It could be difficult to tell the difference between a lot of hither-and-yon movement of files around the internet vs simple noise requiring lots of resent packets. In my case, I concluding the slowness problem was line-noise because it was often there when we would pick up the phone (not always) and I would consistently get url load times and file transfer times and mail download times that were much slower than the connection speed would otherwise indicate. I can particularly tell the difference now that the noise was eliminated. A 42kbps connection FEELS like a proper 42kbps connection instead of a 9600. I don't get the random disconnect, and I get fewer errors and delays when accessing my pop mail.
I would think one measure of noise would show up in pings. If you select a series of sites to ping for a short while now and again, I would think that if it was a noise problem you would see lost packets. If the pings did a lot of transferring from this router to that router, taking side routes, you would get them back with slow timings and probably no loss. praedor On Tuesday 04 June 2002 12:44 pm, Randy Kramer wrote: > Praedor, > > Thanks for the response! > > Praedor Tempus wrote: > > First, the obvious question...did your high disconnect problem start > > coincident with the new policy from your ISP? > > It's hard to tell. I'd say it this way: I started noticing more > disconnects (or maybe just got more frustrated with the disconnects -- > I'm not really sure which). Then I contacted the ISP, and they told me > about this 4 hour / 12 hour policy. (I'll have to go look for this in > writing again -- they promised to send it to me once -- they insist it > was part of what I signed when I first signed up with them (about 8 > years ago), but it was not -- it is something they added since. > > So now I associate the start of the high disconnect problem with my > "discovery" of this policy. > > > Though you pay for an > > "unlimited" connection, I could see the point of you being disconnected > > after X minutes of inactivity, so long as X isn't too short. > > IIRC, they always had a policy of disconnecting after X minutes of idle > time (15?) -- this additional policy was probably in response to people > like me who started checking for mail every 10 minutes. > > > All the suggestions provided were solid...but none would work for me. I > > had a similar problem to you but it was obvious that it was a line > > problem, not an ISP problem. Our phoneline would produce lots of static > > quite often. We noticed it was worse for a period during and after a > > rain. The phone company came out and, of course the first time, heard no > > static at all. Second time, they came out and heard the static. They > > replaced our old outdoor junction box which coincided with the > > termination (for a few days at least) of static. Finally, the static > > returned as strong as it always had been before and they came out again > > and located the problem...the connection from our house into their main > > line, ground level. The twisted pairs were old and the insulation had > > worn off in a few places. They fixed this and the static (and slow modem > > connection speeds and disconnects) went away. > > I'll switch my voice and data lines and spend more time listening. > > > This brings to mind another test to suggest whether you have a line > > problem or an ISP problem...how are your connection speeds? When I was > > having the line problems, my 57kbps modem would often be connecting at > > sub-20kbps (sometimes even 9700). > > Almost every time I check, my connection occurs at 33,400 IIRC. On very > rare occasions I can remember some other speeds, but that's over the 8 > years or so I've been using the ISP. > > > Even when it made a reasonably fast connection, the > > actual performance was much poorer than the speed would otherwise > > suggest. Even though I might get a 33kbps connection on occassion, the > > actual data transfer rate would be MUCH slower. Do you see any of this? > > If not, then it is not likely a line problem. > > We often see file data transfer speeds way below 33 kbps (way below kBps > or whatever that translates to) -- I've always attributed that to the > Internet "hops" and, when I've looked closer, I've seen a reasonable > (inverse) correlation between hops and transfer speeds, or sometimes > attributed slowdowns to an overworked server. Do you have any > suggestions on how to tell the difference between one of those causes > and line problems? (I guess I could "bounce" a big file back and forth > between me and my ISP -- I do have a 5 or 10 mB free web hosting site > available. > > Randy Kramer
Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
