Replies to This Post

Permalink Reply by John Hoffman on February 7, 2009 at 22:09

Well its cheaper and smaller. But I see some problems. WHICH OF THESE IS THE BAD GUY?

1. Technically the problem is a straight forward one, albeit, a challenging one. The physics can be calculated ok, and test beds in the labs and field can have test results recorded.

2. Economically there is problem for a client in taking a solution that is outside of a warranty specification. It is a risk! This cost of risk would be too high for the customer or the installer to support.

3. Human Momentum. A barrier to change would be momentum of users and engineers being familiar to the 90 metre rule. E.g the simple rule of ABCbit/s speed of ethernet needs and CategoryXYZ "grade of cable". This is simple inertia and adherance to the familiar. Look up the Easter island coconut headphone story.

4. Market Economics. During the death of the electric typewriter IT has enjoyed a sustained period of growth fuelled by new technology creating new business solutions and value propositions to businesses. And importantly these customers have been able to justify the huge relentlessly increasing costs of IT as it gave them a competitive advantage, shareholder value, or increased profits. Times are changing It is sometimes said to me by procurement managers and Finance Directors that even IT is now a cost of business for the most part. The old analogy being that of " the green shield stamp schemes of the 60s", as soon as everybody had the advantage it was no longer any advantage.

Now is the time for somebody to underwrite a solution that supports high speed on restricted lengths. Although it might seem counter-intuitive that a manufacturer would want to support a warranty that does not drive another IT upgrade cycle, now is the time that somebody with a deep pocket, a good product and a steady nerve could if they wanted steal the lead in the market.

Business would love this innovation as it is brave, beautiful, and honest.

It is time for all those that underwrite the thinner cable over short distance to stand up and say:

"I shout
it loud,
I'm short
and proud"

Come on then, Let's be having you!

Permalink Reply by Tony Khoury on December 30, 2009 at 21:54

Why not allow Cat 3 to support 30mtr runs of 100Base-T or 1000Base-T (although I am not sure this will work)?

Unless it is an existing installation, why would anyone want to run Cat 6 to support 10GBase-T anyway? If it is about cost then I can only suggest that we consider the true cost of cabling in comparison to the rest of the data centre and when we realise how small the amount really is, then we can start concentraing on being the best we can be and not just the cheapest. The amount of money being saved on a new installation by only using Cat 6 instead of Cat 6A is such a small amount in comparison to the total cost of the data centre and the associated risks are not worth it.

If it was my money, I would only install Cat 5e or something better than Cat 6 because I don't see Cat 6 as being useful for anything other than marketing from the cabling system vendors.

We need to, as a cabling industry, raise our profile and quality, and these discussions that try and take the industry backwards serve to do just the opposite.

Our recommendations to customer should be driven by the need for higher quality and standardisation and not take us back to the bad old days of specific cabling for specific applications.

Permalink Reply by Thorsten Punke on February 8, 2010 at 17:33

Hi John

the answer is quiet simple: It is maybe working if you follow all guidelines and instructions which are mentioned by some vendors or in standards such as TSB 155.

In the real world, all channels have to be certified and tested including ANEXT!

In other words, there is no guarantee. But the biggest issue is the missing EMI performance.

Cat.6a and especially Cat 7 and Cat 7a have a much better EMI behaviour. Latest 3rd party tests have shown in a comparative testing the importance of EMI behaviours and the sensitive of the 10 Gb/s application. In this testing all UTP systems failed completely against any global EMI requierement, even the the lowest one (E1). Such tests are required in MICE which is part of the cabling standards (ISO/IEC 11801, TIA 568 and EN 50173).

Another issue data center for Cat 6 and Cat 6A UTP systems is the short link problem. UTP systems (Cat 6 and Cat 6) need as a minimum 15m channel length.

Find more information here:

http://www.utp-vs-stp.com/web/Microsites/UTP-vs-STP/

or ask me directly.

Thorsten

Conclusion: 10 Gb/s compliance is guaranteed when the system meets:

transmission requirements + EMI requirements