jackson@ttidcb.UUCP (Dick Jackson) (02/11/85)
OK. So now we have a debate on what factors make a good versus bad dial-up data circuit. Does anyone know? I would guess that if your call is digitized at the serving CO and stayed digital until it got to the SCO at the destination, then you should have a good circuit provided all the intermediate hops were decently within spec. But maybe there are sources of impairment that I am not aware of (e.g. the synchronization slip outage we have been hearing about re. 212 errors). But cross country I believe that calls go analog microwave. In this case there is repeated amplification, which is presumably why you hear thermal noise on most long distance (voice) calls. If on a multi-stage path there are continual conversions between digital and analog there would seem good reason to expect trouble. On what kind of routes would you expect this? When we make voice calls from Santa Monica (Los Angeles) to Huntington, Long Island, you know you're talking long distance; amplitude is down and there's a lot of thermal type noise evident. As I reported in a previous posting, regular 9600 bd modems work well over such lines. Are there circuit impairments (e.g. phase effects) that affect modems but are not detectable by ear? If so, what type of paths are most likely to produce them? One more anecdote: we have done a lot of calling on 224 modems between Santa Monica and Sioux Falls, SD. I don't know whether that qualifies as a rural end point but we have had no trouble at all.
lauren@vortex.UUCP (Lauren Weinstein) (02/15/85)
I think the bottom line is that unless you are paying for data circuits, you have no way to be sure what sorts of circuits you'll be getting, or when existing circuit patterns will be changed. A very large percentage of long distance toll traffic is (and has been) digital for quite a long time, but sometimes there are analog hops at various points along the way. At least theoretically, every time there is an A/D or D/A translation it's another potential trouble spot for dialup modems. There is no guarantee that phase relationships will be stable on voice calls, and these higher speed modems are extremely vulnerable to phase variations. Many of these variations cannot be detected by ear -- the sync slip problem is frequently one such case. Since you're not paying for data circuits, there isn't much you can say if a route you're currently using stops working due to an equipment/path change, so long as the path still provides reasonable voice-grade service. This may be another reason why investing in expensive voice-grade modems may be rather risky--you may suddenly find that calls are no longer working well (due to path/equipment changes) and you frequently will have no recourse since the path is still providing good voice-grade (though not data) service. The variations, within a single call and across calls to various regions, over time, can be quite dramatic. --Lauren--