EC2 Network Performance Cheat Sheet
EC2 Network Performance Cheat Sheet
This vast compilation of benchmark runs considers almost every EC2 instance type and gives you an idea about what kind of throughput to expect.
Join the DZone community and get the full member experience.
Join For FreeLearn how to migrate and modernize stateless applications and run them in a Kubernetes cluster.
What is the maximum network throughput of your EC2 instance? The answer to this question is key to choosing the type of an instance or defining monitoring alerts on network throughput.
Unfortunately, you will only find very vague information about the networking capabilities of EC2 instances within AWS' service description and documentation. That’s why I ran a network performance benchmark for almost all EC2 instance types within the last few days. The results are compiled into the following cheat sheet.
| INSTANCE TYPE | MIN (Gbit/s) | MAX (Gbit/s) | AVG (Gbit/s) | STDDEV (Gbit/s) |
|---|---|---|---|---|
| c4.large | 0.662 | 0.663 | 0.662 | 0.000 |
| c4.xlarge | 1.337 | 1.338 | 1.337 | 0.000 |
| c4.2xlarge | 2.519 | 2.522 | 2.521 | 0.001 |
| c4.4xlarge | 5.036 | 5.037 | 5.037 | 0.000 |
| c4.8xlarge | 5.035 | 9.861 | 5.840 | 1.877 |
| c5.large | 10.043 | 10.046 | 10.045 | 0.001 |
| c5.xlarge | 10.044 | 10.046 | 10.045 | 0.000 |
| c5.2xlarge | 10.044 | 10.049 | 10.046 | 0.002 |
| c5.4xlarge | 10.043 | 10.102 | 10.055 | 0.020 |
| c5.9xlarge | 10.044 | 10.050 | 10.047 | 0.002 |
| c5.18xlarge | 24.607 | 24.614 | 24.611 | 0.002 |
| d2.xlarge | 1.336 | 1.338 | 1.337 | 0.000 |
| d2.2xlarge | 2.519 | 2.523 | 2.521 | 0.001 |
| d2.4xlarge | 5.035 | 5.038 | 5.037 | 0.000 |
| d2.8xlarge | 5.036 | 9.861 | 9.216 | 1.525 |
| g3.4xlarge | 10.094 | 10.103 | 10.099 | 0.003 |
| g3.8xlarge | 10.097 | 10.103 | 10.100 | 0.003 |
| g3.16xlarge | 20.640 | 24.635 | 23.755 | 1.324 |
| h1.2xlarge | 10.097 | 10.102 | 10.100 | 0.002 |
| h1.4xlarge | 10.097 | 10.103 | 10.100 | 0.003 |
| h1.8xlarge | 10.097 | 10.103 | 10.100 | 0.002 |
| h1.16xlarge | 20.739 | 24.636 | 23.805 | 1.458 |
| i3.large | 8.706 | 10.099 | 9.945 | 0.405 |
| i3.xlarge | 9.634 | 10.097 | 10.053 | 0.126 |
| i3.2xlarge | 10.093 | 10.102 | 10.098 | 0.003 |
| i3.4xlarge | 10.094 | 10.103 | 10.098 | 0.003 |
| i3.8xlarge | 10.094 | 10.102 | 10.099 | 0.003 |
| i3.16xlarge | 17.261 | 24.609 | 23.554 | 2.297 |
| m3.medium | 0.334 | 0.336 | 0.335 | 0.001 |
| m3.large | 0.707 | 0.711 | 0.709 | 0.002 |
| m3.xlarge | 1.027 | 1.032 | 1.029 | 0.002 |
| m3.2xlarge | 1.027 | 1.031 | 1.029 | 0.002 |
| m4.large | 0.488 | 0.489 | 0.489 | 0.000 |
| m4.xlarge | 0.840 | 0.841 | 0.841 | 0.000 |
| m4.2xlarge | 1.031 | 1.035 | 1.033 | 0.002 |
| m4.4xlarge | 2.064 | 2.068 | 2.066 | 0.002 |
| m4.10xlarge | 5.036 | 9.860 | 5.725 | 1.751 |
| m4.16xlarge | 20.120 | 24.633 | 23.913 | 1.348 |
| m5.large | 10.016 | 10.047 | 10.041 | 0.008 |
| m5.xlarge | 10.043 | 10.047 | 10.045 | 0.001 |
| m5.2xlarge | 10.044 | 10.049 | 10.047 | 0.002 |
| m5.4xlarge | 10.044 | 10.050 | 10.047 | 0.002 |
| m5.12xlarge | 10.043 | 10.050 | 10.047 | 0.003 |
| m5.24xlarge | 24.580 | 24.613 | 24.597 | 0.015 |
| p2.xlarge | 1.310 | 1.313 | 1.312 | 0.001 |
| p2.8xlarge | 10.161 | 10.169 | 10.165 | 0.003 |
| p2.16xlarge | 19.294 | 24.635 | 23.385 | 2.126 |
| p3.2xlarge | 10.092 | 10.102 | 10.098 | 0.003 |
| p3.8xlarge | 10.094 | 10.103 | 10.099 | 0.003 |
| p3.16xlarge | 19.719 | 24.636 | 22.619 | 1.885 |
| r3.large | 0.538 | 0.540 | 0.539 | 0.001 |
| r3.xlarge | 0.790 | 0.795 | 0.792 | 0.002 |
| r3.2xlarge | 1.030 | 1.032 | 1.030 | 0.001 |
| r3.4xlarge | 2.057 | 2.059 | 2.058 | 0.001 |
| r3.8xlarge | 5.037 | 9.860 | 5.641 | 1.646 |
| r4.large | 7.119 | 10.099 | 9.770 | 0.893 |
| r4.xlarge | 10.011 | 10.099 | 10.092 | 0.021 |
| r4.2xlarge | 10.094 | 10.102 | 10.098 | 0.003 |
| r4.4xlarge | 10.094 | 10.102 | 10.099 | 0.003 |
| r4.8xlarge | 10.095 | 10.103 | 10.099 | 0.003 |
| r4.16xlarge | 20.773 | 24.636 | 23.758 | 1.279 |
| t1.micro | 0.093 | 0.094 | 0.094 | 0.001 |
| t2.nano | 0.496 | 0.497 | 0.497 | 0.000 |
| t2.micro | 0.988 | 0.992 | 0.991 | 0.001 |
| t2.small | 0.985 | 0.994 | 0.992 | 0.002 |
| t2.medium | 0.989 | 0.997 | 0.996 | 0.002 |
| t2.large | 0.997 | 1.004 | 1.003 | 0.001 |
| t2.xlarge | 1.008 | 1.012 | 1.011 | 0.001 |
| t2.2xlarge | 1.017 | 1.020 | 1.019 | 0.001 |
| x1.16xlarge | 10.163 | 10.168 | 10.166 | 0.002 |
| x1.32xlarge | 17.250 | 23.972 | 20.611 | 2.835 |
| x1e.xlarge | 10.039 | 10.099 | 10.094 | 0.015 |
| x1e.2xlarge | 10.097 | 10.099 | 10.098 | 0.001 |
| x1e.4xlarge | 10.096 | 10.102 | 10.099 | 0.002 |
| x1e.8xlarge | 10.097 | 10.103 | 10.100 | 0.003 |
| x1e.16xlarge | 10.096 | 10.103 | 10.099 | 0.003 |
| x1e.32xlarge | 15.899 | 23.933 | 20.751 | 2.900 |
I’ve run the network performance benchmark in us-east-1 from December 29th, 2017 to January 5th, 2018 using iperf3. A benchmark run took 30 seconds. Each EC2 instance was used for nothing more than to benchmark. The table summarizes at least 10 benchmark runs per instance type. A c5.18xlarge was used as counterparty for the network benchmark of each instance type.
Join us in exploring application and infrastructure changes required for running scalable, observable, and portable apps on Kubernetes.
Like This Article? Read More From DZone
Published at DZone with permission of Andreas Wittig . See the original article here.
Opinions expressed by DZone contributors are their own.
{{ parent.title || parent.header.title}}
{{ parent.tldr }}
{{ parent.linkDescription }}
{{ parent.urlSource.name }}