Optimism Benchmarks

Benchmark Date: 08/24/2023

Providers included

• Nodies

• Blast API

• Quicknode

RPC Method Tested

eth_getBlockByNumber

Flood Command

flood eth_getBlockByNumber <endpoint1> <endpoint2> <endpoint3> <endpoint4> <endpoint5> --rates 10 100 1000 10000 --duration 30

Benchmark Summary

When evaluating the Optimism mainnet endpoints across varied request loads, the following observations stand out:

Nodies:

• Demonstrated unmatched performance across all metrics. At 10,000 rps, its P50 latency remains a mere 5.843ms, and its P99 was only 11.353ms.

Other Providers:

• BlastAPI: Offered consistent performance across varying request loads. At 10,000 rps, its P50 latency was 119.213ms, P90 was 181.613ms, and P99 managed to stay under 3.166 seconds.

• QuickNode: Showed competitive results at lower loads but faced considerable challenges at 10,000 rps. Its P50 latency soared to around 20.5 seconds, P90 to approximately 31.7 seconds, and P99 reached a staggering 37.9 seconds.

In conclusion, Nodies leads performance across all percentiles and request loads. Among the other providers, BlastAPI holds up well, consistently delivering respectable latencies. QuickNode, on the other hand, starts strong but struggles to maintain performance at higher request loads.

Latency Benchmarks

Metrics: Latency percentiles (P50, P90, P99) against load rates (10, 100, 1000, and 10000 rps).

P50 Latency (in ms, lower is better):

rate (rps)	BlastAPI	QuickNode	Nodies
10	105.547	94.738	6.550
100	9.540	90.386	6.239
1,000	9.685	91.280	5.951
10,000	119.213	20547.582	5.843

P90 Latency (in ms, lower is better):

rate (rps)	BlastAPI	QuickNode	Nodies
10	112.086	95.598	6.902
100	24.233	92.821	6.525
1,000	21.027	121.969	6.209
10,000	181.613	31681.039	6.115

P99 Latency (in ms. lower is better):

rate (rps)	BlastAPI	QuickNode	Nodies
10	130.303	113.069	7.354
100	109.190	140.482	6.962
1,000	111.537	256.351	6.604
10,000	3166.259	37909.239	11.353

1MB

report.html

Open

Full report includes more indepth benchmarks such as success rates, throughput, etc.