Router Stats

pcombo

GunJack wrote: »

what a load of tosh !! Wouldn't alter the sync speed, and would slow actual download speed achieved due to the interleaving overheads on the data packets !!!!!

not a load of tosh as you put it, Interleaved does give you a faster connection but a lower ping, And it's also good becuase it fixes errors on the line.

kwikbreaks

8128 is the maximum possible sync on ADSL. The maximum sync with interleave on is 8096 although with many routers you don't get that you get 7616.

With lines that cannot deliver maximum sync interleave will often provide a faster sync as the router figures in the error correction when negotiating the connection.

You need a good line or a higher noise margin to get the most from fastpath otherwise the retransmits because of the very much reduced error correction counteract the lower ping. For download throughput interleave is a better option but gamers seek low pings so they always want fastpath.

Regarding the stats - although it is sitting at maximum sync the noise margin is only 7dB and I'd have expected to see higher with 25dB attenuation (in my ADSL days I got that on 35dB attenuation). It could be that the DSLAM transmit power has lowered itself (not given in those stats) but I'd be more inclined to suspect noise pickup - most likley on the internal wiring.

Speeds lower that the profile figure could be over subscription by the ISP or problems with user equipment. Is the computer used to test speeds wired or wireless for instance. If wired then I'd be more incline to suspect the ISP although before pointing any fingers more testing is needed.

GunJack

kb raises some good points there....testing should really be done from the test socket (the one behind the master socket faceplate) to eliminate internal wiring issues. Also about relative snrm, although how picky the DSLAM is may affect this, especially if the profile is unstable. Monitoring the line with Routerstats or DMT would give a better picture of how the sync/snrm is behaving over a period of time rather than just a snapshot.

My money's still on congestion in lieu of more detailed info

rmg1

Apologies for borrowing this thread, but what can be done if you don't have a master socket (which is true in my case)?

kwikbreaks

GunJack wrote: »

My money's still on congestion in lieu of more detailed info

Mine too although as Plusnet shaping prevents any serious (ab)use of their bandwidth I would expect to see good speedtest results.

I suppose it could be a congested VP from the exchange but blaming the exchange is usually the last resort of a shyster ISP. I think this problem is usually fixed quickly too.

kwikbreaks

rmg1 wrote: »

Apologies for borrowing this thread, but what can be done if you don't have a master socket (which is true in my case)?

You will almost certainly have a master socket although it may be the old style one that has no removable faceplate. If you have one of those then you can't remove the ring wire from extensions without breaching your BT contract. In that case it will make no difference which socket you use for tests. In fact the IPprofile takes so long to react it doesn't matter anyway other than it may affect the stats.

rmg1

Thanks for that.
The only socket I can find (and I've taken the front plate off the wall), is just a faceplate and the wires disappear into the wall. There's nothing behind the face plate.

JJ_Egan

rmg1 wrote: »

Thanks for that.
The only socket I can find (and I've taken the front plate off the wall), is just a faceplate and the wires disappear into the wall. There's nothing behind the face plate.

BT Logo on the front of it ???


My supposed master socket is a small BT connection box wonder if yours is similar .
Just inside the premises .


jje

rmg1

Nope, no logo.
This is the right size for a socket (about 3 inches square) and it's the only socket I can find.

ChiefGrasscutter

pcombo wrote: »

not a load of tosh as you put it, Interleaved does give you a faster connection but a lower ping, And it's also good becuase it fixes errors on the line.

I'm with Gunjack on this.
In fact BT openreach initally stated that lines which were interleaved could not obtain syncs higher than something like 7616 due to the check bytes needed - though they also claimed that it shouldn't reduce your download speed!
In fact routers which are additionally ANSI specifications compliant can go better and some/most will do near/at the max of 8128 if the line allows.
I'm on an interleaved line and my Voyager 2110 can only do 8096 sync max and will not go any higher no matter how high the SNR margin goes, but my Dlink 2640 can easily do 8128 sync. (My line normally has a SNR of around 10). ADSL2 is not enabled as my exchange so I can only comment on ADSL1/max. My download speeds are lower with the Voyager than the Dlink because of the reduction in the sync speed, which reduces the IP setting at the Dslam in the exchange.

Interleaving does indeed help the stability of the line by splitting up the packets - and was actually the initial default design concept of ADSL - fast path was added later.
So either your download speed will be slightly reduced by the inability of the router to connect as fast as it might if it was on Fast path, or you will loose something in the time taken by the router in re-assembling the split packets and if necessary doing the Reed-Solomon error correction function to correct for the errors. These are sometimes listed a FEC errors (Forward error correction)