Why Run FTTH When You Can Pretend You Do? > 100Mb is 100Mb

100Mb is 100Mb

I think that fiber speeds are more sustainable than copper because most cable providers have fiber to the node system where multiple users are using one strand of fiber for service so it flucuates depending the demand of the system. but I do have to say that docsis speeds aren't bad they just have problems keeping a constant speed.

reply to battleop

said by battleop:

This is stupid. Both DOCSIS and and FIOS are capable of doing 100Mbps+. I would love to hear a reasonable explanation of how 100Mb fiber is any faster than 100Mb copper or how 100Mb copper is faster than 100mb fiber.

Fanboys this is your chance.

reply to battleop
GPON, which is the primary technology used in modern FiOS deployments, has a net total of 2.4 Gbps (2457.6 Mbps) on the downstream that can be distributed over 32 or 64 homes.

Each downstream channel bonded in a DOCSIS 3.0 yields approximately 38mbps. To match GPON's total bandwidth they would need I believe about 64 channels, which would basically require dumping most TV channels. I've seen, however, a claimed maximum of 3 channels that can be bonded at once (not sure on the source) in current implementations, which would give a maximum systemwide bandwidth of 114mbps. Cable systems are usually shared over a larger number of people than GPON systems (the number varies), but cable companies have been doing a lot to upgrade their networks and lower the number of subscribers on a node, at least on the ridiculously huge ones.

Where this becomes relevant is oversubscription. Both networks can provide a theoretical 100mbps connection, yes, but the GPON connection would have a much lower oversubscription rate assuming more than one person is buying your 100mbps connection. This would make the GPON connection more stable and so more likely to actually reach the 100mbps speeds.

reply to battleop

said by battleop:

This is stupid. Both DOCSIS and and FIOS are capable of doing 100Mbps+. I would love to hear a reasonable explanation of how 100Mb fiber is any faster than 100Mb copper or how 100Mb copper is faster than 100mb fiber.

Fanboys this is your chance.

reply to nasadude
A relatively minor correction- FTTH, at least in the BPON and GPON forms most commonly used in the United States, isn't symmetric- GPON provides 2.4 Gbps downstream but only 1.2 Gbps on the upstream, BPON provides 655 Mbps down and a 155 Mbps up. (These numbers are for an entire node)

Now, this is far more upstream capacity than most implementations of DOCSIS currently provide, which allows Verizon to offer symmetrical packages that cable can't match, but in the end the overall technology is still asymmetric.

reply to battleop
Faster is NOT the issue.
It's latency and QOS that is inherently better with fiber that makes Fiber by far the better choice. In addition fiber is expandable to provide up to 10 Gig speed, or anything in between...

Can Cable do that ?

And BTW, how many systems ARE on Docsis 3.0 to support this ?

reply to battleop

said by battleop:

This is stupid. Both DOCSIS and and FIOS are capable of doing 100Mbps+. I would love to hear a reasonable explanation of how 100Mb fiber is any faster than 100Mb copper or how 100Mb copper is faster than 100mb fiber.

Fanboys this is your chance.

reply to nasadude

reply to raptor1418
"I can only go by what technologies are being looked into now"

How do you know just what technologies the cable companies are looking into? Do you really think they are going to max out at 100Mbps and give up? I doubt it. They will adapt as the market dictates.

*Disclaimer* I do not work for a cable company nor do I subscribe to service from a cable comapny.

I am not saying they are going to stop at trying to max out at 100Mbps. I am pointing out that your original comment seemed pretty short sighted. Plus cable companies and Telco's are not the only people that develop network technologies to improve bandwidth, latency, etc... Look at cable makers as they design better insulation or chip manufactures or network gear vendors. Most of those type of companies publish white papers on theoretical designs that sometimes become reality.

reply to battleop
Uhm, you can't break the laws of physics, doesn't matter what the 'market wants'. Spectral efficiany can only go so high before cable companies have to upgrade networks to full FTTH. Reason? GPON delivers 2.4 Gbps/sec down and 1.2 Gbps up, to 32/64 homes. On top of that, it supports a 1550 nm RF overlay, with enough bandwidth to deliver hundreds of HD and SD channels. Even with 1 Ghz systems underway, cable will have to reallocate more channels for DOCSIS 3.0 to work, and use SDV. Even then, DOCSIS 3.0 delivers 343 Mbit down with 8 channels allocated. So, to reach 2.5 Gb/sec downstream, that's approx. a bit more than 56 channels, maybe 64 to get a bit more bandwidth than FTTH. On top of that, the MSO's STILL have to keep increasing HD channels to compete with FTTH, not to mention many are starting to b***h about cable's craptastic HD quality.
I'm all for some kind of miraculous modulation technique to increase cable's RF efficiency, but as of today, I'd say that DOCSIS 3.0 is pretty much pushing the general limits of Shannon's law and other predictions used in telecom.

From one Mathematics major to another saying that DOCSIS is pushing Shannon's law is absolutely, positively, 100% false.

»en.wikipedia.org/wiki/Shannon-Ha···_theorem

Assuming we are talking about a clean moderate sized wire (RG6), the absolute maximum assuming a reasonable distance, is very high. Don't believe me, plug in the numbers.

(assuming of course by that fundamental theorem you have on display, you are/where might be a mathematics major.

Your confusing total "node" bandwidth with cable medium bandwidth. DOCSIS is a node limitation, which indirectly limits cable.

DOCSIS 3.0 is no where near the limits of cable. Koreans who use cable internet have been getting 40-100-1,000 mbit lines for about 3 years now. Also, these are per subscriber speeds. Also, i should mention that Japan has very few cable lines, but 100 meg per subscriber was also deployed there using a technology completely different than DOCSIS.

DOCSIS 3.0 is very "narrow" , sacrificing huge amounts of speed for compatibility.

Spent a long time running the numbers, so check it out:

Assuming 40 dB SNR, the raw power/noise is 10000 (L sub dB= log10(P1/P2). Plugging into original theorem: C= (6*10^6)*log2(10001), or (6*10^6)*(13.3), which is approx= 80 Mbits/sec on a 6 Mhz channel. Sure, you're right they've not YET filled up all of the theoretical capacity, but it's pretty close. with 4-channel bonding, it only yields about 320 Mbits/sec, no even counting that SOME of that must also be allocated upstream. So, assuming 100% spectral efficiency, to match FTTH's 2.4 Gbit/sec downstream, cableco's need (2400/320)*4, or 30 channels. Ouch....
(And take into acct. that this assumes some miraculous new tech that actually can USE this theoretical capacity)

BTW Koreans only get "up to" 100 Mbit access , at least over cable (and they're also bound to pre-cert DOCSIS 3.0 gear). And the majority of Japanese urban centers get FTTH, while the rest get VDSL/ADSL2+/Cable-D3.
And no, I'm not a math major(although that's my target), I'm a sophomore in HS who happens to know calculus and other advanced topics.

Peace,
a333

Korean companies used Pre cert DOCSIS equipment in 2006. Most faster carriers use TDM (time division multiplexing) between 800-1,000 mhz range to supply dedicated per user 100 mbit/s lines. These are DEDICATED, there are no "nodes" on TDM systems.

This is also deployed right here in the states, by NARAD on behalf of cablevision.

from there website

"Narad's overlay technology lives on the existing HFC plant, but taps into spectrum above 860 MHz and below 1.1 GHz. Although Cablevision is starting with a 50 Mbps dedicated data service, the Narad platform the MSO is deploying is capable of raising that figure to 100 Mbps. Narad said its architecture roadmap is built to eventually support symmetrical 1 Gbps and 10 Gbps services"

I should mention that these speeds are symmetrical over RG6 cable using a tiny fraction of the available "spectrum". Assuming one could use a dedicated line we are, in theory, well above the 100 gigabit range.

»Cablevision's Advanced Network Delivers 100 Megabit Dedicated
A trial of 100 mbit/s DEDICATED (per user) from calbevision.

Another note: ethernet over coax is common in commercial applications, standards come in 1,2,5 and most recently 10 gigabit applications. As far as i can tell none of these use frequencies above 625 mhz.

I also suggest you research ultrawideband. Tests conducted have already passed the 100 gigabit mark over coaxial.

a light read:

»www.wipo.int/pctdb/en/wo.jsp?IA=···LAY=DESC

Also, It should be noted that your still not using Shannon's law correctly. For one, different frequencies have different potential, which vary with power and length and temperature. I would suggest not even quoting that law.I have chosen to disprove your argument by example, because its easy, and also to avoid an overly complicated argument which by your last post suggests is hopeless.

Yes, I know about ultrawideband and NARAD... have been reading up on those for years now. Fact is, UWB is only in the labs as of now, and having committed themselves to DOCSIS 3.0, fat chance that cable providers will consider anything else. Deviating from the current upgrade scheme would entail a major overhaul of established node structures and headend equipment, by the end of which FTTH providers would manage to have at least half of their footprint covered by full fiber. And don't forget, while cable struggles to achieve possibly-gigabit speeds in the near future, Verizon can simply upgrade their CO equipment to WDM-PON, or 10GEPON in the next few years, pushing 10-gig speeds to new subscribers, while at the same time coexisting with older GPON subs. Finally, cable's node sizes and oversubscription ratios remain a huge problem, considering the FTTH 32/64 users per splitter, compared to cable's average node size of about 150-300 users. Sure, I'll concede that there exist technologies like UWB to push the limits of copper ever farther, but the same schemes applied to fiber can achieve magnitudes more.
As to my use of Shannon's Law, I merely used the theorem in it's pure theory, without considering real-life conditions, to simplify the argument a bit. Of course real-life limits come in, and they can also reduce potential bandwidth available on coax.
Cheers,
a333