Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations Mike Lewis on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Backbone :bigger "pipe" 3

Status
Not open for further replies.

waynesworld58

IS-IT--Management
Nov 27, 2003
2
US
We have a cat5e going from 1 closet to another as the "tie" or backbone, but now we are adding 100 more workstations to one of the wiring closet racks.
How much volume can a cat5e handle? We use 10/100/1000 (gig) switches.
Also is there a way to add another cat5e to the backbone or would I have to step up to fiber in order to add more flow.
I understand the anology of plumbing, i.e. 5" pipe can handle more water volume than a 2" pipe,(assuming the pressure is enough to fill the pipes) but with data on copper I don't quite understand it.
 
Properly installed Cat 5e can do gig on 1000baseT.

If you use Link aggregation, you can bond 2 to 8 Cat 5e lines together to get 2 to 8 gig of throughput.

Your life will be easier if both switches are from the same vendor, ideally the same model. There is a 'new' standard, 802.3ad however many vendors used their own versions.

Electronics for 10 gig over Fiber is available, but pricey. Cat 6A copper can also do 10 gig, but is both rare and pricey.

I tried to remain child-like, all I acheived was childish.
 
Run a few extra cables, one cable is too risky---Opinion only. Like jimbopalmer said it will handle speed, but capacity to handle throughput for numerous users---?

Regards
Peter Buitenhek
ProfitDeveloper.com
 
Am I missing something here. Why not run fiber and be done with it.
 
yea , fiber is really not that expensive

if you can afford to add 100 workstations you certainly can afford fiber
 
Fiber isn't that expensive at gig speeds but he already has that with copper. As jimbopalmer noted link aggregation is the way to go. Many managed switches have this capability and it works well. On Dell switches it is called port trunking and I use it all the time.
 
op says he has "a cat5e"

how much to pull more copper vs fiber ?

Id still do the fiber rather than a "work around
 
I would also opt for fiber. The price difference is minimal (often cheaper) and it's simply a better.
 
If you have unlimited money who wouldn't "opt for fiber" and 10G uplinks between the switches.

He needs to add 100 computers. Switches capable of 10G uplinks start at around $75 per port vs $14 per port for 1G uplinks. Add the 10G uplink modules at around $800 each. The cost of networking hardware comes in at around $9.1K for 10G fiber uplinks vs $1.4K for a copper link aggregation solution. Add to this the cost of the fiber which may be substantial depending on contractor.

The primary consideration is network configuration. What workstations connect to what services via what path. If all this uplink between closets does is share a T1 internet connection then what is gained by 10G?

If there are servers involved then where are they and who uses them? If the servers for each "closet group" are not across the uplink there is no gain in a 10G connection. If there are only 2 servers each with a 1G port then a 2G link aggregation uplink will work well.

10G uplinks are the current "cutting edge" technology and as such are expensive. Link aggregation is not a "work around" but is the way bandwidth issues are addressed now and in the future. You can bet that we will be grouping 10G ports together as bandwidth requirements continue to expand.

There is currently no difference in the maximum speed for ethernet over copper or fiber. Both can run at 10G. The practical difference is how far the signal can travel. In common usage 10Gbase-SR goes 269' on MM fiber while 10Gbase-CX4 goes 49' on copper. 10Gbase-LR which will go 6 to 15 miles on SM fiber is available but even more expensive. New standards are being developed all the time as the market has not stabilized on a favorite. All this reminds me of the older 100 and 1000 megabit technologies that now litter the landfill. My bet would be on 10Gbase-SR, 10Gbase-LR or the emerging 10Gbase-T which should go 328' on cat6a.

Simply saying fiber is the solution does not fix an undefined problem.


 
Also is there a way to add another cat5e to the backbone or would I have to step up to fiber in order to add more flow.

I thought we were talking about one backbone connection, rather than one per machine.
 
we are talking backbone , connection from IDF to
MDF

choices are fiber or pulling additional copper

where it my job I would recommend Fiber .

(well actually I would know all the particulars and would cost it out , but based on what we have been presented i would recommend fiber )

 
We are talking about the backbone connection (uplink) between two wiring closets each with "10/100/1000 (gig) switches".

The issue is how much bandwidth is required across this backbone.

The OP (waynesworld58) wanted to know "is there a way to add another cat5e to the backbone". Jimbopalmer answered correctly by pointing out link aggregation as a technology that would do exactly what he wanted. Buitenhek added a good second point as to the failover nature of link aggregation which is a very good reason to employ this technology on important links.

Since that point things have gotten muddy (as they tend to do around here) with the usual chorus of "text message" type responses with "knee jerk" solutions. These experts tout running fiber as a solution to a problem that is best solved by the first response regardless of media type (copper or fiber).




 
with the usual chorus of "text message" type responses with "knee jerk" solutions. These experts tout running fiber as a solution to a problem that is best solved by the first response regardless of media type (copper or fiber).

[nosmiley]
 
maybe we need a moderator here to delete all but the "correct answer " and then lock the thread from those they consider posting knee jerk solutions ?

any volunteers ?
 
Yeah skip555, I'm unhappy also.

Time was that this forum was a place to exchange ideas and explore technology. Lately it has gotten to where fixed responses are given to many problems.

"you shouldnt be crimping anything"
"don't we all ?"
"contact BICSI and hire a RCCCD"
"id still do the fiber rather than a "work around ""

These sort of responses don't add to any discussion. In this case you particularly dismissed the most appropriate solution as a "work around" out either ignorance or pride.

I did appreciate your appropriate response to "Extending Cat 5 Cables" but I feel as though you get a lot of mileage around here for less than thoughtful comments.

A little self moderation would go a long way...
 
(I like fiber, most of my outside plant is connected via fiber, for runs like 17,000 feet, I would not consider anything but SM fiber, I am not bashing fiber)

The OP asked a question, I answered it as best I could, several others have added additional useful info.

I use a LOT of link aggregation over a Lot of fiber. Why?

1) In 1992 when I laid down my fiber plant, the only 100 meg network topology was FDDI, over fiber. 100 meg ethernet was not delivered yet. (1995)

2) In 1998 when I went to gig ethernet, all of the working gig switches were fiber.

3) My LAN is about 4 miles long, so some runs have to be fiber or wireless. (I use both)

4) When my bosses decide I need to do 10gig, I bet fiber gear is more robust and available than copper. Each new speed has come first in fiber. (much of my current fiber plant will be too old, it is 62.5 nm MM, and needs to be 50 nm, some distance will need to be SM)

5) All of the other 1999 employees at my site think their job is to move dirt, so my goal is to use routes and technologies that allow a link to be cut without my users knowing.

6) My network is also our VoIP phone system so it needs to be up more than a pure data network would.

But these reasons will not be universal: some of you are not in 55 separate buildings; some of you have to wait until a technology is cheap, rather than when you boss has a whim; some of you are not working in an open pit mine and can expect your cable to remain whole; some of you do not have 1000 nodes on your network; some do not have to be the 911 service for a dangerous profession.

If the OP has a network that is 100 meters end to end and is happy with copper, I want to keep him happy. I use a lot of fiber, but am not a fiber bigot.



I tried to remain child-like, all I acheived was childish.
 
you know a lot of really knowledgeable people have quit posting here ....

enjoy
 
you know a lot of really knowledgeable people have quit posting here ...."

Discussion of this topic is best continued in thread575-1355380
 
Thanks everybody for all your help & input.
I need to price out both fiber and Link aggreation and then let the "higher up" people decide.
Although we have alot of workstations and servers, the network mostly is used for office apps, with pdf being the biggest load. No voip in the future.
The layout of the building is the IT server/ t1 room is in the center (of building)while the northern end already is done with only 1 cat5 which so far has been okay, while the southern end is the newest area that I'm presently asking about.
Presently 2 of the 3 switches are managed and capable of
link aggregation, the 3rd would have to be upgraded.

Once again thanks and I'll post a follow up after they decide.
 
forget the "no voip in the future"

you never know how long it takes but that will be the new standard for telephony


______________
Women and cats can do as they please and men and dogs should relax and get used to the idea!
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top