this case, the plasma operator will gain the power to manipulate user’s data unilaterally without any checks. Therefore, Plasma introduced an measure called Exit or Exit Game to prevent this problem. As long as this works, honest users will be able to safely exit from the invalid plasma chain whenever the plasma operator withholds the data.

Security of Plasma eventually depends on how efficiently and effectively this exits could be done, and many researchers have tried to improve and develop it.

Unfortunately, many of the proposed solutions did not go far beyond the starting point of the discussion. In particular, most exit game required for all users to watch all data processed by the operator (3. Monitoring Issue) and to exit within the time limit (2. Mass Exit). These two were fatal in terms of the usability of Plasma. If all users did not monitor the operator every moment, there was a risk that some users who attacked with data withholding could not exit, and that the exit itself might not be handled because many users had to simultaneously send exit transactions in a short period of time even if everyone could monitor (and this also depends on the scalability of Ethereum itself, Layer-1).

As a result, the Ethereum community began to explore other possibilities other than Plasma. That’s how it appeared: Rollup, or Optimistic Rollup. In fact, Rollup is not a very new story. It’s just a variation of the shadow chain that was mentioned a few years ago. Nevertheless, there is a reason why Rollup is getting attention again in recent years.

We can get away from the data unavailability that we were sick and tired of.

Rollup uploads all transactions data that executed in layer-2 for verification to layer-1 (not directly stored in storage of layer-1, make anyone to check those transactions by CALLDATA). This makes data of Rollup is always fully available. This means no complex exit games are needed, and anyone can verify the state transition of the layer-2 chain at any time.

Although Rollup significantly limits scalability compared to Plasma and pay more expensive costs, elimination of data unavailability has made it safe for users to use Layer-2 without having to continue monitoring layer-2 or exit within the time limit.

So far, Plasma has failed, and Rollup seems like a successful tool.

But is it really true? Let’s talk about other story for a moment.

Homi vs Excavator

Sometimes appropriate questions help you understand the nature of things and phenomena.

Which is more useful tool, Homi or excavator?

(Note: Homi is a traditional small farming tool in Korea.)

The answer is ‘We can’t tell which is more useful.’

If you didn’t get the answer, at least you should have asked again “Exactly what aspects are you evaluating for usefulness?”. (Readers who ask themselves like this are as good as correct answers!)

At first glance, the excavator itself looks far superior to homi. The excavator, a collection of numerous mechanical devices, can be seen as the definitive technology of modern engineering. Homi, on the other hand, is shabby in shape and performance. However, if a specific situation is given to this question, the answer will be overturned immediately.

Which is more useful, homi or excavator (to grow small garden in house)?

It is not easy to plant small seedlings with an excavator. It is clear that homi is more useful in gardening than excavators.

However, excavators can be much more useful than homi in situations where large buildings are to be built. It is not easy to mine soil with homi, which weighs hundreds of kilograms. Therefore, there is no room for homi at the civil engineering site. On the other hand, excavators are a must-have tool at there. From this point of view, excavators are more useful than homi.

The preceding questions and answers imply that it is hard to say which tool is absolutely useful or superior. The important point here is to recognize that the advantages and disadvantages of each tool may or may not be useful in certain situations. We should not fall into the trap of judging the usefulness of a tool on a single point.

With this in mind, let’s go back to Plasma and Rollup, the tools we designed for Ethereum’s scalability.

Plasma vs (Optimistic) Rollup

Previously, we mainly discussed the shortcomings of Plasma, but in fact Plasma also has its own powerful strengths.

The biggest advantage of Plasma is scalability (which can be achieved without much sacrifice of decentralization and security).

Plasma is a proposed framework for incentivized and enforced execution of smart contracts which is scalable to a significant amount of state updates per second (potentially billions) enabling the blockchain to be able to represent a significant amount of decentralized financial applications worldwide. — Plasma Whitepaper

The Plasma white paper emphasizes this from the beginning, which is clear to be the most scalable of all layer-2 solutions that exist. This is a great advantage, because the scalability issue, a chronic disease of all blockchain, including Ethereum, can be resolved by Plasma without sacrificing decentralization and security — even though certain conditions are required.

Let’s compare this to a Rollup. As mentioned earlier, the Rollup uploads all transaction data to the root chain, unlike Plasma, so that anyone can check it. Thus, in Rollup, with only one honest verifier, all the remaining users can safely use the layer-2 chain. Unlike Plasma, Ethereum-level security can be guaranteed without additional conditions.

However, the Rollup eliminated all of Plasma’s key shortcomings, while at the same time sacrificing its advantage, scalability. Rollup uploads all transaction data to calldata in transactions when storing state roots in layer-1 to ensure data availability of layer-2. Specifically, the current cost of using calldata is 16 gas per 1 byte(non-zero). Based on this, the gas required for the transfer of ERC-20 token is about 1,257 gas. Although the cost has been significantly reduced compared to the past, the fact that each transaction still requires a certain level of commission has not changed. (Of course, it is clear that the transaction cost has decreased significantly from the existing ~50000 gas.)

As a result, when using rollups in terms of scalability, TPSs can be increased at least 4 to 50 times more than Ethereum. The problem is that this is the upper limit. No matter how many Rollup chains are used, the results are the same. No matter how optimized the Rollup is, scalability of layer-1 cannot be increased exponentially like plasma. If s is Ethereum’s scalability, Rollup can only increase scalability as much as n*s, which is the constant multiple of s. This is because the increase in scalability of Rollup depends linearly on the scalability of layer-1.

Plasma, on the other hand, can increase scalability to s^m, which is the exponential, rather than the constant multiple of s. The important point at this time is that m can be a variable, not a constant. Because the root of plasma blocks that are stored in the root chain can contain countless transactions, even can store hundreds of blocks as a single root. If the scalability of the Plasma is the ocean, the scalability of the Rollup is a lake.

As mentioned earlier, however, additional conditions are needed to ensure security in Plasma. On the other hand, Rollup is clearly superior to Plasma in terms of security because it does not require those complex conditions.

Plasma, Rollup, and Blockchain Trilemma

Based on the previous discussion, a simple comparison of Plasma and Rollup with other solutions in terms of blockchain trilemma, can be expressed as follows.

Comparison of each solution in terms of the trilemma

(Note: There may be differences in the specific assessment of each solution in relation to the figure above. But the important thing is to figure out approximately where Plasma and Rollup are on this graph.)

Rollup can provide security and decentralization near Ethereum levels, while also providing improved scalability over Ethereum. On the contrary, Plasma is able to provide higher scalability than consortium blockchain represented by EOS, depending on the situation, instead of sacrificing security compared to Rollup.

In fact, the initial community’s expectation for Plasma was that fully resolved the blockchain trilemma, as shown in the figure below, but it actually fell short of that expectation. Using the analogy, the early Plasma was not an all-around tool for both large-scale civil engineering and small seedling.

Dream of the Ethereum community — Unfortunately, the dream didn’t come true.

However, that does not mean that Plasma’s unique strengths have disappeared. Just as both homi and excavator have its own advantages, the fact that Plasma is a layer-2 solution with enormous scalability does not change. However, even before we get to use the scalability of the tool properly, we put it in a coffin and buried it.

Now let’s get this tool back out and find out where it’s used.

Plasma and Optimistic Rollup

The pros and cons of Plasma and Rollup contrast sharply with each other. Scalability, the strength of Plasma, is the disadvantage of Rollup, and security, the strength of Rollup, is the disadvantage of Plasma.

This is in fact similar to the previous comparison of homi and excavator. Homi is specialized in small gardening, but on the contrary, it is not possible to be used in large-scale construction sites. On the contrary, excavators are virtually impossible to use for small garden landscaping, but are essential tools for most civil engineering works. Plasma and Rollup can be used for that purpose, just as we use homi or excavator where we need each of them in real life.

For example, suppose you have a decentralized Reddit. High scalability will have to be guaranteed for countless sub-reddits, posts and comments to be created and stored. High levels of decentralization should also be ensured so that certain entities cannot censor and block social activities of users. In addition to this, security should also be ensured to prevent posts or comments from being forged or deleted arbitrarily. Although none of the components is unimportant, the necessary priorities based on the importance of each component can be summarized as scalability > decentralization > security. This is because although decentralization and security are of course important, the service itself is impossible to operate unless scalability is guaranteed.

On the contrary, consider a decentralized exchange(DEX). High scalability must also be ensured, as countless transactions occur simultaneously. Decentralization is also essential to prevent certain users from censoring transactions or the sudden shutdown of exchanges by third parties. But the most important thing is security. A situation in which a user’s assets are taken over by an attacker is that must be prevented at all costs. (This is why the initial Bitcoin and Ethereum sacrificed scalability for security and decentralization) Therefore, the priority of each factor will be security > decentralization > scalability.

The priorities of each element required to operate DReddit and DEX can be briefly illustrated in the following figure.

What we need to do here is to choose the most appropriate solution to meet this requirement to the fullest.

For DEX, where security is more important than any other factor, it seems more appropriate to operate in Rollup rather than plasma. Conversely, in the case of DReddit, scalability takes precedence over security, so Plasma is a more appropriate solution than rollup.

Of course, if you want to handle more transactions, even at some cost of security, DEX can also be operated in Plasma, while DReddit can also be operated in Rollup if you sacrifice scalability and prioritize security. The key is that both Plasma and Rollup have different strengths and weaknesses, and that the appropriate solution should be selected based on the priority of requirements.

In addition to the examples used here, DReddit and DEX, there are already lots of existing DApps and even whole new ones can come out. Each of these DApps may have different priorities for each requirement in terms of scalability, security and decentralization. Plasma and Rollup can perform their respective roles admirably to meet the needs of various DApps, and even more diverse DApps can be created that have never been imagined in Ethereum through a powerful Ethereum community based on its strong interoperability with Ethereum.

Using two effective solutions, Plasma and Rollup, as shown above, various DApps (purple and orange areas) that were not possible with existing Ethereum can be brought to the Ethereum community.

This is not the end. It is also possible to use Plasma and Rollup at the same time. To be exact, it is possible to select and use a solution of either plasma or rollup depending on the timing and purpose, and to change to other in the middle. (How it specifically works will be discussed in Part 2. However, this means selecting and using a solution between plasma and rollup, not the ultimate solution that combines the advantages of the two solutions. )

We would like to serve a means to use both tools of Plasma and Rollup flexibly for the purpose. We are sure that this type of architecture will certainly be of great help to DApp developers and users who choose a layer-2 solution for the needs of various spectra (scalability, usability and security) that Ethereum does not meet.

Conclusion

Not all layer-2 solutions that have emerged, at least so far, have been a means to fully overcome the blockchain trilemma. Perhaps this trilemma is a challenge that cannot be solved perfectly in the future. However, at least we would like to introduce a partial optimization scheme for the trilemma to maximize the utilization of each means. This will ultimately maximize the freedom of choice of developers who use blockchain to create applications to meet market requirements, further expanding the use of blockchain.

And let’s talk more about the technical details in Part 2.