Looking Past the Blockchain Trilemma
Imagining a world beyond the blockchain trilemma of scalability, decentralization and security
Considering the current efforts put in place to solve issues arising from the ‘blockchain trilemma’, it’s interesting to imagine what impact sorting these challenges out might have on the blockchain ecosystem. The three issues to balance here are scalability, security and decentralization.
While I believe solving the blockchain trilemma boosts the technology adoption and is beneficial for the space, it does not guarantee the fundamental paradigm shift once promised by industry players, namely the rise of scalable and efficient decentralized sub-economies that benefit a greater number rather than the few.
This essay explores potential scenarios that could occur if the blockchain trilemma was solved.
The Blockchain Trilemma
Public blockchains have limitations; this is nothing new in the blockchain community. In fact, these limitations have been described as the “scalability trilemma” (for further information regarding the trilemma, please refer to Ethereum’s GitHub documentation). It consists of saying that out of the three properties decentralization, scalability and security, today’s blockchain systems can only optimise two at a time. Meaning that strengthening any two will automatically weaken the third.
Let’s use c as the size of computational resources where n refers to the size of the ecosystem such as market cap, state size (all information available on-chain) or the number of nodes, in the broad sense. O is an optimisation function for c and n:
- Decentralization is defined as a system being able to run even when participants only have access to O(c) resources
- Scalability as being able to process O(n) > O(c) transactions (the network can only process the amount of transactions an individual node can)
- Security as the ability to be secure against attackers with up to O(n) resources.
Note that in the case of Ethereum, transactions can be the execution of any type of smart-contract, which is essentially a piece of code or small program. For software engineers, transaction throughput describes the amount of transactions per second a network can handle. To clarify further, the centralization of validating nodes can lead to a greater transaction speed and higher scalability, since less nodes are required to verify transactions. Ultimately though, this defies the decentralized aspect of blockchain as well as its security, since the amount of validating nodes is reduced.
As a metaphor, think of this problem as the number of safety checks that went into your brand new car before you bought it. The less safety checks, the less sure you can be about the security of your car, but the quicker the car gets delivered to you. On the other hand, a higher decentralization level currently implies longer block validation periods, increasing security but diminishing scalability. In this case, the odds that your car is safe are higher but it will take much longer to be shipped to you.
Solving the trilemma challenge is about ensuring that these properties can be achieved at the protocol’s first layer. The trilemma isn’t impossible to solve, but given the current technological challenges it seems pretty hard to achieve it in the short run.
Why Finding a Solution Matters
We’ve witnessed the risks associated with too-high centralization, greater scalability and diminished security. Just recently, the Ethereum Classic chain suffered from a 51% attack — stipulated as an attack performed by a Chinese manufacturer testing a new 1,400 MH/s ethash application specific integrated circuit (ASIC). ASICs are specific type of hardware optimized for one or a few type of computations. We call “51% attacks”, attacks which can be performed when over 50% of the hashrate is owned by one party. The hashrate is the speed at which a computer computes an operation on the blockchain, in most cases solving a Proof of Work operation. This theoretically allows the attacker to input false information on the blockchain.
In December of last year, Vertcoin, a cryptocurrency intended to be ASIC resistant for greater decentralization, was also victim of a 51% attack. As David Vorick a lead developer of Sia concludes in this excellent piece (discussing the power of ASIC manufacturers and their impact on decentralization): “Manufacturing just inherently leads to centralization, and it happens across many different vectors” — thus even the most popular blockchains out there have a hard time reaching true decentralization. Current data suggests that combined together, the five largest mining pools have over 50% of the Bitcoin hashrate.
For the past few years, finding layer one and layer two scalability solutions has been a top priority for the developer communities of projects like Ethereum and Bitcoin. Scalability solutions like lightning and SegWit for Bitcoin, or plasma and casper for Ethereum, have burgeoned because it is commonly believed in the space that infrastructure, rather than applications, will drive adoption. In October of last year, Union Square Ventures published an article aimed at debugging this myth, arguing that applications should drive infrastructure, and not the other way around.
Eight Potential Scenarios
In this post, I want to imagine a world where the issues associated with the trilemma are solved and explore how this would affect the ecosystem. I imagine a situation where blockchain networks are not constrained by the computational power of single validating nodes, where true hashrate decentralization is taking place, and where security — even against Quantum attacks — is guaranteed. In the sections below, I describe possible scenarios where each section discusses an independent concept:
1. GPU and CPU prices will explode
Assuming Proof of Work (PoW) remains the most popular consensus mechanism amongst public blockchains — largely due to its high level of security and envisioning a mainstream cryptocurrency resistance to ASIC — I believe a drastic inflation in CPU and GPU prices would take place, due to their ability to generate revenue for mass consumers. “ASIC resistance”, though currently fictional, is the idea of developing cryptocurrencies that resist to ASIC hardware, enabling the real decentralization of the network’s hashing power. In the scenario where ASICs become inefficient, less specific hardware can then be employed to mine cryptocurrencies and secure PoW based networks.
As we’ve witnessed during the 2013 Bitcoin and 2017/early 2018 crypto bull-run, the cryptocurrency market drives up demand for general purpose processing units. Cheap and efficient ASICs on the other hand, have pressured GPU prices downward given their superiority in mining PoW SHA256-based cryptocurrencies. Thus, if true decentralization is achieved and anyone equipped with CPU/GPU can secure the network while gaining access to a passive stream of income, CPU and GPU manufacturers could become much more valuable than they are today.
2. Industry-wide efforts to leverage blockchain potential will take place
I also believe coopetition around public blockchains will increase between enterprises, especially in the supply chain, financial and insurance sector. Coopetition can be defined as businesses that engage in cooperation and competition simultaneously. Indeed, while there’s been a plethora of consortiums (EEA, R3, etc.) created to federate enterprise blockchain efforts, little to no tangible and scalable outcomes have emerged from them as of now. Blockchain-based interbank transfers might have been tested, yet they haven’t been adopted.
With the trilemma being solved, we can assume that enterprises will become less averse to, and more certain of, the potential benefits created by development efforts. Solving the scalability problem will allow companies to access high throughput using Ethereum’s Virtual Machine without the fear of network congestion (a network state where transactions are slower and the overall system bugs).
3. Open-source and public blockchains will beat private ones in performance and adoption
If the trilemma is solved, the permissioned and private networks (which often make use of consensus algorithms such as Proof of Authority and Practical Byzantine Fault Tolerance) will lose their added-value of being less resource intensive and more scalable, compared to permissionless systems. For enterprise-wide operating systems, open-source technologies like Linux have demonstrated how useful and dominating free and open software can be. Supercomputers and web servers are only two additional working examples of this success story.
4. UI will need to become the center of development.
Scalability and performance improvements will make the overall user experience better, but the user interface will remain a pain point for blockchain adoption that needs to be solved. Blockchain wallets will still encompass key management, and the irreversibility of transfers is something that customers might still be anxious about.
Third parties and centralized service providers have the advantage of being able to reverse transaction status and recover any lost piece of information, which makes them a safe haven for customers who might not want to be held responsible for all actions associated with their wallet. It remains to be seen whether this shift in responsibility can still meet market needs.
5. Greater value creation for the end user
Greater value will be created for the end user as development efforts move away from infrastructure to applications. We all know that consumers don’t care about how stuff works as long as it works. If developers start to focus on products that create value for end-users regardless of the how, then they’ll be better at answering the what. This goes back to the “infrastructure vs. application” thesis.
6. A handful of universal cash systems will arise
The large majority of cryptocurrencies developed and employed as a means of payment are likely to vanish. Bitcoin, although originally envisioned by Satoshi as a “P2P electronic cash system”, failed in doing so for some time now. Consequently, giving birth to a large number of forks and cryptocurrencies doing the same thing: competing on usability and scalability.
The trilemma being solved, Bitcoin could achieve its initially intended prophecy of becoming pure digital cash. The remaining currencies will be behind, primarily because of Bitcoin’s dominating network effects. Cryptocurrencies focusing on privacy and other specific features however, would still be in the race because of their differentiation in offerings.
7. The payment card industry and payment services providers could face unprecedented disruption.
In today’s payment market, the aforementioned players compete mainly on customer services (chargebacks, fraud prevention, etc.). While they can already manage high transaction throughput with ease, if the scalability issues of blockchains were solved, then the payment processor infrastructure capabilities would no longer be a USP. Merchants could opt for less pricey payment solutions, with, however, less protection for sellers and buyers.
8. The future of utility tokens remains unknown.
While solving the trilemma will lead to greater adoption of blockchain technologies, there is still no guarantee that utility tokens would take off. User interface constraints, along with the inherent risks associated with purchasing digital assets are barriers to adoption. Most internet-based services are currently free of payment. There is no guarantee that there would increased market traction for a token-based economy. Trilemma-free blockchains would rather give birth to the race for “killer ecosystems”.
Beyond Fueling Adoption
Solving the trilemma will help fuel blockchain adoption in many ways. I believe enterprises will be more at ease collaborating, public blockchains will triumph over private ones, lay users will have greater incentives to participate in decentralized systems, and that the payment industry will surely be fundamentally disrupted.
The question remains whether this will be enough for utility tokens and decentralized economies to rise, since important efforts in UI and UX will remain necessary, and their adoption requires a change in how users currently interact with internet-based services.
What do you think a post-trilemma future would look like? Feel free to leave a comment below with your thoughts and follow the NBT Thing Tank for future-facing articles. If you want to chat with Clément Bihorel on blockchain, he’ll be at the governance-focused Aracon conference in Berlin on Jan 29–30, 2019.