And why is Ethereum attracting the most developers?
2. Ethereum has better tools and infrastructure for DApp development than any other platform
Truffle. Infura. Web3.js. OpenZeppelin. Geth. Ganache. MetaMask. CryptoZombies. MyCrypto. Etherscan. ERC20 and ERC721.
These are tools (among many others) that different developer teams have tirelessly poured hundreds of thousands of hours into — and they’re free to use for any Ethereum developer who wants to build a DApp on Ethereum.
And this developer ecosystem is only growing as time goes on. Our team and at least a dozen others are in the process of building even more tools and infrastructure around Ethereum DApp development that will make Ethereum developers’ lives even easier in the future.
It’s Metcalfe’s law applied to developer infrastructure. The more developers building useful stuff, the easier (and more enjoyable) it becomes for new developers to build, and the effect compounds on itself.
If you’re a developer and you wanted to build a blockchain platform that replaces Ethereum, you would need to build equivalent versions of all of these tools on top of your own platform in order to compete with Ethereum’s ease-of-use for developers.
Or you could, you know… just build on top of Ethereum, and take advantage of all these things out of the box ?
But let’s go one level deeper.
Let’s talk about WHY developers would want to spend their time building these tools.
And here’s where we go really deep down the rabbit hole, and see where Ethereum really shines.
When it comes to Blockchains, there’s a fundamental law called The Scalability Trilemma. It’s kind of like a law of physics that says a blockchain can only have 2 of these 3 properties: decentralization, scalability, and security.
What that means is, given the same level of security, if you want to increase a blockchain’s scalability, you must sacrifice its decentralization.
Why is this true?
Due to the nature of blockchains, every single validator needs to run every single computation that occurs on the network to ensure its accuracy.
So if you want the network to be sufficiently decentralized by having thousands of users running validators, the max number of transactions per second needs to be capped by what the average user’s PC and network speed can handle.
On the other hand, if you wanted to have the fastest, most scalable blockchain EVER, you should do the following:
- Require that all validators be supercomputers
- Have as few validators as possible on the network, to decrease the number of connections per node
- Put all the validators in the same geographical area (country, data center) to decrease latency between nodes
Can you see why that would be a really bad idea for a blockchain?
And yet every project I’ve seen that boasts a very high number of transactions per second is quietly making this decentralization tradeoff — most of them are just not transparent about it to their users / investors.
Why decentralization matters
There’s an excellent pair of blog posts from Chris Dixon and Spencer Bogartwith the same title, “Why Decentralization Matters” — both highly recommended reads.
In his article, Chris Dixon makes a simple argument: Developers are a lot more incentivized to build on a platform that they know is NOT going to change the rules on them later down the line, taking away their audience and profits.
If you build your application on Facebook or Apple’s App Store, you have to trust that these platforms aren’t going to ban you in the future, block certain users from using your app or seeing your updates, or start charging you higher fees to keep accessing the same audience.
Ethereum, on the other hand, is what we call permissionless.
Ethereum can be used by anyone for any purpose, without needing permission from anyone.
NO ONE can stop you from uploading a piece of code into the Ethereum blockchain, and NO ONE can stop your users from executing it.
Let that sink in for a minute, because it’s extremely powerful.
For the first time in history, we have a platform that that no one can shut down and no one can censor — not governments, not mega-corporations with tons of money, not the Rothschilds or Bogdanoffs or Reptilians or whatever conspiracy theory you subscribe to.
If you build a DApp on Ethereum, no one can stop your users from anonymously accessing it.
If you buy some tokens or digital items that are stored on the network, you are GUARANTEED they will stay there forever and no one will be able to take them away from you.
And it’s these permissionless and censorship-resistant properties of decentralized blockchains that have enabled us to have, for the first time in Internet history, true ownership of digital items.
More centralized platforms don’t have this guarantee
Pretty much every platform that’s been touted as the “Ethereum Killer” has simply decided to trade off decentralization in favor of higher scalability, and advertise it as if it’s a feature.
And this tradeoff is tempting to make, because it seems like it’s what the market wants.
Users who don’t know better complain about high fees and slow transaction times — so we can’t really blame developers for trying to give the market what it thinks it wants.
In Spencer Bogart’s Why Decentralization Matters, he says:
It’s not surprising that new users and developers have gravitated toward these newer networks: Improved throughput and functionality are things that users and developers can immediately appreciate, whereas the benefits of “decentralization” as a feature are seemingly amorphous.
In the short term, users may be tempted by the performance offered by more scalable blockchains, not realizing the importance of decentralization until a wake-up call brings everything crashing down.
He continues:
The reality, however, is that without decentralization these crypto networks lose their most important qualities of being “permissionless” and “censorship-resistant” — that is, that anyone can use the network and anyone can build on top of them.
After all, the entire point of a decentralized blockchain is to provide a hard-promise — an immutable ledger with open, non-discriminatory participation. In a sense, we bear the inefficiency of decentralization because it is the only way to enable a network with these qualities.
Other blockchains that achieve 1,000 TPS or higher do so by having a small, fixed number of nodes that validate all the transactions — 21 in the case of EOS, 101 in the case of Lisk.
But a network run by 21 nodes requires that you trust these 21 publicly-identifiable nodes to not make changes to the protocol, or restrict certain people from using it for certain purposes in the future.
It’s a lot harder for a malicious entity to influence thousands of anonymous Ethereum nodes to censor certain transactions than it would be for them to gain influence over 15 out of 21 publicly-known block producers. Or for these 15 block producers to form a cartel and change the rules in ways that profit them, like what happens on centralized platforms. Or for a government or corporation to put pressure on these entities to censor certain transactions or users.
As Spencer puts it:
These semi-decentralized platforms are subject to the same social and economic pressuresthat motivate centralized platforms to censor certain users and activities, and therefore will trend toward the same outcome they’re supposed to correct.
developers can’t trust 100% that the base layer will always remain permissionless and censorship-resistant, then there’s very little incentive for them to start building on the platform vs. simply using a traditional web server.
By sacrificing base-layer decentralization in the nearsighted goal of attracting users by increasing throughput, these platforms have ironically undermined the entire motivation for using a blockchain in the first place.
And on top of that, increasing throughput on Layer 1 isn’t even a scalable approach.
It will get you some initial gains, but it’s fundamentally bottlenecked by the nature of blockchains and not the way to achieve true scalability.
Which brings me to my next point…
4. It will be impossible to run all the world’s decentralized applications on a single blockchain: Scaling has to occur on Layer 2
It would be absurd to try to run the Internet’s 100 most popular games and social apps all on one giant supercomputer.
Likewise, it’s absurd to assume all the world’s decentralized applications in the foreseeable future will run on one blockchain.
Let’s look at some quick numbers.
Facebook experiences 30,000+ Likes / Comments per second, the Nasdaq sees 20,000+ trades per second, and MMO games like PlayerUnknown’s Battlegrounds handle over 1M concurrent users updating game state.
It would only take a few dozen apps and games of this size before you’re exceeding one million TPS in aggregate.
And then what do you do when the number of users doubles?
Clearly trying to run every single DApp on the same chain is not a practical approach.
It doesn’t matter if a blockchain can do a thousand transactions per second or a million transactions per second — no single blockchain will be fast enough to handle all the world’s decentralized applications on the same chain.
Scaling has to occur on Layer 2
The solution is obvious — these applications will need to be split up across multiple blockchains.
We realized this early on at Loom, when we presented the idea of application-specific sidechains. We foresaw that eventually some decentralized applications would grow popular enough to reach even 1/10th of Facebook’s scale, and the only possible way to run them would be on their own dedicated chains.
Of course, if you put these DApps that require thousands of transactions per second on their own standalone blockchains, they would be vulnerable to the same issues we discussed above in “Why Decentralization Matters.”
But if you put them on a sidechainto a blockchain that is sufficiently decentralized (like Ethereum) — you get the best of both worlds.
Sidechains provide higher scalability without sacrificing security
A sidechain can use a different consensus algorithm (like DPoS) optimized for DApps that require very high TPS or low-latency, while storing any tokens or data requiring a high level of security on the main chain.
This way, even though the sidechain is less decentralized than the main chain, the amount of trust required by users is minimized, since they have the option to move anything of real value to the main chain for safe keeping. (Even more so if you secure the Layer 2 assets with Plasma Cash).
By putting your DApp on a sidechain to a decentralized mainnet, you get all the benefits of higher scalability offered by a faster blockchain, while maintaining the same trust and security guarantees provided by the decentralized base layer.
Spencer, in his article, arrives at the same conclusion we came to:
The path forward: Highly decentralized base layer with increased centralization (and efficiency) on higher layers
And in fact, this seems to be the same model envisioned by Vitalik Buterin himself:
You could run StarCraft on the blockchain. Those kinds of things are possible. High level of security and scalability allows all these various other things to be built on top. Ethereum is a secure base layer that doesn’t have too many features.
~Vitalik Buterin
Ethereum provides a secure base layer for Layer 2 solutions to build on top of
Now we understand that:
- Scaling needs to happen on Layer 2
- The most important property of a Layer 1 is decentralization
So the real question is, if not Ethereum, what base layer would you build your Layer 2 on top of?
We’ve already seen that very few chains do decentralization as well as Ethereum does.
According to a recent ConsenSys report, there are “just under 17,000 nodes running the Ethereum blockchain across six continents, making it the most decentralized blockchain platform in existence.”
And any additional features another blockchain might offer, such as higher throughput, gas-less transactions, low-latency transactions, etc., can simply be implemented as Layer 2 services on Ethereum.
In fact, those features are exactly what we’re building at Loom Network with ZombieChain — a Layer 2, gas-free, low-latency DPoS sidechain to Ethereum.
And that’s just one of many Layer 2 scaling solutions under development.
It’s hard to fathom why a developer would want to replace Ethereum instead of simply building on top of it ?
Ignoring the obvious motivation that doing so allows them raise hundreds of millions of dollars in an ICO… ?
It’s kind of like reinventing the wheel.
Sure, you might be able to build a slightly better base layer that still provides sufficient decentralization AND throws in some additional features.
But then you’d somehow need to convince all the developers to jump ship to use an untested platform — and in the meantime, Ethereum developers could take whatever good ideas you had, and implement them on a Layer 2 chain on top of Ethereum.
It’s also extremely risky.
If a Layer 2 platform gets hacked or exploited, users’ loss is minimized because the majority of tokens and data of value is still safely stored on the Layer 1 (Ethereum).
But if you build a new Layer 1 blockchain that stores tokens that users pay real money for, the odds that your code will be exploit-free out the gate are slim — and in the case of an exploit, your users potentially have billions of dollars at stake.
I’m going to go ahead and predict that after we see the first major blockchain exploit in which millions or billions of dollars in token value evaporate into thin air, we’ll start to hear a similar refrain among blockchain engineers:
At the time of this writing, there’s $61 billion in circulating ETH.
That’s a LOT of financial incentive for someone to try to hack / exploit the network.
