Starfish adds erasure coding and DAG-based data certificates to improve IOTA’s payload availability under heavier validator load. Sui’s Mysticeti V2 reduces product latency by moving transact
IOTA’s move toward Starfish shows a clear split in how blockchain networks now approach consensus design. IOTA has taken a different path by focusing on stronger liveness, better data availability, and steadier performance during network stress. Meanwhile, Sui has continued to reduce latency via Mysticeti V2.
Mysticeti removed explicit certification from the consensus process and allowed the DAG structure itself to act as a virtual certificate. That design helped cut latency and reduce the overhead created by repeated signatures and certification rounds.
However, the same design also left open technical concerns. Without explicit certification, block availability no longer comes “for free.” Validators may need to fetch missing data from peers, which can strain the network during heavy load. Researchers also raised concerns about liveness, especially when validators move between rounds without producing their own blocks. Those gaps created the split now seen between Sui and IOTA.
Sui’s Mysticeti V2 treats the core consensus engine as already close to optimal. Instead of rebuilding the base layer, Sui targets the extra transaction-validation process around consensus. The network moves more validation work into the consensus flow and uses a Transaction Driver to reduce delays that came from the older Quorum Driver model.
If you want to get deeper technical knowledge of our protocol, and specifically Starfish – @kowei1995 has got you covered with this excellent article. Have a read! https://t.co/wdTklasvvt
— IOTA (@iota) May 26, 2026
That approach fits Sui’s focus on consumer apps, DeFi, and high-throughput activity. Lower product latency matters for users who expect fast swaps, games, payments, and on-chain interactions. Mysticeti V2, therefore, keeps the speed race alive by cutting surrounding friction rather than changing the core dissemination model.
IOTA’s Starfish treats dissemination and liveness as the weaker parts of the system, not the surrounding transaction layer. Instead of pushing only for faster finality, Starfish rebuilds how data moves across validators and how the network keeps progressing when conditions worsen.
Starfish separates block metadata from transaction payloads. Headers carry the information consensus needs quickly, while heavier payload data can move in a more controlled way.
Starfish also uses Reed-Solomon erasure coding to split transaction data into recoverable pieces. Each payload gets split into fragments, and validators can reconstruct the full data from enough valid pieces. This design reduces full-data duplication while still supporting availability. Starfish then uses Data Availability Certificates that grow naturally inside the DAG, rather than adding a separate certification round to every block.
A Push pacemaker also plays a central role. Validators must produce their own block before moving forward, which reduces holes in the DAG. That rule targets the liveness issue seen in uncertified DAG designs, where too many missing blocks can stop the network from forming the needed witness and confirmation patterns.
Starfish does pay a latency cost. Under cleaner comparisons, Mysticeti can sit near four message delays in practical settings, while Starfish sits closer to five. In harsher scheduling conditions, both can stretch further. Even so, IOTA accepts that added delay in exchange for stronger availability guarantees, tighter tail behavior, and a more formal path toward liveness.
Meanwhile, this month, IOTA started the first ADAPT rollout in Kenya, Morocco, and Nigeria to support digital trade under AfCFTA. The project targets Africa’s $100 billion trade finance gap by linking digital identity, payments, and cross-border data exchange.
The post Why IOTA Chose Robustness While Sui Chose Lower Latency appeared first on ETHNews.
