During a firmware upgrade, I've noticed one problem: An equipment keeps linking/unlinking on a specific port of a PowerConnect 3548(stp portfast, switchport access mode). I've contacted the local support of that remote office and they will try to isolate that equipment so, this isn't the problem.

I'm just curious since i could not find this information on this switch specs/manpages and thus, my question:

Is there a generic MTBF for "linking/unlinking" operation cycles on switches? Maybe a MTBF based on the ethernet module chip?

In this specific case, the equipment seems to be linking one time per second.


07-Jan-2000 13:58:11 %LINK-I-Up:  e38
07-Jan-2000 13:58:11 %LINK-W-Down:  e38
07-Jan-2000 13:58:13 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:13 %LINK-I-Up:  e38
07-Jan-2000 13:58:14 %LINK-W-Down:  e38
07-Jan-2000 13:58:16 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:16 %LINK-I-Up:  e38
07-Jan-2000 13:58:17 %LINK-W-Down:  e38
07-Jan-2000 13:58:18 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:18 %LINK-I-Up:  e38
07-Jan-2000 13:58:19 %LINK-W-Down:  e38
07-Jan-2000 13:58:21 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:21 %LINK-I-Up:  e38
07-Jan-2000 13:58:22 %LINK-W-Down:  e38
07-Jan-2000 13:58:24 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:24 %LINK-I-Up:  e38
07-Jan-2000 13:58:24 %LINK-W-Down:  e38
07-Jan-2000 13:58:26 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:26 %LINK-I-Up:  e38
07-Jan-2000 13:58:27 %LINK-W-Down:  e38
07-Jan-2000 13:58:29 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:29 %LINK-I-Up:  e38
07-Jan-2000 13:58:30 %LINK-W-Down:  e38
07-Jan-2000 13:58:32 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:32 %LINK-I-Up:  e38
07-Jan-2000 13:58:33 %LINK-W-Down:  e38
07-Jan-2000 13:58:34 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:34 %LINK-I-Up:  e38
07-Jan-2000 13:58:35 %LINK-W-Down:  e38
07-Jan-2000 13:58:37 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:37 %LINK-I-Up:  e38
07-Jan-2000 13:58:38 %LINK-W-Down:  e38
07-Jan-2000 13:58:40 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:40 %LINK-I-Up:  e38
07-Jan-2000 13:58:41 %LINK-W-Down:  e38
07-Jan-2000 13:58:42 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:42 %LINK-I-Up:  e38
07-Jan-2000 13:58:43 %LINK-W-Down:  e38
07-Jan-2000 13:58:45 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:45 %LINK-I-Up:  e38
07-Jan-2000 13:58:46 %LINK-W-Down:  e38
07-Jan-2000 13:58:48 %STP-W-PORTSTATUS: e38: STP status Forwarding
07-Jan-2000 13:58:48 %LINK-I-Up:  e38
07-Jan-2000 13:58:48 %LINK-W-Down:  e38
  • MTBF as in mean time between failures? If it's not in the spec sheet it's nowhere. – Zac67 Apr 1 '19 at 17:22
  • Did any answer help you? If so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you can provide and accept your own answer. – Ron Maupin Dec 15 '19 at 1:25

Mean Time Between Failures is a statistic value provided for a complete device or sometimes for an assembly unit.

Nobody breaks that down to specific parts of a device or specific types of failures (e.g. link flapping). It's simply not practical.

When looking at the components' MTBFs (if available) you have to take into account that it's not only the component that might be failing but it could also be the power supply, another controlling component or similar.

MTBF calculation is a somewhat complex process and can only be done with considerable insight into all device components, their individual MTBFs and their interaction - the complete design. More complex designs can only be tested in whole and the MTBF statistically extrapolated from a sample error rate. Check out MTBF on Wikipedia for some details.

For an Ethernet link to fail without anything broken (including the cable), the receiver's carrier detection had to lose its synchronization with the transmitter.

Ethernet PHYs aim at bit error ratios (BER) of 10^8 (10 Mbit/s) to 10^12 (10+ Gbit/s). For the carrier to actually fail, several consecutive bits in a PCS code block would need to be mistransmitted. If we assume we need at least four bits to fail that probability is 10^12^4 or 10^48. Even for a 100G interface that won't ever happen.

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy