New Building - Cat6 or Cat6A

P

Protoform-X

[H]ard|Gawd
Joined
Jan 30, 2002
Messages
1,203
So Cat6A is the media you want for long 10Gbit runs; however, it's much more expensive than standard Cat6. We're putting up a brand new building this year and I want to make sure I'm justified in telling the owner that Cat6A is the way to go. In my mind, preparing for a future 10Gbit network is more important at this point than saving a few thousand dollars initially. Rerunning all that cable again later would really suck. Please tell me 6A is the way to go! :confused:
 
A few thousand now or even more than that later rerunning? May as well future-proof as much as you can now. (Within reason)
 
Cat6A is larger cable, have to be more careful terminating..larger boxes, pathways etc. Keep that in mind.
 
How long are your "long runs"? Wikipedia says cat6 is good for 10g from 121 feet to 180 feet (depending on conditions) vs 330 feet for cat6a.
 
Why are you not doing fiber? No one has asked, so thought that I would.
 
It depends. Personally, I wouldn't do it, but in my environment, my users don't push over 100Mb, much less 1Gb; I see us moving out of our existing building long before anybody needs 10Gb to their desktop.
 
Fint said:
It depends. Personally, I wouldn't do it, but in my environment, my users don't push over 100Mb, much less 1Gb; I see us moving out of our existing building long before anybody needs 10Gb to their desktop.
Click to expand...

I wouldn't do it either. Maybe if you were wiring up IDFs back to MDFs (Id rather use fiber in that case).
 
I just went through this for our office. We have a new small location needing about 48 drops. After measuring out the longest runs (60ft), I went with Cat6. Will we push 10g to some desktops? Yes. Will they be within 20-30ft of the switch, yes.
 
What sort of desktops do you have that need more than 1Gb?

What switch will you be using? I can't find any high density edge switches that do 10Gb; I'm not about to deploy a Nexus in my edge wiring closet.
 
Extreme X650 and Arista 7100t are about the only switches I can think of that do 10GBase-T. Other then a couple low density line cards Cisco/Juniper is SFP+.

IMO if you want future proof run Cat6 for the desktops and OM3/OM4 for everything else.
 
corge said:
Why are you not doing fiber? No one has asked, so thought that I would.
Click to expand...

F*ck fiber i would do my hole building in Infiniband 40Gb.
Office 2000 open way faster on our Windows 98SE desktop with Infiniband. :p


Seriously more detail would be needed.

- What size of building are we talking about ? (how many floor, maximum length)
- How long do you think you will stay in that building ?
- What kind of usage ? (is there different department)
- What kind of server system do you currently have in place and plan to have eventually (upcoming 3y)

Honestly the only reason i would see that justify 10Gb cabling for the hole building is if you are thinking of staying more that 10 years or if you are a Film/Video related company that transfer HD content all day long and even there i would only do specific department. Also how your current system ? If you fasted storage barely hit 2Gb in 10 years you "might" have one that barely hit the 10-12Gb mark probably... unless you are only 2 users in the place i don't see the point of having the hole building 10Gb ready, cat6e is still 3x the price.... there not enough informations for us to make a clear and "correct" judgement of your situation.


Edit:
Once you will have those answers you should already have a pretty good answer for yourself.
No offense but if you would be a good IT you should already have a list of for/against and technical/economical justification for that "upgrade". Right now you seem (no offense again i just go with the information we have here) to just want to have the biggest thing out there without a specific reason or need.

Don't forget the 10Gb standard isn't defined yet so is it really worth it ? It's like IPv6 we are talking about it since 2002/2003 everyone is freaking out about getting out of IPv4 and most IT got their company "IPv6 ready".... and still the move to IPv6 won't happen before YEARS from now. Unless all your workstation have 10 SSD Raid0 i don't see the need of a 10Gb building.
 
Last edited:
Back of the envelope math is that cat6 is roughly 50$ drop more (if you're doing wiring properly) than cat5e.

So, do you expect to need 10G to the desktop in the forseeable future?

Look at your current peak utilization to your highest PC, do some math. How much extra bandwidth do you think you'll need?

Lets say I have a PC now that routinely uses 100mb/s. add 25% per year, which tells me that you'll be OK on gig for more than the next 20 years.

The most frequent bandwidth conversation I have with people is asking if they need to upgrade from 100Mb to Gig to the desktop, so they're 5Mb/s DSL will work better.
 
Sorry, I think a lot of you are misunderstanding the reason why we would put Cat6A in now rather than Cat6. Like 99.99999999999999% of companies, our access network is not 10Gbit, and will not be for the foreseeable future. However, just like 10Mbit became 100Mbit, and then 100Mbit became 1000Mbit, 1000Mbit will become 10Gbit some day. This is a 40,000 sq ft 2-story building that we'll be headquartered in for the long-haul, so 10, or 20+ years is not out of the question. Since no one really knows how prevalent video and other high-bandwidth technologies will become in the business place by then, it's a bit of a toss up on what to do.

I have a strong feeling(an opinion) that my desktops will all be 10Gbit in less than 10 years. The back-end will start getting 10Gbit components this year already. So the question becomes, will it cost me more in the long run to rerun all of that cable years down the road, and should I just future proof now? The kicker - What if the networking media changes from Ethernet to something else by then? Then we rerun cable anyway.
 
Then run CAT6a. If you are planning to stay there for that long, you will be rewiring at some point regardless. Cat6a is the best you can do right now to help future proof the building.
 
In our building (1.6 million sq ft) we have 300-500ft runs of cat5e......still works fine...fiber in between the hubs.....

Never hurts to future proof......we are running new drops with cat6. But we are still on a 1mb up 800k down T1 line...feelsbadman.jpg
 
Protoform-X said:
Running fiber to what? My workstations? Brilliant.
Click to expand...

Not any different than running CAT6a to your desktops, but hey whatever.

You have the information that you requested so this thread is dead.
 
Protoform-X said:
Running fiber to what? My workstations? Brilliant.
Click to expand...

You can do that.


Or, since you did not mention the building specs.......you COULD run fiber in between the nodes/hubs/switches whatever you want to call it. Then, that means better reliability and stability, and shorter cat5e runs and still get 10gb


Where I work we have over 24 nodes with 2-5 switches in each node.
 
You're getting drops to go 500 feet reliably? Do tell. Also I hope you meant switches and not hubs.
 
Protoform-X said:
Sorry, I think a lot of you are misunderstanding the reason why we would put Cat6A in now rather than Cat6. Like 99.99999999999999% of companies, our access network is not 10Gbit, and will not be for the foreseeable future. However, just like 10Mbit became 100Mbit, and then 100Mbit became 1000Mbit, 1000Mbit will become 10Gbit some day. This is a 40,000 sq ft 2-story building that we'll be headquartered in for the long-haul, so 10, or 20+ years is not out of the question. Since no one really knows how prevalent video and other high-bandwidth technologies will become in the business place by then, it's a bit of a toss up on what to do.

I have a strong feeling(an opinion) that my desktops will all be 10Gbit in less than 10 years. The back-end will start getting 10Gbit components this year already. So the question becomes, will it cost me more in the long run to rerun all of that cable years down the road, and should I just future proof now? The kicker - What if the networking media changes from Ethernet to something else by then? Then we rerun cable anyway.
Click to expand...

History can help us here. When did 10BaseT over Cat3/5 begin to take over from the earlier standards (Banyan, Thinnet, SmokinRing, etc)? Around 1995-1996 in earnest? What about the next major adoption step to 100BaseT? Call it '02-'03? Gigabit? Call it 2007? That's 12 years, driven by the need for access speed. All of these standards were driven by need, by processing power and read/write capabilites of the Client PC's. By now, most current-gen client PC's can't saturate their Gigabit uplink anyway. It's going to be a while before the need for 10GBaseT is felt and that's what will drive the market. The next possibility is that much will be wireless by that time anyway, if 802.11ac catches on. So, to sum up my input to the discussion, stick with Cat 6, save 20% on the project, and look like a hero for keeping the budget down.
 
klank said:
Not any different than running CAT6a to your desktops, but hey whatever.
Click to expand...
Other than massively prohibitive cost of fiber hardware and media. I've never seen a fiber to workstation configuration in a business... Ever. The only places I would run fiber would be to a switch panel on the far side of the warehouse if we can't make a <300 foot run with Ethernet. Although, in practice you can make runs further than 300 feet without a problem. I've done 450 foot runs of Cat6 that still pushed full GigE no problem- You can try some crazy things when your customer can't afford the right solution. Heh. :rolleyes:

You have the information that you requested so this thread is dead.
Click to expand...
Who are you to tell me when I have the information I need? I really don't appreciate your attitude or crappy advice, so please keep further insights to yourself. Thanks.

Nate7311 said:
History can help us here. When did 10BaseT over Cat3/5 begin to take over from the earlier standards (Banyan, Thinnet, SmokinRing, etc)? Around 1995-1996 in earnest? What about the next major adoption step to 100BaseT? Call it '02-'03? Gigabit? Call it 2007? That's 12 years, driven by the need for access speed. All of these standards were driven by need, by processing power and read/write capabilites of the Client PC's. By now, most current-gen client PC's can't saturate their Gigabit uplink anyway. It's going to be a while before the need for 10GBaseT is felt and that's what will drive the market. The next possibility is that much will be wireless by that time anyway, if 802.11ac catches on. So, to sum up my input to the discussion, stick with Cat 6, save 20% on the project, and look like a hero for keeping the budget down.
Click to expand...
Thanks Nate. That's a well formulated response with good advice.
 
I would personally run fiber to each closet, and stick to cat6 for the individual jack runs. If you cannot run fiber for whatever reason, then run cat6a as the "main" runs between closets. Also you should use conduits for these main runs making future changes easier.

Cat6A is mostly for inter server connectivity such as SANs and what not, even gigabit is usually overkill for each individual PC, but since it's not much more expensive may as well do it.
 
Protoform-X said:
Who are you to tell me when I have the information I need? I really don't appreciate your attitude or crappy advice, so please keep further insights to yourself. Thanks.
Click to expand...

Protoform-X said:
Sorry, I think a lot of you are misunderstanding the reason why we would put Cat6A in now rather than Cat6. Like 99.99999999999999% of companies, our access network is not 10Gbit, and will not be for the foreseeable future. However, just like 10Mbit became 100Mbit, and then 100Mbit became 1000Mbit, 1000Mbit will become 10Gbit some day. This is a 40,000 sq ft 2-story building that we'll be headquartered in for the long-haul, so 10, or 20+ years is not out of the question. Since no one really knows how prevalent video and other high-bandwidth technologies will become in the business place by then, it's a bit of a toss up on what to do.

I have a strong feeling(an opinion) that my desktops will all be 10Gbit in less than 10 years. The back-end will start getting 10Gbit components this year already. So the question becomes, will it cost me more in the long run to rerun all of that cable years down the road, and should I just future proof now? The kicker - What if the networking media changes from Ethernet to something else by then? Then we rerun cable anyway.
Click to expand...

I can tell you whatever I want, free country. You asked a question regarding if you can justify the several thousand dollar price increase, then you answered your own question in this post. So get off your high horse and make a decision already.
 
/usr/home said:
You're getting drops to go 500 feet reliably? Do tell. Also I hope you meant switches and not hubs.
Click to expand...

Cat5e starts to attenuate at 300 feet......it starts to attenuate at 300 feet. Does not mean its 100% stops sending anything you push through it at 300 feet. I took a box of 500ft cat5e, put a rj45 end on both ends.....hooked it to my desktop. No speed difference vs a 6foot pre-made jumper vs the whole box. When I say no, I mean some, but not enough to even bother.


If you would like, I can make a small video of a few 400-600ft runs to another hub.



btw, hub= centralized location of a rack/cabinet with switches and various other network devices.


hub, head end, truck, ...they are all the same.
 
JPF_ said:
Cat5e starts to attenuate at 300 feet......it starts to attenuate at 300 feet. Does not mean its 100% stops sending anything you push through it at 300 feet. I took a box of 500ft cat5e, put a rj45 end on both ends.....hooked it to my desktop. No speed difference vs a 6foot pre-made jumper vs the whole box. When I say no, I mean some, but not enough to even bother.


If you would like, I can make a small video of a few 400-600ft runs to another hub.



btw, hub= centralized location of a rack/cabinet with switches and various other network devices.


hub, head end, truck, ...they are all the same.
Click to expand...

Not saying I didn't believe you, I've never seen a run that long work well.

Have to specify hub when it comes to networking :p.
 
Like someone already mentioned, I really like the idea of fibre/cat6a between floors, then regular cat6 drops to the workstations.

If you do the math, the 120ft max for 10GE over CAT6 still gives you something like 45000 sq-ft of coverage per floor. If you centrally locate your comms cabinets and place a cabinet on each floor, I'd imagine just about every single drop would be well within the 120ft range anyway. Yeah, I realise 120ft isn't the actual distance you can run, but assume 100ft actual (10ft slack at each end), and you get around 30,000 sq-ft of coverage.

Or you could figure out which runs will be over 100ft or so, and do only those in CAT6A, or even just run them in a different colour of CAT6, so in the future it's easy to recognise which drops might have to be redone for 10GE.
 
If your runs to the IDF are long, I think Cat6A is a good investment to make.

Right now I have a 10gig access network for my 10gig server farm with fiber and DAC. For my 1gig server farm and user access, I have gigabit over cat5E. I plan to upgrade general purpose servers to 10gig Cat6A once 40gig/100gig line cards come out for my core. And users sometime after that.

corge said:
Why are you not doing fiber? No one has asked, so thought that I would.
Click to expand...
There's really no point. Why did the industry initially choose fiber and DAC (direct attached copper)? Three reasons:

1) Cat6A products have limited availability now, and not too long ago, had no availability
2) Power consumption for early Cat6A was through the roof (8-10W per port vs 1W for fiber/DAC), which is why switch manufacturers didn't waste their time with it
3) Latency advantage -- fiber/DAC adds about 100 to 200 nano ... where early Cat6A added 4 to 5 microseconds ... but has now fallen between 1 to 2 (this makes a big difference to me, but most people wouldn't care about this)

Now that Cat6A is starting to become available on switches (Arista/Cisco) and servers (Dell R620), they've dramatically improved the power consumption, and they made latency reasonable ... Cat6A will take over the 10gig access market. Finally, Cat6A PHY is much cheaper to make than precise optics for fiber. That doesn't mean fiber/DAC is going away.. I'll still use it and HPC will still use it mainly for reason #3, others will use it only for long distance runs, or for 40/100gig.

Early gigabit over Cat5 had the same problems, they fixed it, and then it became mainstream. Look for 10gig to get real traction once Cat6A becomes standard on more servers.
 
You must log in or register to reply here.
Back
Top