how much bandwidth is t2.micro instance type?
Because I want to do TCP optimize, tune my debian server, like this tutorial and this tutorial
- 最新
- 最多得票
- 最多評論
Hi,
This is actually quite difficult to answer. T2 instances in AWS are "burstable" instances and have the ability to scale with "CPU Credits". Bandwidth can be directly correlated to the credit system since CPU can affect overall throughput. So you would have baseline performance, but that could increase depending on your CPU credits, thus "low to moderate" as far as network performance.
please follow my screenshot to check docs instance type
although you say it is burstable. I still want to know what range does it support. 1M - 1G? 1G - 10G? or 40G - 100G?
Given that the instance is rated "low to moderate"; and the other answers (which boil down to: "performance is variable because it is a burstable instance" you can safely assume that "low to moderate" means less than "up to 10 Gigabit" but it is variable so there's no specific number. The better question here is "what bandwidth do you need?"
You should always benchmark your application to ensure you meet your needs. In addition, we constantly work to improve performance, and they may change (upward) without notice.
Have a look at https://cloudonaut.io/ec2-network-performance-cheat-sheet/ - it's a few years old now but may still be useful. There it says t2.micro tested as 0.06 Gbit/s sustained with 0.72 Gbit/s bursts. The "sustained" performace roughly doubled with each instance size increment.
your article is good. but I don't need benchmark bandwidth, because it is test data. I do need baseline bandwidth. this data is from AWS network hardwares and ISP. now do you understand my meaning?
I think I understand you're after definite numbers provided by AWS, not numbers that third parties have benchmarked which have a lot of confounding factors. But TBH I think you've got all the AWS-provided info you're going to get. We've worked with AWS for many years in a large enterprise and haven't come across anything more reliable to work with that these benchmarks.
I will post this here for you convenience - you can query the network performance for instance types using this command
aws ec2 describe-instance-types --filters "Name=instance-type,Values=c5.*" --query "InstanceTypes[].[InstanceType, NetworkInfo.NetworkPerformance]" --output table U will get something like this
-------------------------------------
| DescribeInstanceTypes |
+--------------+--------------------+
| c5.xlarge | Up to 10 Gigabit |
| c5.4xlarge | Up to 10 Gigabit |
| c5.12xlarge | 12 Gigabit |
| c5.metal | 25 Gigabit |
| c5.24xlarge | 25 Gigabit |
| c5.9xlarge | 10 Gigabit |
| c5.large | Up to 10 Gigabit |
| c5.2xlarge | Up to 10 Gigabit |
| c5.18xlarge | 25 Gigabit |
+--------------+--------------------+
Unfortunately for t2 instance it returns this table DescribeInstanceTypes |
+-------------+-------------------+
| t2.2xlarge | Moderate |
| t2.micro | Low to Moderate |
| t2.large | Low to Moderate |
| t2.xlarge | Moderate |
| t2.medium | Low to Moderate |
| t2.nano | Low to Moderate |
| t2.small | Low to Moderate |
I cannot find the documentation now but I believe I at some point I saw somewhere 10GB (High), 1GB (Moderate) and 100MB (Low). I will dig a bit more and add more info if I find it
where is somewhere?
every one, these days, I did test and test how "low to moderate" is. I just use iperf3 to test my two t2.micro instances bandwidth, and those two vms are still in my free hours. but my aws account incur costs.
Looks like maybe you were running in two AZs, so got some cross-AZ data transfer charges.
相關內容
- 已提問 6 個月前
- 已提問 1 年前
AWS 官方已更新 1 年前- AWS 官方已更新 10 個月前
- AWS 官方已更新 1 年前
Even if it is burstable, it also have a range. I have no money to cost it for unlimit burstable. do you understand? I hope AWS can give me this range. "low to moderate" is very confused to many peoples like me. I am very angry.