Discrepancy in network performance

Have a 100Mbps ATT ethernet that goes straight to internet in AL.

The Enterprise has a 1Gbps MPLS from CenturyLink that takes our site in AL to Seattle where it goes out to the internet.

On the ATT connection i can get 90-95Mbps with wget from softlayer in seattle. ATT line has a 69ms RTT.

On the CL connection I can only get 2-16Mbps from the same exact softlayer site. CL connection has a RTT of 76ms.

I have tested other sources and have consistent results. We can max our our ATT line but can’t even get more then 30Mbps on the CL line.

Same exact machine gets unplugged from one to the other to do test. I have checked MTR and no dropped packets on either connection. On the CL line I do see a lot of ??? but no packets are being dropped.

I’m trying to work with the enterprise to figure out why the MPLS (CL) is so much slower. They do a test by just sending 200Mbps of UDP traffic from seattle or STL to our site in AL and say that it’s working fine. I don’t get how that can validate the connection?

I have no visibility into what’s happening on the MPLS circuit other than it’s throughput is significantly slower trying to get out to the internet than our connection that has 10x less bandwidth.

What does sending 200Mbps of UDP traffic validate?