Speeds of Communication: Fibre versus Cable.
-
Your episode on Network Cables and Connectors compared the speed on communications between fibre and copper cable at time 39:30. Don and Ronnie on the statement:
'In terms of the sheer signal getting from point A to point B its going to happen faster over fibre than it is over copper'.
This is factually incorrect. It is universally acknowledged that the speed of light in a material such as fibre-optic glass is around 200,000,000 meters per second, where as the speed of an electrical current through copper wire is around 300,000,000 meters per second.
How will you correct this?
-
Murray,
Thanks for writing in. I'm really glad you brought this up because we didn't address it in the show. This is the age-old battle of the science of physics versus network engineering. Your statement is 100% correct in terms of physics. In physics speed is expressed in distance over time (miles per hour). In networking, however, speed is typically expressed in terms of bandwidth over time (megabits per second). Fiber optic connections support a higher megahertz signal frequency which means it can carry a higher bandwidth and move more data in a shorter amount of time than copper, thus making it "faster" than a copper based network.
The exception to all of this is measuring the speed in which someone can get fired from a Network Admin position which is normally measured in quantity of help desk tickets over time.
Thanks again,
Don -
I want to piggyback on this question if you don't mind. I have heard you can saturate a fiber link more thoroughly than a copper link in terms of percentage of theoretical capacity? Is there any truth to that? Does that even make sense? Thanks.
-
Paul,
Thanks for writing in!I have heard you can saturate a fiber link more thoroughly than a copper link in terms of percentage of theoretical capacity? Is there any truth to that?
When using fiber optic cabling, you can saturate the receiver of the light signals when you do not have a sufficiently long enough cable to provide for proper attenuation of the signal. This is called "overdrive." You have two options if this happens: One, install a long enough cable so that the proper attenuation is there or two change the receiver of the signal to something that can handle more signals simultaneously.
I'm not sure what the overall capacity of the fiber modes (single or multimode) are on the cable. I can tell you that fiber optic standards are open ended because they haven't found a saturation point yet. Even so, the result, whether the cable can become saturated or the receiver can be saturated, is that we will experience saturation.
Cordially,
Ronnie Wong
ITProTV -
Murray,
Thanks for asking the question: which is faster...copper or fiber optics?
For the idea of one packet vs another packet. I'm not sure if fiber optics is faster vs copper according to your figures, it couldn't ever possible be faster. But...most of the time, when referring to the "speed" on cabling, we're referring to bandwidth or throughput on the cabling. And throughput of data is normally determined by the ranges of frequencies that the cable can handle. Fiber Optic cable can transmit MORE at one time gets more data from point A to B in the same amount of time.
The analogy that I used to help a non-IT friend understand is using the idea of moving a friend out of an apartment. Nate, has a GXR and has offered to help a friend move. I've seen the man fly about 100+ mph on that bike. Yet when he arrives to help to move, He can only hold what that GXR can carry to that the new apartment. He'll take the the item faster but he cannot carry as much. If we go to my friend's apartment with a box truck. It might take time to fill the box truck but when we are done, we can drive all the apartment stuff once instead of expecting Nate to ride back and forth 5000 times carrying one item at a time.
Cordially,
Ronnie Wong
ITProTV -
Thanks Don and Ronnie for clarifying, however when I hear the phrase:
'In terms of the sheer signal getting from point A to point B its going to happen faster over fibre than it is over copper'.
I take this to mean the speed of an arbitrary electrical signal or light beam. I have listened to the segment of the show repeatedly and could not conclude that you were talking about the speed of effective packet transmission as opposed to what I would describe as ' sheer signal getting from point A to point B'.
I have always wondered; from an IT perspective are there any drawbacks to the slower propagation speed within fibre? Are there any applications where one would really prefer to use copper thanks to the faster individual signal speed alone? -
Hello Murray,
Thanks for the question again.from an IT perspective are there any drawbacks to the slower propagation speed within fibre? Are there any applications where one would really prefer to use copper thanks to the faster individual signal speed alone?
I'm not aware of any application or equipment that we choose in IT based on velocity of signal itself along the medium. I don't know of a particular drawback to the slowness that you're referring in data transmission in fiber optics. I've not seen in a datacenter, where if given the option to choose one over the other, they would choose copper over fiber optics just because of the signal speed alone. Now, that being said, people may choose copper over fiber do to other factors (i.e. costs, equipment on hand, current existing infrastructure rather than running new cables...) Let's see if Don chimes in on this one, he may know of one!
Cordially,
Ronnie Wong
ITProTV -
Murray,
Not that I am aware of. Storage Area Networks (SANs) are the most data intensive networks out there and they are almost always fiber based. I think when you factor in the latency introduced by our network cards, HBAs and other network devices, the latency between copper and fiber becomes negligible. I'm not an electrical engineer, though, so that is an unscientific statement and I would gladly defer to someone with more expertise on the matter.
Don -
I read an interesting article a while back about someone paying alot of money to get a shorter fiber link across the Atlantic to Great Britain. It was something to do with financials and the speed of trading and because it shaves down a few ms off the shorter fiber path it gives them an advantage to trade faster. In their case it was worth millions of dollars to shave a few ms so maybe there is a case to be made for copper in the financial tradings industry?
-
I wanted to clarify what I meant by saturating a link even though your description was fabulous and I learned something new and interesting. What I meant was say you have a gig-e link theoretical 1gbps. With the fiber you can come closer to hitting that max than you can in an identical case with copper. So say you could hit 900mbps on fiber and 800mbps on copper. Just throwing out random numbers to demonsrate but is there any truth to that from what you have heard?
-
Paul,
I'm now tracking what you're saying. If the 1Gbit was a theoretical limitation, the achievable saturation of fiber optics would be limited to light interference and heat. Put differently, external sources of light, cable imperfections will affect this limitation and the "melting point" of the fiber cabling itself. In a similar manner, copper cabling is limited by EMI and heat as well. So can we achieve better saturation in fiber optics? We can send more signal at one time over fiber optics, we have more saturation (i.e closer to the max).Cordially,
Ronnie Wong
ITProTV