12

I am under the impression that you should use optical for 10GbE networking because of a BER (bit error rate) several orders of magnitude lower than copper.

I'm pretty sure I previously found documentation stating so, but for the life of me I just cannot find it.

I'd like to see documentation on typical BER (in general, if such a thing exists) for:

  • 10GBase-T ports (with Cat6a cabling)
  • 10GBase-CX4 ports
  • 10GBase-T SFP+ modules (Cat6a cabling)
  • Fiber SFP+ SR modules
  • DAC SFP+ modules
MikeyB
  • 38,725
  • 10
  • 102
  • 186
  • 5
    I think distance and cost requirements may factor into your decision matrix as well. – Mike Pennington Aug 03 '12 at 14:51
  • BER by itself is a meaningless metric when comparing technologies. It only has meaning when comparing the same product from different vendors. By the way: Copper SFP+ is misleading: These are eiter 10Gbase-T or Twinax (which is the same as DAC). – Tonny Aug 03 '12 at 15:03
  • Distance and cost requirements will factor into the overall decision, yes, but the purpose behind this question is to gain insight into the BER of various media. – MikeyB Aug 03 '12 at 15:37
  • BER is absolutely -not- meaningless. There is a two order of magnitude difference between twisted pair and twinax. This is why, for example, FCoE isn't supported on 10GBaseT yet. – rnxrx Aug 03 '12 at 19:36
  • Great! Can you please provide documentation? – MikeyB Aug 03 '12 at 19:47
  • Related: http://serverfault.com/questions/195111/why-would-i-choose-copper-over-sfp-for-10gbe?rq=1 – MikeyB Aug 15 '12 at 22:09

1 Answers1

7

Twinax is rated at 10^-17 at <= 10M

Fiber is generally 10^-18 or better (lot of variables here, though - especially on long haul)

Ethernet's minimum specification is 10^-12. This is where the contention comes in as far as FC goes - it may be that a TP infrastructure can support better than 10^-12 (and hopefully most do) but at 10G speeds even that much error translates to a bad packet every few minutes.

rnxrx
  • 8,103
  • 3
  • 20
  • 30
  • 2
    [*The long-term prognosis forFCoE over 10GBaseT looks very promising*](http://sniaesfblog.org/?p=139) – Mike Pennington Aug 03 '12 at 20:40
  • Yes - absolutely. Smaller die sizes mean DSP's are improving, which improves BER and power issues. Moore's law works for more than just CPU's! I wouldn't be surprised if we saw movement in this regard in the next 12 months. – rnxrx Aug 03 '12 at 20:57