Re: data rate ratios for varying distance, and H-88 loading
<1. Is it possible to have 7Mbps at 12,0000 feet and only have 1Mbps at 18,000 feet?>
The principle of lower data rates at longer distances is certainly correct. Your cited ratio may be a bit extreme, though, and 7Mb/s may be a bit optimistic for 12kf of route distance from the CO. Note the operative distance for ADSL range is the route distance PLUS bridge tap length, so the effective distance from the CO is typically less than the apparent route distance (by the cumulative BT distance, up to 6kf by the design rules.) I don't see that little wrinkle noted very often, and BT is a fact of life in many loops.
<2. Lets say you're installing ADSL, and are under 18,000 feet. Given the fact that early
systems might prove to be incompatible, I want to leave that part out and ask about
"loading coils". Since ADSL doesn't work much after 18,000 feet and since loading
coils are usually beyond that point (right?), then how much of a problem is going to be
generated by loading coils under 18,000 feet.? What are the problematic figurs of coils
being under 18k, and how is the DSL industry going to combat that problem?>
You probably understand H-88 load coil spacing, but I can't be sure from the way you phrased this question. H-88 loading (the vastly predominant exchange load pattern in North America)calls for the first load nominally at 3kf from the CO, and loads thereafter at 6kf intervals +-2%. The end section (loop section beyond the last load) cannot be <3kf, and BT can never be loaded.) Only exchange grade subscriber lines over 18kf of route length to the CO should ever be loaded, with one exception I'm aware of that I'll mention below. Note that older voice grade special services of various lengths may be loaded according to a specific design, but such analog circuits tend not to be out there anymore in any real volume. One "exchange grade" exception to the 18kf rule is CO Centrex. Loading design rules called for any CO CTX loops between 12kf and 18kf to be loaded at 3kf and 9kf, the only exception I am aware of to the rule that H-88 loading always entails a minimum of three points of loading. (The logic here is that a 12+ kf loop on 26 ga copper had just marginal voice band loss performance, and since there was a very high likelihood of two such loops back to back for intrastystem calling, the transmission performance on such calls would have been subpar. Hence, the tighter loading rules.)
The net is: Telco outside plant designed according to the various generations of resistance design rules would very rarely ever have terminals less than 18kf from the CO fed with any loaded pairs. If it ever happens, it is most likely a design mistake, a left-over analog special service loop (rare), or some old CO CTX feeder plant (rare). Voice services work fine over CO CTX loaded plant, so it is entirely possible that no engineer would bother to engineer a job to remove loading when a CO CTX was disconnected. On the other hand, CO CTX has not been a big factor in business telephone service in quite a few years, and it was never much of a factor at all in locations >12 kf, because of distance charges typically in the tariffs. You should be confident that there should only be a handful (say, <2%) of terminals more than 18kf from their CO, fed by loaded copper pairs. However, note that the actual loading of longer pairs occurs at 3kf, 9kf, 15kf, etc. There is loading inside some imaginary 18kf radius circle, but it is not problematic for loops that serve customers inside 18kf.
< I have heard that IDSL is the only alternative, is this also true? >
Alternative for loaded plant over 18kf, or plant over 18kf? IDSL requires non-loaded plant, so you have to remove any load coils that are installed to turn up an IDSL service. If you are going to the trouble to remove loads anyway, you could use any 2B1Q line code technology, most commonly, HDSL (two pairs required) of SDSL (single pair). The operative factor is that 2B1Q lines can be repeatered, but CAP or DMT (ADSL) cannot. It gets to be a pretty ugly install, though, to be removing loads and installing field-mounted line-powered active electronics. Not a volume business you want to get into! Note that the soon-to-be-standardized single-pair, T1 rate symmetric HDSL2 will not be a factor beyond about 15kf. It uses a new non-2B1Q line code (a PAM variation).
< Can DLCs (digital loop carriers) be a pain in the royal a$$ here too?
Does that mean that you need to put DSLAM at the DLC so digital traffic can "skirt" around the problem.?>
Yes! The exact topologies and packaging vary, but at the end of the day, your characterization pretty well captures the net effect. Some of the newer Next Generation DLCs propose a higher degree of integration, so there is no discrete DSLAM in the remote enclosure, but electrically, there is a division in the data path at the remote terminal between the channel switched voice and the broadband data.
This gets really messy when you take into account what is available to a CLEC -- will the FCC decide there must be an unbundled network element (UNE) for the loop from the DLC remote terminal to the subscriber? How do you price it? If the CLEC has the only direct connection to the subloop, how does the ILEC test it? (they are still responsible for the physical layer when they sell a UNE). Where does the CLEC equipment (DSLAM) go -- in the existing ILEC cabinet? How do you apportion the limited space? Do you provide a separate cabinet for the CLEC(s)? Who secures the right of way for the new cabinet and power connection? How do you pay for thousands of new CLEC cabinets (say, $4500 each, turnkey, bare shelves with power) collocated with the thousands of ILEC DLC remote terminals?
I suggest we will need to see what the FCC rules in CC Docket 98-188 before we can really get engineers working constructively on the broadband-behind-DLC issues. It would certainly be a big risk to bull ahead with implementation absent any clear understanding of the ILEC obligations to be established under 98-188 (but note US WEST.) Expect an FCC 98-188 order in February 99.
I agree, no one could reasonably claim that this is not all very confusing. The potential for extremely difficult regulatory rules (on unbundling and remote collocation) as well as the potential for resulting implementation complexity are daunting.
|
