Hi I have heard that the small EC2 instances are limited to around 35mbit/s. Does anyone have any figures for the different instance types and the size of their pipes?

We will be planning on using elastic load balancing but I'd really like to know what the limits of the external interfaces of the instances will be also.

3 Answers 3


I wanted an updated answer, so here's what I found by aggregating a variety of benchmark sources:

Speed roughly scales with instance size for m3/c3/c4/r3 instances, with m1.* being exceptions (much faster):

  • t2.nano = ??? (based on the scale factors, I'd expect ~25-30 MBit)
  • t2.micro = ~70 MBit? (qiita says 63 MBit/s) - t1.micro gets about ~100 Mbit
  • *.small = ~150 MBit (t2, qiita says 127 MBit)
  • *.medium = ~300 MBit (m1 is better at ~900 MBit, m3.medium is ~400 MBit)
  • *.large = ~450-550 MBit (m3 is better at ~700 MBit, m1.large is ~800 Mbit, t2 tracks this at ~500 MBit)
  • *.xlarge = 700-900 MBit
  • *.2xlarge = 1-1.2 GBit
  • *.4xlarge = ~2 GBit
  • *.8xlarge = 10 Gbit (published)
  • This is a really old question that would likely not be considered on-topic nowadays. Mar 4, 2016 at 21:58
  • 3
    @KatherineVillyard Is there a better place to supply the information? I had to go scouring the web to gather it and would like to save someone else the pain.
    – BobMcGee
    Mar 4, 2016 at 22:11
  • This is still very much on-topic. @BobMcGee if you have the latest benchmarks for the latest generation, I would appreciate that. Thanks
    – Houman
    Jan 14, 2018 at 19:35
  • @Houman most of the latest instance types announce their capacity explicitly, and it's often 10 GBit+
    – BobMcGee
    Feb 26, 2018 at 15:08
  • @BobMcGee Thank you, is it right under Ec2 pricing where this is shown?
    – Houman
    Feb 26, 2018 at 15:14

I had running small instance as load balancer with traffic peaks to 80Mbit. So i guess there is no limit. Also thereis a thread, that states https://forums.aws.amazon.com/thread.jspa?threadID=15534 that it is 250Mbit bandwith for all.


Most instances have 1 Gbit shared with the other instances on that server, so your available bandwidth can vary quite a bit if the other tenants are busy beavers.

Cluster compute and GPU instances have 10 Gbit, again shared with any other instances on the host.

  • Upvoting because of the reference to the bandwidth being shared, however the link in the other answer states that each server only has a 0.25gb dedicated link do you have a source for this? Thanks!
    – scsmith
    Apr 14, 2012 at 8:14
  • I'll see if I can dig up the source (it was a Twitter conversation involving an AWS evangelist). The 10 Gbit info can be found at aws.amazon.com/ec2/instance-types.
    – ceejayoz
    Apr 14, 2012 at 13:45

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .