Decentralization in PoS Networks: ๐๐ต๐ฒ ๐๐ฎ๐ฟ๐ธ ๐ฆ๐ถ๐ฑ๐ฒ ๐ผ๐ณ ๐๐ต๐ฒ ๐ ๐ผ๐ผ๐ป
Overview:
โข Analyzing validator decentralization on a geographical & infra scale in PoS networks via Messari's report.
โข Positioning ๐ฃ๐ผ๐น๐ธ๐ฎ๐ฑ๐ผ๐ amongst other networks.
โข Spotting the vulnerabilities.
โข Solutions for creating unstoppable PoS networks.
๐๐น๐ถ๐ฒ๐ป๐ ๐๐ถ๐๐ฒ๐ฟ๐๐ถ๐๐: Most networks, except Ethereum, rely heavily on single-client software. ๐๐ช๐ด๐ฌ๐ด: A bug in this client can compromise the entire network.
E.g. Solana has experienced this issue, which led to a client fork. And now there exist 2 clients with a 3rd in development.
Ethereum also suffered from an error in Geth clients in 2016, causing block confirmation to a halt.
Spotlight on Polkadot: is dependent on Parity's client atm.
But @soramitsu_co's Kagome (C++) and @ChainSafeth's Gossamer (Go) are in development and available in testnet.
๐๐ฐ๐๐ถ๐๐ฒ ๐ฉ๐ฎ๐น๐ถ๐ฑ๐ฎ๐๐ผ๐ฟ ๐ก๐ผ๐ฑ๐ฒ ๐๐ผ๐๐๐ถ๐ป๐ด ๐๐ถ๐๐๐ฟ๐ถ๐ฏ๐๐๐ถ๐ผ๐ป: Many blockchains also heavily depend on a few network providers, especially AWS, OVH and Hetzner.
This reliance, a result of self-hosting difficulties and cost-effectiveness, creates a single point of failure.
Polkadot and Cardano are less reliant on a single network provider, while over 60% of Aptos and Avalanche's validator nodes use 2-3 network providers.
๐ง๐ต๐ฒ ๐ฐ๐ผ๐ป๐ฐ๐ฒ๐ป๐๐ฟ๐ฎ๐๐ถ๐ผ๐ป ๐ผ๐ณ ๐๐๐ฎ๐ธ๐ฒ ๐ผ๐ป ๐๐ช๐ฆ(Amazon Web Services):
- Avalanche: 67%
- Aptos: 43%
- Near: 35%
- Cardano: 30%
- Solana: 17%
- ๐ฃ๐ผ๐น๐ธ๐ฎ๐ฑ๐ผ๐: ~7% ; meanwhile OVH as being the highest concentrated network provider by 17%.
Solana and Polkadot are least dependent on a single network provider. But still, when combined with their top three network providers, they surpass the 33% risk threshold.
๐๐ช๐ด๐ฌ๐ด: An outage, regulation, or crypto-related ban affecting these providers could seriously impact networks.
๐๐ฒ๐ผ๐ด๐ฟ๐ฎ๐ฝ๐ต๐ถ๐ฐ ๐๐ถ๐๐๐ฟ๐ถ๐ฏ๐๐๐ถ๐ผ๐ป: It's crystal clear that there is a dominance of EU and US here for all networks, due to their cost-efficiency and accessibility advantages.
Even though Avalanche has greater presence across low-tier regions, its 31% of active stake is on US.
Together with Germany 46% of stake of Avalanche is concentrated only 2 country.
This is also caused by AWS.
- Cardano: US + Germany = 51%
- Solana: US + Germany = 39%
- Polkadot: US + Germany = 35%
- Near: US + Germany = 33%
- Aptos is not concentrated on US just like Polkadot but if we add Korea to US + Germany -> 50%
Except Aptos all of them are exceeding critical 33% stake threshold only with two countries.
Spotlight on Polkadot: EU is accounted for 68% of public rewards.
There is no concentration on US due to not being reliant on AWS like others. But there is a heavy dependency on Europe.
And still >60% active validators are based on EU.
๐๐ช๐ด๐ฌ๐ด: Client diversity might have a bigger impact if any vulnerability occurs, but there has to be considered the future risks: what if a country where a significant % of nodes located decides to enact regulations or bans any crypto-related activities?
Silver lining: China's banning BTC and Hetzner canceling crypto operations didn't cause any major damage.
๐ฆ๐ต๐ผ๐ฟ๐-๐บ๐ถ๐ฑ ๐๐ฒ๐ฟ๐บ ๐๐ผ๐น๐๐๐ถ๐ผ๐ป๐: Encouraging self-hosting and local validators.
Creating incentivizes to establish infra on other regions beyond US, EU and Asia.
E.g. in Polkadot, a treasury funded project IBP - Infra Builder Program aims to support building infra in the locations typically overlooked due to cost concerns.
Solana Foundation Delegation Program incentivizing the use of data centers hosting fewer than 10% of validators.
๐ช๐ต๐ฎ๐ ๐ฎ๐ฏ๐ผ๐๐ ๐๐๐ต๐ฒ๐ฟ๐ฒ๐๐บ? There exists challenges there as well.
E.g. -> staking pool diversity, with nearly a third (33%) of pool shares are on Lido.
Furthermore, approx. 50% of the Ethereum nodes are operating in the US, and this increases to 67% when combined with nodes in Germany.
And over half of US based nodes are in a single data center, 20 min away from the CIA HQ. Therefore, initiatives like @ether_fi's Operation Solo Staking are significant in mid-long term.Also solutions like DVT (Distributed Validator Tech) are emerging to reduce the chances of a single point of failure.
As a result, @divastaking is preparing to launch its mainnet for DVT this week.
Although Ethereum's validator count is impressively high, ~ 630K, the core devs are now contemplating increasing the upper bound of validator stake to 2048 ETH to sustain network throughput over time.
Hopefully this will calm down the validator number maximalism. After all, the control over the PoS Networks depends on the stake.
๐๐๐ก ๐๐๐๐ฆ ๐กโ๐ ๐๐๐๐๐ก๐ ๐๐๐๐ก๐๐๐๐๐ ๐๐๐๐ฃ๐, ๐๐ข๐ก ๐๐๐ ๐ ๐๐กโ๐๐ ๐๐๐๐ ๐๐๐๐๐๐ก๐๐๐๐ ๐๐๐ ๐๐๐ก๐ค๐๐๐๐ ๐ โ๐๐ข๐๐ ๐๐๐๐๐ข๐๐:
โข Staking Accessibility: How easy is it for someone to become a validator?
โข Network Governance: Who has the "control" of decision-making? Do validators have a say in network gov like in Cosmos ecosystem? What kind of impacts and politics does this introduce?
โข Trade-offs: in terms of providing decentralization?
๐ข๐๐ฒ๐ฟ๐ฎ๐น๐น, ๐๐ต๐ฒ ๐ณ๐๐๐๐ฟ๐ฒ ๐ผ๐ณ ๐ฃ๐ผ๐ฆ ๐ป๐ฒ๐๐๐ผ๐ฟ๐ธ๐ ๐น๐ถ๐ฒ๐ ๐ถ๐ป ๐ผ๐๐ฒ๐ฟ๐ฐ๐ผ๐บ๐ถ๐ป๐ด ๐ฎ๐น๐น ๐๐ต๐ฒ๐๐ฒ ๐ฐ๐ต๐ฎ๐น๐น๐ฒ๐ป๐ด๐ฒ๐ ๐ฎ๐ป๐ฑ ๐ฒ๐บ๐ฏ๐ฟ๐ฎ๐ฐ๐ถ๐ป๐ด ๐๐ต๐ฒ ๐ฑ๐ฒ๐ฐ๐ฒ๐ป๐๐ฟ๐ฎ๐น๐ถ๐๐ฎ๐๐ถ๐ผ๐ป ๐ฎ๐ ๐ฎ ๐๐ผ๐ผ๐น ๐๐ผ ๐ฏ๐ฒ๐ฐ๐ผ๐บ๐ฒ ๐๐ป๐๐๐ผ๐ฝ๐ฝ๐ฎ๐ฏ๐น๐ฒ ๐ป๐ฒ๐๐๐ผ๐ฟ๐ธ๐.
Coming from a background of Computer Science and Cinematography, I've found my niche in this space, blending technical know-how with storytelling.
Much as films tell stories, blockchains craft their own narratives, complete with innovation, culture, believers and emotions.
In this space, I connect the dots, bridging the gap between technology and story, making it relatable for everyone.
As a filmmaker in real life with a background in computer science, here I am, primarily conducting research in this wild wild space.
Often agnostic in my interests and inquiries, I regularly delve into cross-research, exploring both past and future narratives, trends, and developments in the broader blockchain space.
And in this particular space, I try to connect the dots, bridging the gap between technology and story, making it relatable and digestable for all.
0 comments