Some interesting information about ADSL and ADSLMAX for you all toread

I have just posted a few bits via some old webspace of mine.

It has some good information (on the Virata page in particular) about
how ADSL works.

Also an observation on my results after MAXing.

I hope you find it interesting and informative.

Of so, I would love to hear from you.

Its at http://myweb.tiscali.co.uk/excomms/

Mike

m wrote:


> Also an observation on my results after MAXing.

you say "As you can see I am syncinng at over 7M but due to the
corrections needed on my (quite long) and therefore quite noisy line,
am only getting an effective transfer speed of about 1M."

I think the 1M is due to other issues than error corrections, like an
ISP throttle, BRAS rate profile or congestion of some sort. Having
pulled 400 kbytes/s reliably from lines syncing at 4600 with plenty of
errors I don't buy into the theory above.

Phil

PhilT wrote:
> m wrote:
>
>
>
>>Also an observation on my results after MAXing.
>
>
> you say "As you can see I am syncinng at over 7M but due to the
> corrections needed on my (quite long) and therefore quite noisy line,
> am only getting an effective transfer speed of about 1M."
>
> I think the 1M is due to other issues than error corrections, like an
> ISP throttle, BRAS rate profile or congestion of some sort. Having
> pulled 400 kbytes/s reliably from lines syncing at 4600 with plenty of
> errors I don't buy into the theory above.
>
> Phil
>
Thanks Phil

We shall see!

Despite any other problems, am I right in thinking that 'corrected
blocks' are ones where there are errors (due to the line etc?) that the
modem is having to correct, whereas 'uncorrected blocks' are ones that
are OK. Or is it that 'uncorrected blocks' are unmendable and therefore
the modem has to request a 're-send'

I must admit that when it works, MAX is great but, like some others, I
wonder if the old reliable - if slower - standard ADSL was better

Mike

On Wed, 2006-06-14 at 10:43 +0100, m wrote:
> I must admit that when it works, MAX is great but, like some others, I
> wonder if the old reliable - if slower - standard ADSL was better

In my case, the old ADSL was not only more reliable, but also faster --
by a factor of two. Once every few days, 'Max' will sync at just below
1152Kb/s, leaving me with a half-megabit cap for another few days.

Eventually it goes back up (not usually within the allotted three days),
and is dragged down again by another momentary blip in the sync rate.

If we could just impose a lower limit on the sync rate, it ought to be
fine. It was fine at 1Mb/s before the 'upgrade' to Max.

Alternatively, we could abolish the silly BRAS cap and let the ISP
rate-limit to precisely the line sync rate; that would also suffice. At
least it'd only be a shade under 1Mb/s then, rather than half.

Thanks for the information, btw; it made very interesting reading.

--
dwmw2

|
| Despite any other problems, am I right in thinking that 'corrected
| blocks' are ones where there are errors (due to the line etc?) that the
| modem is having to correct, whereas 'uncorrected blocks' are ones that
| are OK. Or is it that 'uncorrected blocks' are unmendable and therefore
| the modem has to request a 're-send'

Is interleaving switched on ? that would account for the blocks being
corrected.
My Draytek always has 0 corrected blocks in fast mode.

Chris..

m wrote:

> Despite any other problems, am I right in thinking that 'corrected
> blocks' are ones where there are errors (due to the line etc?) that the
> modem is having to correct, whereas 'uncorrected blocks' are ones that
> are OK. Or is it that 'uncorrected blocks' are unmendable and therefore
> the modem has to request a 're-send'

the latter, blocks that couldn't be corrected with error correction.

Phil

PhilT wrote:
> m wrote:
>
>
>>Despite any other problems, am I right in thinking that 'corrected
>>blocks' are ones where there are errors (due to the line etc?) that the
>>modem is having to correct, whereas 'uncorrected blocks' are ones that
>>are OK. Or is it that 'uncorrected blocks' are unmendable and therefore
>>the modem has to request a 're-send'
>
>
> the latter, blocks that couldn't be corrected with error correction.
>
> Phil
>

Thanks Phil - thought it was that.

I have now been leaving the router on overnight and the relative number
of uncorrected blocks seems to be falling. Not sure if thats cos the
network is more stable at night or cos the router is a 'happier bunny'

Back to about 4.5M on a 6M connection and so uncorrected blocks don't
show so much.

Mike