[c-nsp] 10g Copper Transceivers for SPF+

[Tom Hill]

On 26/05/17 15:24, Nick Cutting wrote:
> I got a couple of greedy replies from traseiver vendors, but nothing
> from the wise old network wizards.

ProLabs were very nice to me; greed isn't really the problem - these
things won't be getting made at any scale yet.

> The GLC-10G-T - which seems to fool the switch into thinking it is SR
> , so yes I agree with the naysayers it sounds like a bad idea from
> the get go.
> 
> Is anyone using them, or has been using them?

Yes, and their reason for masquerading is generally because most vendors
haven't yet caught up with the idea yet, or don't have their own parts.
Whereas they will commonly have 10G DAC, or SR products. The reported
part number is just read from the EEPROM, and doesn't really affect how
they work in the device.

I have a pair of them. One pretends to be a DAC and the other pretends
to be an SR optic, and there's very little difference in behaviour. The
3560E whinged a bit about the transceiver not being valid, but then
pulled it up and worked just fine anyway - and that was through
(official) OneX converters, too.

The ASR9000s (Typhoon cards) had no problems with them at all, and
worked a treat over ~20m of internal patching within the DC. Didn't use
it for long, as I was just testing them out whilst we were commissioning
that hardware. No errors though!

> Reason being it would be a great way to uplink our old switches with
> SPF+ modules to our 10g copper nexus Borders, without using a
> breakout cable on the 100g ports.

Given their cost, you might find 40G breakout cheaper. Handy things to
have about for a pinch though, I must say.

I had plans to test them on slightly more daring cable runs, but I
haven't yet gotten around to it. :)

-- 
Tom