Hi pass pass,
This responds to your uplinked and message 9996.
"Why can't each home be treated like a cubical in a company? In the office, you plug your PC to ethernet and you get high-speed internet."
Commercial building wiring standards (EIA/TIA 568 and appendices) have been largely responsible for allowing the plug and play characteristics you've cited. Here, maximum distances are 100 meters, including tail connections at the desk and closet, which make it a wholely controllable environment for universal vendor adaptation at all approved speeds and protocols.
The local loop, on the other hand, enjoys no such pre-defined data handling characteristics. Its properties are conducive to handling analog voice under the constant presence of battery (-48v d.c., which has proved to have a binding, or sealing, effect on the pair that makes voice more reliable than data services on some types of splice sections) over a range of from several hundred feet to over 30,000 feet at times, aided by ancillary network elements, such as repeaters, termination sets, and load coils. Pre-historic voice lines (prior to the proliferation of carrier systems) actually extended cross-country, fully thousands of miles in length, using vacuum tube amplifiers every 6000 feet. Here a single cable pair, in combination with several others, was able to support several voice conversations each through phantom circuit techniques through the manipulation of magnetic flux characteristics in transformers. But I digress.
In other words, the local loop was never designed to handle the types of data transmission that would be considered desirable today, as witnessed by your questions.
The loop was designed to handle voice, which is why modems use analog modulation schemes that have characteristics and properties based on the use of frequencies in the same range as the human voice, mostly.
Some local loop adaptations are conducive to digital forms of line coding, though. These fall under digital data categories which are mostly derivatives of T1 bipolar return to zero, and the V.90 class of modems you alluded to, which are capable of roughly 56 kb/s under the right circumstances.
Not all local loops will support the 56 k, remember. Those with load coils, and those that attach to subscriber carrier systems, particularly, wont work, relegating the 56 kb/s V.90 modem to operate at the lower defaults of 28 k and below.
"Why can't a residential access be that simple? I think given the bandwidth achievable by DSL or cable, shouldn't ethernet be doable over telephone line or coax?"
DSL, like 56 kb/s V.90 modems, also demand that certain criteria be met. If those criteria are not met the telco has two choices. They can either go out and fix the problem that prevents DSL from operating (remove the load coils, or swing the loop off of the subscriber carrier system and onto a discreet pair), or they can declare the service unavailable at this time.
The plug and play Ethernet that you are asking about is the goal of several FTTC/FTTH initiatives now under way. But until they have established a foothold, it's still pretty much going to be hit and miss, depending on where you live and who your provider is, as to whether or not you will be able to get your 56 k or DSL service to function as desired.
As for Ethernet being deliverable over Telephone lines, yes they are. But at speeds lower than standard 802.3, which would ordinarily be 10 Mb/s. Some vendors have last mile Ethernet solutions that function at 1 or 2 Mb/s, I believe. Does anyone recall their names? I think that NT has an interest in one of them...
As for cable, DOCSIS already makes use of a modified form of frame delivery, which inherits some of its characteristics from Ethernet. I'll let ftth or Denver Techie, or someone else, explain this model further. Right now, my soup is getting cold ;-)
FAC |
