[BBLISA] Rent/Borrow Fluke or similar for Cat6 horizontal testing

[Jon Young]

Thought I'd post a follow-up to this message... We got lucky and were able
to borrow a decent fluke to test a sample of the questionable runs.  Many
devices on this brand new network were experiencing various complex issues
that might be attributable to network errors.  These particular horizontal
runs were questionable because the devices were auto-negotiating at speeds
below 1Gb and the primary goal was to determine if the physical layer was
the likely cause or if these were auto-negotiation issues that might
require forcing the ports.  I have no direct access to the network switches
at this location and must rely on 2nd hand information for switch
configuration and logs... in this case I'm informed but cannot verify no
line errors were reported.

To Mr. Harvey's suggestion:
1.  These devices were already not auto-negotiating to 1Gb.  Putting a
different NIC on each end only tells me that particular combination can
auto-negotiate and does not confirm the reason the required devices are
having issues isn't physical layer.  A particular combination of chipsets
may very well succeed where another fails when cable doesn't meet spec.
 Also, a physical layer problem might be experienced intermittently if the
cabling is at the edge of 'working', a proper test suite should give vastly
superior data.

2.  Expecting a cabling contractor to return and remediate cables at their
own expense will not happen with EIA/TIA based testing.  Cabling
contractors contracts are written to these specs and giving them evidence
of issues that does not include testing to these specs will not result in a
productive conversation, just a large bill.  There are only a hundred or
two runs that need to be checked on this site but consider the cost of a
pair of union contractors in a major city testing and remediating a hundred
cables... I would anticipate a cost of at least $100/cable if not more.  No
contractor will provide $10k+ of labor for free without good reason to
believe their contract requires it.

3.  In some large distributed enterprises, one cannot have access to
everything directly and must coordinate and work with data provided by
another group in that enterprise and sometimes those other groups choose be
less helpful than is ideal.  This particular project is part of a $2B
construction project in a government sized entity on sovereign ground.  The
nature of that beast requires very careful handling and working to a
particular and contracted spec is exceedingly important.

Your suggestion implies the inverse.  Would you accept a pair of laptops
passing some packets as an adequate cable test report on new horizontal
cabling?  I sure wouldn't... EIA/TIA specs are there for a reason.

FWIW, the sample testing I did showed the vast majority of the questionable
runs fail cat 6 spec testing.

Thanks, Jon


On Fri, May 31, 2013 at 3:16 PM, Edward Ned Harvey (bblisa4) <
bblisa4 at nedharvey.com> wrote:

> > From: Bob Webber [mailto:webber at panix.com]
> >
> > If that's your idea of reasonable test gear, it's not surprising that
> you think
> > that you need access to both ends of the cable and that using two
> laptops is
> > better than using proper test equipment.
>
> I am telling you, after the contractors used professional equipment, it's
> not uncommon that I still find problems, and the most reliable way to test
> it is in fact with ethernet cards.  Because that's what you care about
> using anyway.
>
>
> > Will you please do everybody a favour, and particularly do your own
> > reputation a favour, and stop posting about network subjects on this
> list?
> > *snip*
>
> The personal attack is not only complete bullshit, also completely
> inappropriate and not permitted on this list.  Go back under your bridge.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.bblisa.org/pipermail/bblisa/attachments/20130606/be7b1717/attachment.htm