MTBF

[w.t.kim]

Hi, 

I have a question.

The MTBF of the C6059-E chassis is 348,935 and the MTBF of the NEXUS9508 is in the datasheet at 928,910 (Hours).

928,910 / 24 (hours) = 38704 (day) / 365 (year) = 106 (year)

I do not understand.

Please let me know.

[Jonathan Unger]

The MTBF value for the 9508 is listed for the chassis only (which is mainly a hunk of metal with a bunch of traces inside). This is why the value is so high (there isn't really much to fail there).

The MTBF on the other components such as the power supply is much lower at 287,097 hours (N9K-PAC-3000W-B).

Does this help?

[Joseph W. Doherty]

What don't you understand?  For example, why the Nexus MTBF is so much better than the 6500's or why both run into years?

If the former, MTBF is a statistical engineering estimate.  Often newer equipment has better MTBF because the build technology has been approved (and/or there's more field experience for the equipment and MTBF can be extended).

For the latter, you have to carefully consider what the MTBF "covers".  For instance, in a chassis like the 6500, the IOS, the chassis, the sup, the line card and the power supply all have their own MTBF, each might be very high, but if one of them fails, the whole stops, so when you have a "series" of components, each component added to the series tends to lower the whole's MTBF.

PS:

There's a Cisco white paper that explains how to compute overall expected uptime, for a 6500 chassis, using component MTBFs.

[w.t.kim]

Thank you for answer.

I understood how to explain to the customer.