Hi, I recently installed an SAF Technika microwave system that has links which are 2+0 (2 parallel microwave links). Each link is 450 Mbps and one side connects to a IE-5000 and the other side to an IE-4000.
As long as the bandwidth is 400 Mbps or below, there is no problems and I get steady 400 Mbps throughput, but as soon as the volume of traffic requires operation of both radios, say for 600+ Mbps of traffic, my effective throughput oscillates between 20-100 Mbps.
So I looked at a Pcap and I could see that I was getting lots of retransmissions and noticed that the radios themselves were showing CRC errors in the SAF error logs and the Cisco switches were showing input errors.
Of course I called SAF and they said try it with unmanaged switches (so I got a couple of 5 port netgear switches) and placed them in front of the Cisco switches. And behold, I could get 900 Mbps through the system. Their viewpoint is that see it works with Netgear, the Cisco switches are the problem.
So I called Cisco, we tried using TAPs, etc and could not find the root of the problem either.
Here are my questions:
1. What can explain as to why they don't work with Cisco switches, but do with unmanaged Netgear as pre-switches?
2. Have spent many hours with Cisco on this issue and to their credit, they have always responded as they want to find the root of the problem. I have also spent considerable time with SAF and their approach is always, it isn't a SAF issue.
2. The radios themselves have an internal switch and do a layer 1 link aggregation at each end to combine the bandwidth. I believe that something is wrong with this in the SAF radios at this point, how do I show that? Further explanation is here:
https://www.saftehnika.com/files/downloads/24c36b5f-26da-e511-a0d1-0050569a8c0f/Link_Aggregation%202016%20Sep.pdf
There are a lot of details about this situation, I tried to limit my description here to the areas that I thought were relevant. If there are any other details that would help, please let me know.
G. Brown
