Interesting White Paper - Cat5e is good for Gig, 10/100 may suffer...

[ScottMac]

Here's a link to a white paper done by the Anixter Lab.

Anixter Structured Cabling Lab, Technical Library


The short version: Chacteristics that make a cable work well for Gig E may cause more errors on 10/100 network systems. Choose your components carefully to ensure a good match.

FWIW

Scott

 

[Mark R]

I don't think that's what it says. If you read it in detail, it doesn't say that Cat5e is worse than cat5 - rather all it says, is that cat5e doesn't verify several parameters which are important for gigabit, and less important for 10 and 100 Mbps.

In short, the whole paper seems to say: 'Just because it's cat5 or cat5e doesn't mean it will work properly. Please buy our super-expensive network cable kits, which are tested much more rigorously.'

 

[ScottMac]

I'm not seeing anything in the article that alludes to either of your points.

Can you reference them specifically?

Thanks!

Scott

 

[spidey07]

Thanks for the heads up Scott.

That paper stresses the age old fact of "Use a matched cable system". I too have seen cat5e instalations not perform well because the installation used BlackBox panels, with billy bobs outlets and joes discount cat5e horizontal cable.

Time for another spideyism - "don't muck with the physical layer"

 

[Mark R]

Nowhere does it say that Cat5e is inferior to Cat5. In fact it actually says:

<FONT face=Arial size=2>Category 5e is today?s Category 5 with additional testing of various parameters added. If Category 5e reflects the performance of today?s Category 5 components, then all components must be installed </FONT><B><I><FONT face=Arial size=2>perfectly </B></I></FONT><FONT face=Arial size=2>to support 10BASE-T, 100BASE-T and 1000BASE-T. There is simply not enough additional performance headroom in the standard to assume that total support of 10BASE-T, 100BASE-T or 1000BASE-T will be guaranteed.

Which suggests that neither Cat5 nor Cat5e are actually adequate specifications. It goes on to list the enhancements Cat5e offers over Cat5.

The main thrust of the paper is that it is impedence matching between all components, the patch cables, the connectors, and the horizontal (backbone) cables, that is critical, crucially it is not the quality of the individual components, but that they are accuracy of the match between the components.

<FONT face=Arial size=2>As shown above, a matched ALC 7 system produces few or no frame errors, ALC 6 produces slightly more and Cat 5e even more, Cat 5 produces the greatest number of frame errors.

The results of their tests, above, don't support the claim that Cat5e may be inferior in a conventional ethernet environment (in this case 100-BaseT).

My claim that this is essentially an advertisement for their products is perhaps somewhat facetious, but they make numerous references to how their Level6 or Level7 products are superior, and give recommendations for their installation in place of Cat5 or Cat5e systems. e.g.

<FONT face=Arial size=2>If the user is not expecting to recable within the next two years, or is not expected to move, a Level 7-based cabling system is the best choice. This cabling system will provide the platform needed to reliably support 1000BASE-T in the future.</FONT>

.
I can't help being skeptical of a report which extols the virtues of the same companies products, even if the claims are well founded.</FONT></FONT>

 

[ScottMac]

I don't believe the intent was to try to discredit Cate5 / 5e cabling. I took it as a caution that all structured cabling components, even with a Category rating, are not equal. Further, I believe their intent was to show that (kinda like software) an enhancement in one area may have an adverse affect in another area (in this case, the "enhancement" to Cat 5 to make it suitable for GigE -may- cause some degradation to a 10/100 system, especially if the components are not matched).

I know the folks that did the tests, I know they do a good job and do their best to produce fair & valid results.

The other thing to keep in mind relating to not trusting a company that "extols the virtues of the same companies products, even if the claims are well founded." Anixter doesn't produce cabling products. They sell 'em. The "product" here IS the "Levels" specification, no doubt about it. They order their product to "Levels" specification, and monitor the stock to see that it's compliant. The promotion of the "levels" program in the paper is, if anything, some justification for the higher price that Anixter may ask for a given manufacturer's product (because they see to it that the product performs at or above the spec).

Aside from that, there is other useful information in the paper. I believe it goes toward disproving the wide-spread belief that "wire is wire" (or by extension, "as long as everything is rated the same, it'll be OK"). There is at least some demonstration of WHY it's a good thing to match your components.

I can appreciate your skepticism, personally, I operate on the "An optimist is never pleasently surprised" side of the line; your grains of salt are noted. Of course there is some promotion in the paper, but there is a lot of good information in there too. My recollection was that Spidey presented some issues relating to Cat 5e / Cat6 months ago, and this paper sounded like it suported / answered some of the issues he brought up.

I know that the average user couldn't possibly care any less about the cabling details, but there are some corporate networking types that hang here too ... and this is good information from a credible source (with a little promotion thrown in, big deal).

Anyway, that's my story, I'm sticking to it.....Thanks for the response.

Scott