Partnered With:

Multicoin Capital: How does technical scalability create social scalability

Read this article in 31 Minutes

This paper assumes that the most important function of the blockchain is not a non-sovereign currency per se, but rather a DeFi.

原文标题：《 Technical Scalability Creates Social Scalability 》

Kyle Samani is co-founder of Multicoin Capital

Original compilation: 0x33, beat Blockbeats

Editor's note: This article assumes that you are familiar with Nick Szabo's article. Social Scalability, Vitalik Buterin's paper "Weak Subjectivity", and Haseeb Qureshi's paper "Why Decentralization Isn't as Important as You Think".

This article is not intended as a rebuttal to Szabo's paper.

Szabo concludes his paper by defining social scalability as "sacrificing computational efficiency and scalability -- consuming cheaper computing resources -- in order to reduce and better utilize the massive human resource expenditures required to maintain relationships between relationships involving modern institutions, such as markets, large corporations, and governments."

Bitcoin's victory over BCH and BSV certainly supports this theory.

However, crypto ecosystems have come a long way since Szabo published his seminal paper in February 2017. Although Szabo came up with the concept of smart contracts some 20 years ago, it is only in the last 24 months that the world has begun to appreciate the most useful application of smart contracts: Defi.

Defi changed everything.

This paper assumes that the most important function of the blockchain is not a non-sovereign currency per se, but rather a DeFi.

Changing the primary functionality of the system from strictly censorless, non-sovereign money to a powerful and programmable development environment for financial applications should have an impact on every layer of the technology stack in the network layer (e.g. gossip With & have spent Turbine)) through the execution environment (e.g. EVM With & have spent SeaLevel)). (For a brief overview of the differences between Solana and Ethereum, see the appendix to this article.)

Therefore, the blockchain should first be designed and managed as the DEFI development platform.

Predictability is power

"The question I often hear is, 'What's going to change in the next 10 years? '& NBSP; it's a very interesting question; it's a very common question, and yet I'm almost never asked: "What's not going to change in the next 10 years?" & NBSP; I'm here to say that the second question is actually the more important of the two, because you can build your business strategy around things that are time stable...

[In our retail business, we know customers want low prices, and I know that will be the case 10 years from now. They want fast delivery; they want a lot of choice. It's impossible to imagine a customer 10 years from now coming out and saying, "Jeff, I love Amazon; I just want the prices to be higher," or "I love Amazon; I just want your delivery to be slower. 'No way. So we put our energy into those things and break them down. We know that the effort we put into it today will still pay off for our customers 10 years from now. When you know something is right, even in the long run, you can afford a lot of energy."

-- Jeff Bezos, founder and CEO of Amazon & NBSP;

Coinbase has about 50 million registered users, as does Robinhood. So did most of the major U.S. banks.

Let's assume that Coinbase's strategic imperative is to migrate 100% of its users to DEFI as soon as possible and that the regulatory environment supports this. How does Coinbase do this on Ethereum today?

Right now, that's an unanswerable question. That's not to say it's technically impossible, it's possible. But there is no single person or organization in the world today that can answer that question. Why is that?

Because no one really knows how Ethereum will scale. To take a simple example, Vitalik stated that he believes that in the near to medium term, optimistic rollups may be the best expansion solution, while in the long term, ZK aggregation will dominate. When and how does this happen? What infrastructure needs to be rebuilt? How will the flow of money between these different rollups affect smart contract developers, wallets, users, liquidity providers, cryptocurrencies, etc.?

Also, it seems that no matter how important the extension solution may eventually become, it is unlikely that the extension solution will have a single global instantiation (meaning, for example, a single Optimistic Rollup). The future of Ethereum expansion will be heterogeneous.

In the long run, this could be a good thing for Ethereum. Each of the various extension solutions has some tradeoffs, and it is not clear which set of tradeoffs is the best or the best way to combine these extension solutions. So, in the long run, the best option for the Ethereum ecosystem is to try multiple scaling solutions, figure out which solution is best for which application, and then figure out bridging and other interoperability solutions and latency issues.

In addition, all the teams building the extension solution are well funded and are already online and user-facing. As a result, none of the teams will give up anytime soon.

So how does Coinbase get over 50 million users to join Defi?

At this scale, the most important thing to optimize is certainty. Companies of any size will demand certainty now and in the future.

Large organizations simply cannot bet on the wrong technology stack. The opportunity cost of error and the explicit cost of later migration/bridging are enormous.

I think the only blockchain protocol that can answer that question or will answer that question in the next 24 months is Solana.

All Rollup-based extension solutions suffer from the same problems I outlined above. The same is true of sharding. Despite billions of dollars of research and development from a range of capable teams (Cosmos, Polkadot, Avalanche, etc.), none of the sharding systems actually reached a meaningful scale (most didn't even work in their daily output). Even if they start with POC (proof-of-concept ) Consensus work, and there will also be many emerging issues that must be managed (for example, cross-shard transaction failure, transaction platform integration, etc.).

To be clear, I'm not saying that scaling through sharding and rollup won't work. In fact, I'm ultimately optimistic about both solutions. However, neither of these scaling strategies really works today and creates a number of secondary and tertiary issues that must be addressed. This makes it hard to see an impartial organization that desperately needs scalability, certainty, and can get what it needs in the next 24 months, because there are so many intertwined components to scale Ethereum.

Fragmentation of social scalability costs

In addition to the uncertainties outlined above, spreading the application across shards and rollups creates significant new social coordination costs that do not exist in a single-shard system. Some examples:

Block time and computational throughput vary between Layer 1 and each Layer 1. This is directly related to how risk should be managed in all of the DEFI protocols that manage risk, which includes almost all of the major primitives except Uniswap/Sushiswap. The DEFI team has committed to deploying contracts on multiple Layer 1 and Layer 2. However, each execution environment will require unique risk parameters. This will increase the amount of social coordination required for each agreement community and slow down the pace of development.

From the optimistic rollups (ORU) exits take a long time. Market makers are widely expected to bridge the liquidity gap between Rollup and Layer 1. However, the implementation details of this operation are tricky. Should the protocol front end be provided locally? If so, should they "sign up" to specific market makers (for example, how Citadel Securities signed the PFOF contract with Robinhood)? Or should the front end let the user figure it out? What if a user wants to move from one ORU to another... How can the user tell the application so that the user can use Connext or Thorchain instead of exiting to Layer 1?

For MetaMask users, most of whom are power users, it may make sense for the users to manage this complexity themselves. But for a select wallet (such as Exodus or Arent) that tries to abstract away these complexities, how much additional development time will the wallet team spend addressing these issues? How much actual new feature development will be forgone? What if, for some reason, market makers stop providing liquidity in that particular bridge/segment? What are the backup options? The developer tools must be updated to handle the new data structures (unprocessed transactions for the ORU, ZK output for the ZKR). The indexing and query layers will require significant upgrades, and application developers may need to rewrite their subdiagrams to handle new data structures (for example, it is impossible to map EVM subdiagrams to StarkWare's Cairo).

Developers will be forced to rewrite a large number of applications across a variety of heterogeneous extension solutions.

As sharding and rollups proliferate, Devex will become even more challenging. None of these problems are intractable, but they can slow down development and frustrate many developers who don't want to solve all of them.

Predictable, boring expansion

Solana currently supports 50,000 transactions per second. The global network will soon exceed 600 nodes. But most importantly, Solana offers an infinitely expanding predictable path. Because it can execute transactions in parallel on the GPU, it can take advantage of the great benefits of parallelism provided by the GPU.

Moore's Law is probably the most important economic force of the past 50 years. But today it looks more like an illusion.

About 10 to 15 years ago, Moore's Law stopped the work of single-threaded computing. Because heat is generated every time the clock's speed increases superlinearly. This is why sector devices (desktops and laptops) stagnate at around 3.5-4 GHz, while fanless devices (phones and tablets) stagnate at 1.5-2.0GHz. Although various optimizations have improved single-threaded performance over the past decade, single-threaded performance has not doubled every 18-24 months.

Over the past decade, almost all of the computing gains have come from chip specializations (FPGAs and ASICs) and parallel architectures. Modern desktop graphics cards have more than 4000 cores. The number of cores per card has been increasing in a Moore's Law manner for the past decade, and this trend will continue, as the heat generated by increasing the number of cores will have little of the same effect as that generated by increasing the clock speed.

Solana is the only blockchain that supports intra-shard parallelism through its native support. SeaLevel. Sealevel performs transactions locally on the GPU. When Nvidia releases a new GPU with 8,000 cores in 12 to 24 months, the computing bandwidth of the Solana network will roughly double. Developers don't know, don't care, and don't have to change a line of code.

This is the definition of predictability: write code now, knowing that it will last forever, and knowing that the code will cost less to execute tomorrow than it does today.

The main physical limitation of scaling calculation is heat dissipation. Solana literally expands within physical limits.

Quantified decentralization and reduced subjectivity

On the face of it, many people think the Solana protocol is not decentralized enough. They don't really quantify the claim, but they repeat it anyway. Let's quantify what we can quantify.

I. Hardware cost:

When Bitcoin runs on a $25 Raspberry Pi, it requires a weak Internet connection.

Ethereum runs on $500 laptops (although that's questionable given current Gas limits) and requires a broadband connection.

Solana runs on servers that cost around $3,500, assuming a gigabit connection.

The final major consideration is state size. With 50,000 TPS and billions of users, Solana's scale will explode.

That's good. Why? because

1) Assuming Solana is running on a server with upgradable storage (not a non-upgradable laptop),

2) NVME SSD extends read and write performance through RAID 0 linearly, as well

3) Multi-TB NVME SSD for less than $300. The cost and logistical costs of scaling the state store are negligible.

If you know this well, and everything I've written so far, the odds are better than 50% that you're using a MacBook Pro larger than $2,000 +, which is the high-end computer preferred by 50 to 100 million or so developers around the world. I suspect that tuning for $500- $1,000 hardware is optimal. What's so special about the $500- $1,000 price point?

The upper limit for considering hardware requirements should be fair. $25,000 must be too high, since developers don't own the hardware. Instead of specifying arbitrary hardware costs, a better way to think about this is to think about how many nodes are sufficient to achieve adequate censored capability. Obviously, the word "enough" is inherently subjective, but if you assume that 1M is the correct number, then the natural question is: "Are there enough $3,500 servers in the world connected to gigabits so that there is a reasonable path to the 1M node.?"

Given the number of gamers, developers, and companies in the world with high-end hardware, it's hard to see why the answer is no.

Hardware cost cannot be considered in isolation. It must be considered within the framework of the system's design goals.

Earlier, I thought the blockchain should be optimized for DEFI, not for the biggest censorship resistance (which means optimizing for 100M or 1B nodes instead of 1M nodes). There's no need to optimize for $25 or even $500 hardware, because the vast majority of people will never run a node. So why bother to optimize hardware costs and protocols for them?

Everything is subjective

This will result in The subjectivity is weak. And realize that decentralization isn't as important as you might think.

The world is subjective. What does that mean? Consider the following example.

When you last walked into a high-rise building, did you first inspect the building and interview the main contractor to make sure it didn't collapse and kill you?

Or put the same concern on a plane, a car, or your house.

Everything in the world is based on some degree of trust. The world wouldn't work if everyone had to independently verify the structural integrity of everything they interacted with.

Instead, it can be stated that the world works because everyone knows that enough other people have tested the system's assumptions to keep the system safe with a high probability.

This is the basic assumption of less subjectivity. When applied to node counting, the key question is, even if you are not running your own node, can you reasonably assume that there are enough other people and institutions running the node to trust the system?

Today, Solana has about 600 nodes, up from about 100 a year ago. Like almost all blockchains, this number has grown over time. As long as the ecosystem continues to grow, there is no reason to believe that the number of nodes will decline. Like all other major blockchains, the Solana network naturally decentralizes over time as more people use it and more value flows through it.

That's why Qureshi is Right, and decentralization isn't as important as you might think. Decentralization is crucial to achieving censorship resistance. It's not clear where that threshold is (there's not enough counterfactual to infer from that fact), but the actual threshold itself is not important. It is important to:

1) The shape of the curve is actually a reverse S-curve, as well

2) Know that blockchain has become more decentralized over time, thus increasing the resistance to censorship. As long as the chain decentralizes at a sufficient rate (and you believe it can maintain that rate), users are likely to maintain the censorship resistance they need.

Image Credit: Haseeb Qureshi, Why isn't decentralization as important as you think

conclusion

The basic question at hand is: what is a dog and what is a tail?

In the first decade since Bitcoin came out in 2009, it was clear that the censored, non-sovereign currency was the dog, and other applications were the tail.

But that is changing. Defi is now going to completely reshape finance for everyone who uses cryptocurrencies. I think the roles have been reversed: the DeFi is the dog, not the sovereign currency is the tail.

Both require some degree of censorship resistance, but given the engineering constraints, there is a fundamental trade-off between maximizing the utility of DEFI and maximizing the censorship resistance of the system.

When the basic design goals of the system change, the technology stack should change with it. To scale DEFI to billions of people, each layer of the stack must be reconsidered from first principles.