Appendix A
Technical Limitations of a Native L1 Exchange On Solana
1. High Latency
Round trip latency is a crucial factor in the operation of an exchange, referring to the time it takes for an order to be submitted to the exchange plus the time taken for the result to be communicated to the user. Low latency is essential because it enables market makers to adjust their quotes in response to market movements in real time. In optimal conditions, the best case round-trip latency for order placement on the Solana L1 takes a few hundred milliseconds at the least secure 'processed' commitment level, and between 1-3 seconds at the 'confirmed' level. In contrast, fully off-chain exchanges like Binance offer latencies as low as 5 milliseconds, allowing for a faster and more reliable trading experience.
2. Congestion
Whilst we are excited to see Solana usage grow so substantially, we are acutely aware that such growth-induced congestion can diminish the quality of trading on Zeta's Layer 1 implementation. This includes elevated gas fees, longer confirmation times and reduced transaction success rates.
In practical terms, these challenges mean that traders have struggled to reliably cancel orders or execute take-profit and stop-loss orders. Many traders resort to spamming nodes with repeated requests or paying exorbitant tips to validators to prioritize their transactions. Unfortunately, these practices further degrade the trading experience for all users of the network.
3. High barrier of entry for liquidity providers
The liquidity on a fully on-chain order book on Layer 1 cannot rival that of centralized exchanges as market makers tasked with providing liquidity encounter several obstacles that hinder efficient quoting:
Non-deterministic order placement and cancellations result in transactions that may take 20-30 seconds to confirm.
Long confirmation times mean market makers are hesitant to offer tight spreads, which reduces the quality of liquidity available to users. To encourage tighter quoting by market makers, the exchange must compensate with more substantial rebates.
Non-deterministic order placement also necessitates additional complexities in exchange infrastructure, such as adjustments to blockhash utilization and custom order book solutions like time-in-force and sequence enforcers, aimed at mitigating issues arising from confirmation times. This increased complexity in exchange architecture not only heightens the risk of bugs but also prolongs integration processes for market makers.
Expensive gas fees for market makers
While Solana offers lower gas fees compared to many other Layer 1 solutions, the costs can still accumulate quickly for market makers quoting on an on-chain exchange. Market makers on Zeta have reported spending 2-3 $SOL per day trading across 16 assets, which at time of writing is approximately $300-$500. This is a non negligible cost that continues to grow with activity and number of markets on the venue.
Operational complexity is higher, as trading APIs differ markedly from those of centralized venues, adding layers of integration complexity for exchanges and market makers, including:
Market makers must establish their own RPC infrastructure, as opposed to merely sending an HTTP request to centralized exchange servers.
JSON RPC is standard among decentralized applications, in contrast to the REST/Websocket protocols used by centralized exchanges.
This situation is further complicated by the unique nuances of each blockchain on which market makers operate. Implementing our proposed rollup architecture would allow us to abstract away many of these complexities, thereby lowering the barrier to entry. This would enable more market makers to participate, ultimately enhancing the liquidity available to users. It's essential to recognize that improvements in liquidity can be meaningful beyond their dollar value; increased liquidity works as a catalyst within exchanges, to drive trading activity which, in turn, drives increased liquidity. Therefore, ensuring optimal liquidity provisions mechanisms is critical for the growth flywheel of the exchange.
Last updated