Solana
Rethinking the Throughput Ceiling
When the Solana whitepaper was published in 2017 by Anatoly Yakovenko, its central claim was audacious: a blockchain could achieve throughput in the hundreds of thousands of transactions per second without sacrificing decentralization, by treating time itself as a cryptographic primitive. Most blockchain researchers at the time accepted that high throughput required either trusted validators or reduced node count — Solana's thesis was that both assumptions were wrong, and the real bottleneck was coordination overhead in establishing a shared sense of time.
The mechanism Yakovenko proposed was Proof of History (PoH): a verifiable delay function that produces a cryptographic sequence proving that specific events occurred at specific points in time, without requiring nodes to communicate to agree on timestamps. Combined with Tower BFT (a PoS consensus built on top of PoH's clock), Gulf Stream (a mempool-less transaction forwarding protocol), and Sealevel (a parallel smart contract runtime), Solana assembled a system where the throughput ceiling was set by hardware rather than protocol design.
This framing — replace coordination overhead with hardware requirements — has shaped both Solana's successes and its most persistent criticisms.
Transaction Records and Real-World Performance
Solana's testnet demonstrations and early mainnet figures were striking by any comparison available in 2020. Where Ethereum processed 12-15 transactions per second and Bitcoin around 7, Solana was demonstrating 50,000+ TPS in controlled conditions. The theoretical maximum, often cited as 710,000 TPS, reflects the bandwidth limit of the PoH leader's hardware under ideal conditions.
Real-world mainnet performance has been more modest but still dramatically higher than competing L1s. During periods of normal load in 2021-2022, Solana processed 2,000-4,000 non-vote transactions per second — the vote transactions that validators submit for consensus accounting for a large fraction of total on-chain activity. During the NFT mint crazes of 2021, Solana demonstrated sustained throughput that Ethereum mainnet could not approach, enabling mint prices below $1 per transaction during periods when Ethereum gas costs made comparable mints cost $100-500 each.
The performance enabled use cases that were practically impossible on Ethereum mainnet. Serum, the decentralized order book exchange built by Sam Bankman-Fried's team, required the ability to place and cancel limit orders at low latency and cost — a transaction profile that suits Solana's architecture far better than Ethereum's. The Solana whitepaper's claim that its design enables "real-time" financial applications found its first major test in Serum's central limit order book.
defi-growth-story">The NFT and DeFi Growth Story
The 2021 crypto bull market provided an unplanned stress test of Solana's architecture. NFT collections on Solana — Magic Eden emerged as the dominant marketplace — achieved mint volumes that Ethereum's gas market could not accommodate affordably. Projects like Degenerate Ape Academy and Okay Bears drew hundreds of thousands of mint attempts within seconds, producing a transaction load that exposed both Solana's strengths and its fragility.
DeFi on Solana grew from near zero to over $10 billion in TVL during 2021, driven by Marinade Finance (liquid staking), Orca and Raydium (AMM DEXes), Mango Markets (margin trading), and Solend (lending). The ecosystem's ability to offer sub-cent transaction costs attracted retail users who were priced out of Ethereum DeFi.
The velocity of ecosystem growth also attracted development talent. The Anchor framework, which provides Rust-based smart contract development with an interface similar to Ethereum's Hardhat/Foundry ecosystem, significantly lowered the barrier to building on Solana. By mid-2022, Solana had the second-largest active developer ecosystem among smart contract platforms.
The Outage Controversies
Solana's performance record is inseparable from its outage history. The network has experienced more than a dozen significant outages since mainnet launch, with several lasting multiple hours and the longest stretching past 17 hours in September 2021.
The September 2021 outage was caused by an NFT mint that generated 400,000 transactions per second of attempted traffic — not on-chain transactions but transaction requests — overwhelming the network's ability to reach consensus. Validators ran out of memory and the network stalled. Recovery required a coordinated restart by validator operators following a patch.
Subsequent outages followed similar patterns: a high-demand event generates a spam flood of transactions, the network's mempool-less design (Gulf Stream) does not apply sufficient backpressure, validators become overwhelmed, and consensus stalls. Solana's core developers implemented several mitigations, including QUIC (a UDP-based transport protocol replacing the original UDP implementation to enable better congestion control) and fee market improvements in the SIMD-0096 proposal, which introduced local fee markets to allow high-demand programs to set their own priority fee tiers.
The reliability criticism cuts to a fundamental tension in Solana's design. By minimizing coordination overhead and optimizing for throughput, the protocol reduced its tolerance for traffic spikes beyond design parameters. Ethereum's gas auction mechanism, often criticized for making fees expensive, serves an implicit rate-limiting function: when demand exceeds capacity, fees rise until enough users defer their transactions. Solana's original design lacked this pressure valve.
The Hardware Requirements Debate
Solana's throughput claims rest on hardware requirements that are substantially higher than other blockchain validators. A competitive Solana validator in 2024 requires a server with 256-512 GB of RAM, high-performance NVMe storage, a powerful CPU, and ideally a GPU for faster transaction processing. The monthly cost of operating such a validator is estimated at $1,500-3,000, compared to $500-1,000 for a comparable Ethereum validator.
Critics, including Ethereum researchers, argue that this hardware requirement centralizes the validator set. If only well-capitalized operators can afford to run validators, the network's censorship resistance depends on the interests of a relatively small group of institutional participants. Ethereum's design philosophy, as articulated in the Ethereum roadmap documentation, explicitly targets a validator requirement that a consumer laptop could eventually meet.
Solana's response has been partly philosophical and partly technical. The philosophical argument is that hardware costs follow a predictable deflationary trajectory — what requires a $3,000 monthly server today will eventually run on equipment costing a fraction of that. The technical argument is that Solana's validator count (around 1,900-2,000 active validators as of 2024) is already comparable to Ethereum's effective decentralization when weighted by the concentration of stake among large stakers.
The Nakamoto Coefficient
One metric used to evaluate validator decentralization is the Nakamoto Coefficient: the minimum number of entities that would need to collude to halt or corrupt the network. Solana's Nakamoto Coefficient has varied between 19 and 33 depending on measurement methodology, which is competitive with Ethereum's coefficient and higher than Bitcoin's mining pool concentration.
This metric does not capture all dimensions of centralization — geographic concentration, shared infrastructure, and hardware vendor dependencies matter too — but it offers a useful corrective to the simple claim that "fewer validators means less decentralized."
Impact on Blockchain Design Philosophy
Solana's influence on blockchain design extends beyond its own ecosystem. The demonstration that a blockchain could achieve meaningful throughput gains by parallelizing transaction execution at the virtual machine level informed the design of Aptos and Sui, both founded by former Meta engineers who worked on the Diem (Libra) blockchain. These networks adopted the Move programming language and parallel execution frameworks, applying Solana's lesson that EVM's sequential execution model is a performance bottleneck.
Ethereum's own roadmap was influenced by Solana's performance evidence. The shift toward a rollup-centric scaling strategy — treating Ethereum L1 as a settlement and data availability layer rather than an execution environment — was in part a response to the demonstrated demand for high-throughput execution environments. If users wanted fast, cheap transactions, Layer 2 rollups could provide them while inheriting Ethereum's security.
Solana proved that the throughput ceiling was not a fundamental physical limit but an artifact of specific design choices. That proof, even amid the outage controversies, permanently changed the terms of the debate about what a public blockchain could achieve.