Network Upgrades: CAT7 and IPv6

I received a message from a follower asking me about my current network configuration, both in reference to a post I made a few years ago and the fact he too lives in the St. Louis area and uses Charter.  Funny you should ask….

I currently pay for Charter’s highest tier at 100/5 (the fastest in our area… /praying for Google Fiber to come to the St. Louis area).  I’ve never really achieved the speeds I pay for with my hardware and never felt the need, but as my hardware was nearing 5 years old (it was future-proofed at the time) and my family’s demands on the hardware are more reliant and strenuous, I felt it was time to future-proof once again.

First, my old configuration:

And for my newly purchased upgrades:

The purpose behind this whole project was to future-proof with reasonable expectation of some slight speed increase over the network… nothing spectacular and not terribly expensive.

New Network

I received my new Asus router first and had the chance to set it up and tinker before I received any of the new cabling.  This is a good thing as there were significant speed increases once I installed the cable.  And after testing in a real world scenario with CAT5e, CAT6 and CAT7, I can vouch that CAT7 is some awesome stuff.  Yes, your cables matter and they do make a difference.  And anyone who says CAT7 cables won’t make a differenc has never used them.  My internal network speeds are dramatically better from CAT5e to CAT7.  File transfers from computer-to-computer are blazing fast!

Network Test

My CAT5e Maximum speeds were in the 125s over both the Asus gigabit router AND the my older Linksys router (love the sales pitch on “tested 1Gbps” CAT5e cables… MY ASS!)… and yes I tested on multiple CAT5e cables.  The differences in speed [on paper] were marginal between CAT5e and CAT6, getting about 200Mbps out of the CAT6 cables; I certainly couldn’t feel the difference between the two.  The Linksys router kinda crapped out at speeds above 200Mbps anyhow.  So yay for CAT7 baby (results above)!  So what exactly is CAT7? Well:

  • CAT5 = 100MHz
  • CAT5e = 100MHz (it may claim it’s 300MHz or 350MHz… it’s not!)
  • CAT6 = 250MHz
  • CAT6a = 250MHz
  • CAT7 = 600MHz

Translate, noise is bad.  And the CAT6, CAT7 cables offer more shielding to prevent noise, hence faster communication and less errors.  The primary advantage of CAT7 is supposed to be maintaining speeds over longer distances… ie, if you’re gonna run a 100-foot line of cable, CAT7 is probably the way to go in order to prevent transfer speeds from diminishing.  For the geeky details on category cables, this link breaks it down really well.  I primarily went with CAT7 because it seems like it may be the last generation of category cable to be backward compatible over copper.  Whether that’s true or not remains to be seen with CAT8 on the horizon.  I’m more interested in the hands on feel as opposed to the “on paper stats” anyway, and CAT7 is zippy as can be.

Don’t forget, that hardware plays a HUGE role in network transfer speeds as well.  Albeit a bit dated (2009; pre-SSD era) Tom’s hardware has a great article that discusses the hardware aspect very well, although I didn’t even come close to achieving the results the article did.  In my particular case, I’m testing transfers from my Windows 7 workstation (12GB RAM, solid state drive) to my Windows Home Server (8GB RAM, multiple 7200 RPM WD Black Drives in a pool) and my Windows 2012 development server (8GB RAM, 10000 RPM raptor Drive).  The CAT7 cables these tests were run on were all 15-feet in length.

After configuring my router (BEFORE installing the cable), I ran some tests and was only getting slightly better speeds [on the web] than my old router (about 80Mb/s vs the roughly 70Mb/s I was getting before)… exactly what I expected.  However, once I enabled and configured IPv6, some speeds over the internet seemed to dramatically improve, like even my wife and teenagers noticed.  Google services in general are noticeably faster as well as Wikipedia and FaceBook (a simple ping for these sites from a command prompt will return IPv6 addresses).  Most exciting, I can now pay my Verizon wireless bill over IPv6 (yes, Verizon’s compliant too).  The other IP services only pickup my IPv4 address, and don’t really seem any faster than they were before, as is the vast majority of the web that has still not adopted IPv6.  So I’m assuming most of the “zippiness” or “zest” I’m feeling are from sites who’ve already adopted IPv6.  After running speedtest however, although my speeds were reliably faster, As I stated before I still wasn’t achieving the 100mb speeds I was paying for with Charter.  Why?  Turns out I’m a victim of “peak hours”.  How do I fix this?  My modem.  8 bonds instead of 4 to be precise… which Charter just pushed on supported modems on August 29th!  So although I didn’t plan on upgrading my modem, I went ahead and upgraded to the SB6141 which has 8 bonds as opposed to the 4 bonds in the SB6120.  And what was the result?  Although I had to wait 24 hours for Charter to push the 8-bond firmware update to my modem… low and behold I have the speed I’m paying for! (insert angelic sound and bright white lights)

Now I’ve never had trouble with bandwidth before, but naturally I now want to stress test everything and flex my muscles a bit.  So I put the whole family on our devices all at once.  With two simultaneous netflix streams, four simultaneous network HD video streams off the Windows Home Server, and downloading two 1GB files simultaneously… neither the network or the internet even hiccuped… and speeds were fast and zippy the whole time.  This is the hands on result I was looking for.

I was impressed, so lets play with IPv6 some more!

World of Warcraft has an enable IPv6 setting.  Yep… lower ping and I’m no longer seeing detailed scenery load… it’s loaded before I get there.  Another nice unexpected perk.  Sweet!

My HP CP2025DN Printer has detected an IPv6 address in its control panel.  Although Windows didn’t pick it up automatically, I was able to copy the IPv6 address and add the printer manually.  I honestly have no idea if there’s any speed increase with the printer and don’t even care, but it’s cool so I set it up on all the computers as IPv6. :)

As for wireless, I didn’t even bother to test it.  I don’t have any AC devices yet which makes that feature of the router and the network pointless at the moment.  And simply put, tasks on my wireless devices don’t really consider speed to be critical.  The phones, tablets and laptops simply need a solid connection anywhere in the house for basic surfing, streaming video/music and maybe some PDF downloads.  The heavy network tasks are performed by the machines that work… the servers (I have two) and the workstations (again, two on the network).

I don’t profess to be an expert at this and I’d certainly like to tinker some more to see if I can squeeze out faster speeds, which I think lies in tweaking the NICs on my network as opposed to any other hardware.  And I’ll admit I got a bit carried away, but really enjoyed this two-week project… the quest for more speed is addicting!  Not bad for a home network in my humble opinion.  Remember those archaic things called wires and desktop computers… you know… used to do actual work?  Yeah… they’re still VERY relevant.  Granted… this network’s setup for both work [at home] and the new toys.  Here’s to the five years before the next upgrade.

9 comments

  1. James says:

    So how did you ground the end of the cat7 cables? And on which side?

    • Rick says:

      The cables were all premade lengths – ie. I didn’t buy a reel of cable, cut to size and fit my own connectors. I let the manufacturer do the grounding for me. :) In this case, I purchased Rosewill CAT7 cables from Newegg.

      • James says:

        I understand that. But what did you ground the end of the cable to? The end connector (the metal part) needs to be grounded, otherwise it’s not providing a shield in any sense of the word.

        • Rick says:

          To be honest I’ve never heard of “grounding” an end connector and had to look it up. In terms of “not providing a shield in any sense of the word”, everything I can find online disagrees with that statement. This led me to research STP vs. UTP cables, and from the research I’ve come up with, I can only find that you “MAY” run into problems if your cable isn’t grounded. The only people who seem to have run into problems with CAT7 (since CAT7 doesn’t come in UTP) are those who cut their own cables and made their own connectors. I have not had any problems [that I’m aware of] in the 13+ months I’ve been using the CAT7 cables I purchased and my speeds are the same today as they were a year ago. CAT7 cable reviews on Newegg echo and confirm the same results. However, if I can squeeze any more speed out of my system by doing something as simple as grounding the cables, I’ll try and come up with something. I do intend to look into this further.

    • Rick says:

      Alright you had me stumped, but after several hours I think I have an answer.

      In my post I linked the Rosewill cables I purchased. In that description it states “In compliance with TIA/EIA-568-B.2-10 specifications”. Google “TIA/EIA-568-B.2-10” and you’ll find a lengthy PDF of the spec (note the Wiki entry underneath for dates/revisions of the spec). Search the PDF for “ground” and you’ll find the spec addresses grounding with: “The system must be bonded and grounded as per ANSI/TIA/EIA 606.”

      Google “ANSI/TIA/EIA 606” and you’ll find another lengthy PDF, which I read through, and after studying all too many Google images to understand the PDF, I’ve determined that the grounding is in the cable.

      The articles I kept reading that discussed manual grounding of CAT7 cables are old articles… like years old . From what I can see, as of 2014 [revision C] of the standard, if the cable is TIA/EIA-568-B.2-10 compliant, it’s grounded. No extra grounding required.

      Hope that helps.

  2. Sam says:

    I really like this article and the fact that you do a real world A/B test with the only change being the network cables. Nicely written as well. I have a similar setup except I have Category 6 patch cables and I see write speeds that are cut to around 100Mbps due to the server’s hard disk write speeds. But my read speeds are 800Mbps.

    How can you get such a high write speeds given that your test used traditional hard drives on the server side? Those speeds are in solid state hard drive territory. I tested a 100ft run of Category 6 cable with two laptops both with SSD’s, one Windows 7 the other Max OS X. I got 700Mbps+ up and 900Mbps+ down. I also tested two windows laptops with SSD and the results were not as good but still 500Mbps+ up and 940Mbps down. That shows that Category 6 can have much better results what what you got. Category 6 is capable of gigabit speeds.

    • Rick says:

      The server’s OS drive is an SSD, and I think that makes the difference for the communication, although the drives in the pool are still disks. I’ve long upgraded my hardware and software since this article and still getting the same speeds which is nice.

      I just recently ran a whole office in CAT6 and the speeds are sufficiently within the standards and speeds outlined in my post, but nowhere near achieving the speeds of my CAT7 network (regardless of OS) which is just blazing fast. So I can’t say I’ve shared the same real world results.

      Glad to see I’m not the only one still trying to push network limits. :) Thanks for sharing.

  3. Richard Tenamore says:

    I have 45 years experience as an Electrical Circuit Design and Systems Engineer. I have designed computer and communications systems. All shielded circuits and systems need to have their shield grounds and system power grounds tied together at one point within each system. Shielded cables need to have their shields connected to ground at one end only. If both ends get grounded; this will create ground loops between multiple systems. This will produce a great deal of system ground noise. The cable(s) shields should be connected at the end of the Cable; that serves as the source for the signals in the cable. Ideally, the shield for each twisted pair should be treated separately. The shield, for each pair, should be grounded at the end of the cable where the signal source originates. It appears, to me; that pre assembled CAT7 cable configurations do not allow for the proper grounding scheme to be created. The thinking that pre-made cables already take care of all grounding concerns, is totally irresponsible, B.S. Take a look into the Ethernet sockets on your equipment. Do you see any special grounding fingers inside of the sockets that serve to connect with cable shields? Probably not!!! These will leave you with a totally ungrounded shield(s) on your CAT cables and this configuration actually generates more noise on your signal lines; than a cable without any shields at all. If you do have grounding fingers on the inside of your Ethernet socket; you loose again. Now you will have both ends of your cables grounded and creating a great deal of ground loop currents running between different pieces of equipment in your installation.

    • Rick says:

      Cool… thanks for your input Richard. Do you have a good article (or better yet a diagram) that demonstrates proper grounding? Maybe a how-to article if you will? Thanks!

Leave a Reply

Your email address will not be published. Required fields are marked *