Ethereum Benchmarks

Benchmark Date: 08/24/2023

Providers included

  • Nodies

  • Blast API

  • Infura

  • Alchemy

RPC Method Tested

eth_getBlockByNumber

Flood Command

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

Benchmark Summary

Benchmarking the Ethereum mainnet endpoints across various request loads reveals the following observations:

  • P50 Latency: Nodies consistently showcased optimal performance, maintaining sub-10ms latency even at the highest request rate of 10,000 rps. BlastAPI also demonstrated commendable latencies, particularly from 100 to 1,000 rps. Infura's latency rose drastically to nearly 77.7 seconds. Alchemy maintained latencies between 106ms and 131ms.

  • P90 Latency: Nodies remained the top performer with its latency under 20ms across all request loads, except at 10,000 rps where it still remained below 4.5 seconds. BlastAPI and Alchemy demonstrated similar trends, though BlastAPI consistently exhibited lower latencies. Infura's latency surged to 85.9 seconds at 10,000 rps.

  • P99 Latency: The differences become more pronounced. Nodies' performance remained under 24.3 seconds even at the highest load. BlastAPI and Alchemy showed relatively low latencies compared Infura. Notably, Infura's latency at 10,000 rps was close to 89 seconds.

In conclusion, the Nodies endpoint demonstrated top-notch performance across all percentiles and request loads. BlastAPI also fared well, followed by Alchemy. Infura's performance, especially at the P50 and P90 percentiles for 10,000 rps, requires attention due to its extraordinarily high latency.

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)AlchemyBlastAPIInfuraNodies

10

117.363

27.781

103.125

8.920

100

112.768

9.051

96.685

8.247

1,000

106.054

9.440

98.897

7.823

10,000

130.745

112.028

77680.280

165.787

P90 Latency (in ms, lower is better):

Rate (rps)AlchemyBlastAPIInfuraNodiesGrove

10

132.747

58.542

137.230

17.094

45.268

100

122.150

21.912

102.596

10.637

37.459

1,000

119.703

16.032

131.296

11.265

37.298

10,000

179.931

159.137

85865.594

4473.021

21106.584

P99 Latency (in ms. lower is better):

Rate (rps)AlchemyBlastAPIInfuraNodiesGrove

10

184.018

92.841

270.613

32.245

78.539

100

380.943

51.079

460.038

13.988

71.516

1,000

376.071

33.736

879.284

18.376

75.879

10,000

1426.869

1108.877

88990.159

24312.414

30004.877

Last updated