I'm having network issues as well. I see:

3. G1-7-4-6.WASHDC-LCR-22.verizon-gni.net 0.0% 77 13.9 11.3 6.9 22.3 2.7
4. so-3-1-0-0.RES-BB-RTR2.verizon-gni.net 0.0% 77 13.8 19.7 7.0 85.6 19.5
5. 0.xe-4-2-0.BR1.IAD8.ALTER.NET 0.0% 77 8.1 12.6 6.9 79.2 11.1
6. dcp-brdr-03.inet.qwest.net 94.7% 77 11.6 11.4 10.6 11.9 0.6
7. dca2-edge-01.inet.qwest.net 98.7% 77 11.3 11.3 11.3 11.3 0.0
8. 65.120.78.82 0.0% 77 11.0 12.2 8.9 37.6 3.5
9. 72.21.220.151 14.5% 77 13.5 13.0 9.2 27.9 2.8
10. 72.21.222.87 11.7% 77 10.9 12.3 9.1 19.1 2.1
11. ???
12. ???
13. ???

And connecting via ssh etc. is taking up to 30s

I am getting around 3MB/s from the link you posted. However I am having throughput issues from time to time between my AWS servers and my location here in nyc via verizon fios.

Its been going on for me for at least a couple of months now.

I am wondering if it has anything to do with verizon trying to squeeze netflix to pay them for data transfer.

I'm seeing exactly the same thing. Verizon claims it's not them.

At home I have internet from Time Warner. Connections to AWS are fast.

Seeing this also from Verizon FiOS as well as their gigabit service feeding our company.
Have reported on NANOG and will be reaching out to Verizon and AWS tomorrow.
Are all of you still seeing this performance hit?

Adding to the queue here. I'm seeing unstable and unreasonably slow connections between my local network (Verizon FIOS) and remote hosts on all N. Virginia regions. Outbound and inbound traffic to/from other networks (Linode, Digital Ocean, et al.) are running fine. Traceroute output attached. I've opened an issue with AWS support and will refer them to this thread.

Seems to be a Verizon/NTT thing. We were able to work with Amazon's Peering Team to route certain traffic a different way and things vastly improved. This was for our Gig circuit, however. Consumer Verizon FiOS circuits still seem to be passing through the "bad" path and as a result are slow.

Not sure how this gets resolved except to have the peering providers figure something out. We certainly are anticipating that the issues for our Gig circuit could return at any time even via the new path we're on.

We've been seeing the same exact problem too, Verizon FiOS to Amazon AWS: the connection just completely stops working for several seconds at a time. Then picks back up again, but then drops again for several seconds, over and over. We've been seeing this problem for a couple of months now.

I'm hoping that someone at AWS or Verizon can take a look at this issue.

After suffering with this problem for months and months, just now it seems to have cleared up. Are the other posters seeing the problem cleared up too?

I can only hope this fix is for good. Shew...

I created an S3 bucket in the N. Virginia region last month. D/L times to my Verizon FIOS home acct are horrible… 40-100KB/sec. If I bring up a VPN (to practically anywhere), I can d/l at full speed. Traceroute times look fine… RTT looks actually worse over VPN, as would be expected -- but maxes out my bandwidth regardless. Verizon says it's not their problem.

I'm sure this is more corporate slap-fight, a la the Verizon v. Netflix disputes. However, it's got me rethinking using S3, since the content I was going to store in it is primarily serving users in Verizon territory...

I am wondering whether anyone has found a solution to the slow connection speed via FiOS to EC2 servers in the us-east region.

I am a business start-up, connecting to EC2 from New York City via two residential internet providers, TimeWarner and Verizon FiOS. I have a second ISPs as back-up.

My FiOS connection should work much faster and indeed does for most places on the web, but not to the EC2 server I am using.

For instance, when I download data via my RDP connection to my local client, FiOS gives me currently a download speed of 60 kbytes/sec while TimeWarner yields about 7000 kbytes/sec.

Using speedtest.net, however, I get for FiOS a ping rate of 6ms, download speed of 84 Mbytes/sec and upload speed of 92 Mbytes/sec. The TimeWarner results are: ping rate of 16ms, download speed 122 Mbytes/sec, upload speed of 14 Mbytes/sec (so the FiOS ping and upload results are much better, but the download result is slightly worse).

Thanks in advance.

Best, Heiko

Edited by: heikoebens on Feb 28, 2015 9:23 PM

Edited by: heikoebens on Feb 28, 2015 9:33 PM

This has been an ongoing issue for me, too. Using a VPN which sometimes gets 8 MB/sec, which is fine. Most times over our 75/75 Mb FiOS connection though, I am getting 200-400 KB/sec.
FiOS business package, just west of Phila. proper. Elsewhere on the web or via FTP, 4 to 12 MB/sec.

-Jeff B.

I've had this issue for a little over a year now.
I've been after VZ for quite some time to resolve this issue.
I kept complaining to them until they sent a network technician to my home. After I've showed him all the evidence i've gathered we ran a few tests (to prove that i wasn't bsing him). He made a phone call back to VZ, and I was basically told that this is a "corporate politics issue". He suggested that the more people complain the more they'll be willing to "fix" it. I'm so frustrated by this issue. We have app servers deployed on the AWS cloud that are heavily affected by this.

Observations:
-garbage download speeds (5k to 70Kb/s download speed during normal business hours)
-upload speeds seem unaffected
-region US-NORTH EAST
-routing traffic through VPN connection seems to fix the download speed issue.

****TIP: call VZ and insist on a level 2 or "network" technician to come to your place and test AWS connection speeds. Do this even if you know they won't resolve the problem. The goal is to open a ticket with them. They wont be able to close the ticket until YOU are satisfied with the answer. The more open tickets relating to this issue, the more noise we'll create. We pay VZ enough money on a monthly basis, we really don't need to be wasting our time on internet bandwidth speeds, we're not in 90's anymore!

While I lay blame with Verizon based on my attempts to raise the issue with them, I am curious to hear if Amazon is actively working to try and resolve the issue through its partners. I recall seeing a post from an Amazon rep on the same issue from a year or two ago in which they did adjust the peering to provide some relief. Unfortunately, this persistent issue is soon forcing me to migrate away from AWS to another provider simply as a practical matter in order to achieve a resolution.