Is it time to embrace hardware in blockchain?
Is our ideological obsession with decentralization killing Web3’s future? While we endlessly debate the purity of decentralization, the real world waits for no one. It’s time we confront an inconvenient truth: Our ideological rigidity might be the very thing holding back Web3 technologies from mainstream adoption. In this post, I’ll argue that it’s not about ‘how decentralized we can get,’ but rather ‘how decentralized do we actually need to be?’, particularly when it comes to hardware and why a more pragmatic approach is key. Buckle up.
The illusion of perfect decentralization
Consider Ethereum, a beacon of decentralization:
Multiple clients (node software) exist ✅
No one group of developers can overly influence the network as users are free to move between clients. In fact, recently a bug surfaced in one of the clients and this diversity saved the network from any sort of outage.
Network integrity maintained through decentralized PoS ✅
Proof of Stake maintains its predecessor's (Proof of Work) decentralization. As long as you’re able to come up with the staking requirement (and it can be collectively), you can participate in the consensus algorithm of the network.
No single entity controls the network ✅
Ultimately, no single entity controls the network. Certainly, people look towards Vitalik for inspiration, but there is no CEO, and nobody can shut down the network.
We can safely conclude the software and economic incentives are ideally decentralized, right?
But, the hardware isn’t. Three cloud providers account for over two-thirds of all of Ethereum’s nodes. And do you know what? That’s absolutely fine. This level of centralization at the hardware level is surprisingly practical. Running infrastructure is costly, and cloud solutions provide an economical alternative that arguably makes node operation more decentralized.
Economies of Scale: A Necessary Evil?
Economies of scale are cost advantages experienced when production becomes more efficient as it increases in size, mainly due to the more efficient allocation of capital over a larger number of goods or services. This impacts crypto too, because, if we were to start decentralizing hardware, we’d immediately lose the cost benefits of economies of scale.
How would we reconcile this increase in cost with one of our goals of providing financial inclusion for those who are unbanked, mostly in developing nations? Pushing for extreme decentralization obliterates these cost benefits, making crypto services prohibitively expensive.
This may sound extreme, but it’s the path we’re on if we start forcing decentralization down to its smallest denominator. We need a balanced approach.
For example, a piece of hardware that often comes under fire for its supposed centralization is Trusted Execution Environments (TEEs).
Trusted Execution Environments (TEEs) are secure, isolated regions within a CPU that provide a fortified sanctuary for executing code and processing data. They guarantee that the code running is precisely what it claims to be, free from tampering or unauthorized modifications.
While there’s some truth to them being centralized — given that a handful of large manufacturers like Intel, AMD, Nvidia, Amazon, Apple, IBM and ARM produce most TEEs — it’s worth scrutinizing this claim more closely.
Ubiquity and Security
The device you’re reading this blog piece on is probably running on a CPU from an even smaller group. My bet is, if you’re on a laptop, you’re running an Intel or AMD chip and if you’re reading this on your phone, it’s probably an Apple or ARM chip. TEEs are embedded in the core architecture of virtually all modern computing devices — from smartphones to laptops to cars and data centres. They are a cornerstone of digital security beyond just the blockchain world. In fact, pretty much every institution you interface with where keeping data secure is a priority is using TEEs.
Changing Sentiments
It’s time we reconsider the black-and-white perspective on decentralization and appreciate the security layers that TEEs add to blockchain networks and how it’s the only solution today that can provide encryption and open up the Web3 design space.
I’ve seen firsthand how ideological rigidity can kill innovation. A Web3 game I’ve been following for several months is dying at its hands. It was a game that required hiding some information, like the cards you hold. Rather than focus on getting an MVP out, the game developers were discussing ZKP polynomials, matrix algebra, whether a 30-second wait time for each move was acceptable and 3 GB proofs that needed to be passed around.
Three things sprung to mind:
- There was hardly any focus on the game itself being fun.
- The game could have been built within days using an encrypted network.
- Enthusiasm is going to dither as roadblock after roadblock gets hit and the project will die. Unfortunately, that is now happening.
The saddest part is, it was a game I was really looking forward to, a port of one of my favourite card games. So now, rather than have a neat game out there, that could help us on our journey to mainstream adoption, we’ve now got a pile of unfinished code. And for ideological over pragmatic decisions.
Why not get the game out there on proven technology, start gathering feedback, and make the game fun? And revisit decentralizing the hardware down the line?
Luckily, the winds are changing. Projects like WorldCoin and Flashbots are embracing TEEs, and even Vitalik Buterin has softened his stance on the technology, as evidenced by his recent post.
Conclusion
Decentralization is a spectrum, not a binary state. Striving for an ideal state of decentralization without any compromise can hinder the mainstream adoption of blockchain technologies.
We debate constantly what the next big use-case in Web3 might be. If we adopted hardware for the short to medium tem, it’d be obvious. There’s a whole universe of Web2 applications that we could be porting over but without the type of encryption (particularly, through TEEs), they’re a non-starter.
With this wave of new dApps in motion and further adoption, we could then turn our attention to further decentralizing away the reliance on hardware.
TEEs have had tens of years of battle hardening on billions of devices. This differs from most ZKP solutions that assume their underlying tech is exempt from any vulnerabilities. This is obviously not true with the Nova attack, and Vitalik is now suggesting they also take measures.
Pragmatic solutions, like Ten, a project I contribute to, are built from the ground up to withstand attacks on the TEEs because we know what they look like. Ten already has a fault-proof mechanism designed such that a single non-malicious user renders a successful attack on the TEE useless.
Ten’s balanced, secure, and pragmatic use of TEE technology illustrates a pathway towards a more decentralized yet practical blockchain universe.
Find out more
For further reading on TEEs, read Tudor Malene’s post here.
To learn more about Ten, dive into the whitepaper, chat with the community on Discord, and follow us on Twitter.
