A new Cardano scaling test from Input Output compared performance on Hydra, the protocol's layer-2 solution. The October 22 report replayed real application transactions to measure improvements in the newer Hydra version, offering insight into why benchmark gains differ from theoretical maximum speeds.
Test Results and Methodology
The test processed 24,576 generated claim transactions containing scripts from Cardano's distribution pipeline. Throughput rose from 42 to 79 claims per second between versions. Total completion time fell from 590.4 to 339.1 seconds, a reduction of approximately 43%.
Input Output held the same host and workload constant across versions, making the software comparison more reliable than demonstrations run on different hardware with different transaction complexity.
What the Benchmark Measures
The result describes performance on a specific workload, not a new transaction-speed figure for all of Cardano. Hydra processes transactions among its own participants before settling with the main Cardano chain, reducing the amount of work replicated globally for each application interaction.
The test used script-heavy claim transactions, which require validity checks during execution. This makes the benchmark more informative than counting empty transfers, as it reflects actual computational work applications perform.
Limitations of Controlled Testing
A controlled replay does not recreate every production condition, including dispersed participants, unreliable connections, or changing demand. Application builders assessing a specific job would benefit from reproducing their own scripts, number of participants, and expected submission patterns.
The published result supports a specific conclusion: this claim pipeline completed materially faster on the newer Hydra implementation. The improvement is meaningful for teams planning similar distributions, though additional testing would be needed to validate deposits, withdrawals, and recovery procedures.
