MacLeech
The one and only
Premium,MVM
join:2001-07-14
SoCal
edit:
December 19th, @04:16PM
| reply to N3OGH
Re: No good on an overloaded node
The fiber connection between the node and the headend (CMTS) isn't hitting capacity... those links have several gigabits of capacity. The DOCSIS RF data channels have much less capacity then the fiber it's using.
You have to realize that the DOCSIS modem channels are just a couple of the 117+ channels that the average cable system is broadcasting to each and every customer. The fiber from the headend to the node carries all of it without a problem.
Maybe your thinking of the connection the headend has upstream to the nearest peer or transit point... it has to handle the capacity of everybody being serviced by that headend.... that's where GigE or 10G links come in handy... |
|
N3OGH
Will it all be Obama's fault now?
Premium
join:2003-11-11
Philly burbs
 ·Verizon Online DSL
| So, what you're telling me is that the problem is on the coax side, correct?
IE, the implementation of DOCSIS 3.0 will resolve a lot of the problems folks like me experienced with cable modem service by opening up more capacity between the node and the end user???
--
Petty people are disproportionably corrupted by petty power… |
|
majortom1029
join:2006-10-19
Lindenhurst, NY
| Its not the medium either, coaxial can handle speeds of 270mbps atleast. Its how docsis is done on coaxial that is the problem. Other companies not using docsis tech (like the long gone narad tech) were able to get high speeds over coaxial using switching and other technologies.
The thing is that docsis is a standard thats why the cablecompanies sue it. |
|
MacLeech
The one and only
Premium,MVM
join:2001-07-14
SoCal
| reply to N3OGH
No the problem isn't on the coax side either.... it also has gigabits of capacity.
The issue has been how many customers (and how they were configured) were connected to the CMTS ports and how much capacity those ports had or how much backhaul capacity that CMTS has.
The CMTS is the cable equivalent of a DSLAM. |
|
majortom1029
join:2006-10-19
Lindenhurst, NY
| A cable node has only a certain amount of bandiwth, and cable companies connect something like 250 - 500 houses to that one node. Docsis technology even with docsis 3 only gives a small amount of bandwidth. Tech like narad used switches at everybodies houses and at the node.
So how is it that the bandwidth problems arent at the coaxial side of the node? Its due to docsis tech that give the node such small bandwidth.
Other techs can be used to give better bandwitdh. |
|
MacLeech
The one and only
Premium,MVM
join:2001-07-14
SoCal
| said by majortom1029  :So how is it that the bandwidth problems arent at the coaxial side of the node? Its due to docsis tech that give the node such small bandwidth. Its how the DOCSIS tech is implemented... a node and the coax connected to it handles multiple channels, usually 115 or more. The DOCSIS channels are usually just a couple of those, more channels could be run for more bandwidth.
Imagine if cable went to a IPTV type distribution system where your video signal just used a couple channels and the rest of the 115+ were dedicated to data.... it'd be a few gigabits per second of capacity per node and attached coax. |
|
floyd007
join:2004-06-07
Manassas, VA
| reply to majortom1029
said by majortom1029 :Its not the medium either, coaxial can handle speeds of 270mbps atleast. Its how docsis is done on coaxial that is the problem. Other companies not using docsis tech (like the long gone narad tech) were able to get high speeds over coaxial using switching and other technologies. The thing is that docsis is a standard thats why the cablecompanies sue it. Every body here are forgetting the physics of metals. Metals do not conduct well to electricity or electrical magnetic radiation naturally and very poorly to boot. You will have to force a huge amount of EM radiation just to get it to propagate through the wires. (line extenders, amplifiers, etc.,) Metal / Coaxial is NOT the new way to go. Cheaper but ultimately will have same issues with any other HFC protocol |
|
MacLeech
The one and only
Premium,MVM
join:2001-07-14
SoCal
| said by floyd007 :Every body here are forgetting the physics of metals. Metals do not conduct well to electricity or electrical magnetic radiation naturally and very poorly to boot. You will have to force a huge amount of EM radiation just to get it to propagate through the wires. (line extenders, amplifiers, etc.,) Metal / Coaxial is NOT the new way to go. Cheaper but ultimately will have same issues with any other HFC protocol Wow... don't let just about every communications medium in existence in on THAT secret, they may just stop working when wiring stops conducting electricty... it's amazing they call most metals conductors to begin with... |
|
MOTO6809
join:2007-11-05
Springfield, MA
| The big question is how many nodes share the same downstream QAM? In some cases it could be 4 nodes to 1 downstream QAM. In other cases it could be 1 QAM down and 1 QAM up per node(approx 250modems per node) using DOCSIS1.x and 2.0.
When they start using DOCSIS3 I would imagine nodes would be much smaller in respect to how they are combined/split within a headend. |
|
bunklung
join:2002-07-13
Northampton, MA
| reply to MacLeech
said by MacLeech :said by majortom1029 :So how is it that the bandwidth problems arent at the coaxial side of the node? Its due to docsis tech that give the node such small bandwidth. Its how the DOCSIS tech is implemented... a node and the coax connected to it handles multiple channels, usually 115 or more. The DOCSIS channels are usually just a couple of those, more channels could be run for more bandwidth. Imagine if cable went to a IPTV type distribution system where your video signal just used a couple channels and the rest of the 115+ were dedicated to data.... it'd be a few gigabits per second of capacity per node and attached coax. I have a hard time imagining if cable went IPTV and suddenly the entire node of 250-500 customers only needed a "couple" of channels for their video and the rest of the 115+ channels being used for data.
I hope the entire node doesn't decide to watch all different channels.
Does not compute  IPTV and SDV has it's place, but it's certainly not going to be a "couple" of channels OR the utopia of not having anymore broadcasting/multicasting of channels. It'll be a mix. There's no sense in broadcasting MTV2 in the node when 1% of the node statistically is watching it. While prime time TV (major networks) would still be broadcasted in the node.
The cable companies need more bandwidth and they know it. DOCSIS 3.0 is still a band-aid to their bandwidth crunch. They can bond channels, but it's still QAM256! DOCSIS 3.0 doesn't give the cable company more bandwidth in the node. When they bond a channel, that's one less channel for video.
DOCSIS 3.0 just gives cable another way to ADVERTISE higher speeds. Advertising and delivering are two different things.
The cable company needs to split the node, and that costs money. How does 32 customers sound? How much would that cost them? It's costing Verizon BILLIONS to do it. |
|
