Kitz Forum
Broadband Related => FTTC and FTTP Issues => Topic started by: NewtronStar on September 10, 2017, 08:06:14 PM
-
As that's 91% it could be you're on high retransmission
What is the difference between G.INP High re-transmission and Low re-transmission have seen this cropping up more often here.
lets take two lines one has a INT:4 INP:54 the other line has INT:8 INP:45 which one is high or low ?
Admin: Split from original thread at - 3dB Target SNRm Live? (http://forum.kitz.co.uk/index.php/topic,17498.0.html)
-
INP value into the fifties, probably Retransmission High profile.
Interleave depth not useful for distinguishing the two profiles, interleave depth is set to match the DTU size and probably varies with a line's bandwidth more than anything else.
INPRein of 1 or higher on Bearer 0 possibly also a sign of Retransmission High.
There's some irony in that after reducing the target SNRM to increase bandwidth, this can then result in a Retransmission High profile, with a lower IP profile and lower throughput than expected for the line rate that the modem reports.
-
Thanks EJS look I'm not trying to derail this thread just trying to get answers that others may be also asking themselves.
My own Bearer on 0 shows INPRein: 0.00 0.00
-
INPRein of 1 or higher on Bearer 0 possibly also a sign of Retransmission High.
I'm of the opinion this might be right.
Going by my last 5 or 6 resyncs, anytime I have had Bearer 0: INPRein: 0 - I have had an IP profile matching that of retransmittion low.
When I have had Bearer 0: INPRein: 1 - I have had an IP profile matching that of retransmittion high.
-
Wish I had this info a year ago was banging my head against a brick wall as to why the IP profile was stuck at 91-92% the INP: was 50-52 and INT:was also stuck at 4, unfortunately don't think i've any records of the INPRein status at that time.
-
On my phone so screenshot cuts half the graph off, but I can see you had Bearer 0 - INP-REIN 1 for a few months till mid December. It's been 0 since then. Does that sound about when the IP profile increased? That would certainly add to the theory that INP-REIN on Bearer 0 is an indicator as to what retransmission profile a line is using.
Anyone else able to confirm if they have an IP profile around 91% and INP-REIN is 1 on Bearer 0. Or anyone who can debunk it.
Might be worth splitting the last few posts to their own thread?
-
Well done John yes that was before I moved from EE to Vodafone the missing part and back to EE somehow it got stuck on High after G.INP was activated on the upstream for a few days
EDIT: I'll be honest it became stuck when I was testing the BT80 RF3 and wired it up incorrectly and the DLM had already recorded my actions even after I remedy my action :blush:
-
Wish I had this info a year ago was banging my head against a brick wall as to why the IP profile was stuck at 91-92% the INP: was 50-52 and INT:was also stuck at 4, unfortunately don't think i've any records of the INPRein status at that time.
iirc I thought we'd already noticed that your 'R' value (http://www.kitz.co.uk/adsl/linestats_errors.htm#R) was higher and therefore the level of redundancy would be higher, which would in turn affect the overhead allowance?
~ NGA FTTC IP Profile (http://www.kitz.co.uk/adsl/IPprofile.htm#FTTC_ipProfile)
I guess I should also add a note to the other sections such as the calculator to make it clearer. :/ I'll do that later if I get chance.
-
On my phone so screenshot cuts half the graph off, but I can see you had Bearer 0 - INP-REIN 1 for a few months till mid December. It's been 0 since then. Does that sound about when the IP profile increased? That would certainly add to the theory that INP-REIN on Bearer 0 is an indicator as to what retransmission profile a line is using.
Anyone else able to confirm if they have an IP profile around 91% and INP-REIN is 1 on Bearer 0. Or anyone who can debunk it.
If that can be confirmed then I willl add that info too. It could be an easier way to identify for G.INP lines.
In fact thinking about it for non ginp lines the easy way to id would be if INP >3 ?
Might be worth splitting the last few posts to their own thread?
I'll do that in a mo.
----
ETA - Thread split as requested.
Not going to get chance tonight to edit the main site. Will someone please remind me if I forget.
-
iirc I thought we'd already noticed that your 'R' value (http://www.kitz.co.uk/adsl/linestats_errors.htm#R) was higher and therefore the level of redundancy would be higher, which would in turn affect the overhead allowance?
It's unrelated to R. The FEC overhead has already been subtracted to get the "sync speed" (Net Data Rate).
I think non-retransmission, on the high FEC+interleaving setting, can have INP 8, and if they have a higher proportion of FEC data, then the Net Data Rate will be lower, but the IP Profile will still be the usual approx. 96.7% of the Net Data Rate.
There is a table giving the parameters of the retransmission profiles in SIN 498, in section A.4.2.2.8, although the table is possibly incorrect (it gives INPMIN_REIN_RTX of 1 for both low and high retx profiles) or at least does not correspond to what's actually in the live network. But it's possible that the lower 91% IP Profile is factoring in the higher SHINERATIO_RTX.
-
I have an INP value of 52 with the ~91% IP profile.
Wonder what parameters DLM looks at to change retransmission to low.
Or it is always high with a sub 6 target margin? Mine is 3dB.
-
Or it is always high with a sub 6 target margin? Mine is 3dB.
I would think so as the SNR is lowered the level of background noise also increases so more INP will be needed on that line to overcome errors.
-
IP profile is and has been 92.4% for my last 2 sync rates, 5dB then 4dB target. Bearer 0 INP-REIN = 1 still.
Can't see Bearer 0 INP-REIN on MDWS multi-user page so can't see every line but having had a quick look yesterday seems every user I checked that was on a sub 6dB profile had Bearer 0 INP-REIN = 1.
So with Retx low and 6dB target snrm you can have higher throughput than 5dB with Retx high. It's only with the step to 4dB that my throughput/IP profile has increased from what it was with 6dB. Anyone who's line settles on 5dB is likely getting a speed drop, that's well designed.
-
I would think so as the SNR is lowered the level of background noise also increases so more INP will be needed on that line to overcome errors.
So is there a balance that can be struck by forcing the modem max sync rate? i.e I am syncing at 74+, but my profile is at 68. Say I forced sync to 72 max, would that result in an increased profile or does this mean that if you go sub 6db you are likely to end up with no effective increase in profile?
-
Anyone who's line settles on 5dB is likely getting a speed drop, that's well designed.
Unfortunately for me that has come true this morning after sub SNRM took affect at 5dB lower throughput with a higher sync Retransmission High nowhere to go here as the max for this service is 40/10
-
Probably a daft question !
Is the data percentage calculated on actual DS data rate (as obtained from a speed tester) / DS synch speed or DS IP profile / DS synch speed ?
I believe the latter as in my case this gives me 91.1%.
I am G.Inp'ed with INT, 16 & INP, 48, my B0 Inp Rein has been zero for as long as I have been DSLStats/MDWS monitoring.
-
IP Profile / Modem Sync Rate X 100 = Throughput ratio, When looking at your stats your sync rate is higher than your attainable by 3 Mbps, my guess is the IP profile is stuck at your current attainable sync rate
-
IP Profile / Modem Sync Rate X 100 = Throughput ratio, When looking at your stats your sync rate is higher than your attainable by 3 Mbps, my guess is the IP profile is stuck at your current attainable sync rate
Thanks for the confirmation.
Makes sense, using the speed test obtained DS figure would be a very variable factor.
After G.Inp implementation on my line some 7 weeks after ADSL/VDSL conversion my DS synch increased quite considerably but the data throughput actually slightly decreased which surprised me.
After raising this on the forum was advised by some of the more knowledgable members that this was to be expected due to lower throughput ratio associated with G.Inp, my IP profile/data throughput ratio was approx. 91%.
Having read the posts on this thread I am more inclined to believe that I am on Retx high profile, even though the indicators as mentioned would appear to indicate that I am Retx low, rather than a stuck IP profile, this would explain my current throughput ratio of 91.1%.
I also appear to be on the SNRM decrease program, changed from DS 6 to 5dB. on 27/09, probably not relevant to the data throughput ratio in my circumstances.
-
I went through the motions of asking my ISP, Plusnet, if they had any control over G.Inp retx parameters, one of the technical staff gave the expected answer, currently no, control is at the mercy of DLM, absolutely no reason to doubt this prognosis.
Having always been on retx high profile since G.Inp implementation (verification by calculation of data throughput ratio) I wonder if DLM will at some stage decide to change this to retx low and so boost my DS data speed, all be it very slightly ?
Indeed, what is DLM criteria for retx high/low profile selection ?
My DS SNRM is currently 5dB, recently reduced from 6db on what I would consider to be a very stable and noise free line (see MDWS stats.)
-
It tries retx low within a couple days if the line is stable enough. Your line perhaps wasn't stable enough.
Since my line went to a lower xdB profile I've been on retx high. Looks like most on a lower xdB profile are on retx high.
I have a theory that:
Bearer 0 INP_REIN = 0 is retx low and
Bearer 0 INP_REIN = 1 is retx high
Mouselike is a very handy site for getting your IP profile quickly.
https://windows.mouselike.org/be/index.asp?DoAction=BrasChecker
-
j0hn, thanks for your feedback.
Yes, I have been following the thread regarding possible markers for retx high/low state identification.
In my case, bearer 0 INP_REIN has always been zero, before and after G.Inp application on 27/07/17, see my MDWS recordings.
I would consider my line to be very stable and noise free, see MDWS records, I was virtually always ILQ green even before G.Inp application.
I am certainly convinced that I went straight on to retx high profile as I immediatly queried the decrease of DS data throughput when my DS synch had increased my approx 3 Mbps, it was only then that I became aware of the 91% (approx.) throughput ratio penalty associated with G.Inp and later became more informed regarding the retx high/low parameter from forum posts.
My throughput ratio has never been anything other than 91% since G.Inp application, I do monitor this quite religiously and would have noticed any changes to a 97% (approx.) figure.
I also run a Sam Knows White Box, any change in DS data speeds would show up there.
Tried the "mouselike" link but was concerned about the warnings displayed so did not proceed.
-
Tried the "mouselike" link but was concerned about the warnings displayed so did not proceed.
Don't worry about the warnings, they apply to the BT speedtester too and just mean you can't get your IP profile twice in a short period of time.
Here's another quick link for IP Profile but needs Flash to run http://www.speedtest.btwholesale.com/PerformanceTesterWS/diagnostics.do
-
Mouselike is a well trusted site that's been around for years. It's the quickest, most convenient way to get your IP Profile.
-
@tiffy
Can you give us a full set of stats please? Perhaps there is some other indication of a retransmission high profile, not necessarily a value that gets graphed.
-
@Dray:
Yes, the BT WS Performance Tester is what I normally use, recently started to give incorrect results when used with Firefox (see other topic) still works OK with IE..
@j0hn & Dray:
Went ahead and ignored the warning, very handy, easier than BT WS tester, thanks.
@ejs:
Ashamed to admit I don't know how to produce a stat listing from my HG612 other than in Win. command window which I can't easily snapshot, don't know how to dump to file ?
-
The raw data should be available from the Telnet Data tab in DSLstats.
-
@ejs:
DSLStats Telnet Data, yes, of course but the only way I can manage to snapshot the stats is a window at a time cut & pasted to a Word file, not very elegant, sure there must be a much easier way ?
Won't permit a (*.docx) Word file attachment (MS Office 2010)
-
type telnet
type set logfile myfile.txt
type open 192.168.1.1
proceed as normal, session is logged in myfile.txt
-
What was your IP profile? and what's your current sync?
-
Just showing what IP profile I get with a DS sync of 39999 Kbps @ XdB 4.0 SNRM INP 52:00
Download speed achieved during the test was - 36.31 Mbps For your connection, the acceptable range of speeds is 20 Mbps-37.05 Mbps .
Additional Information: IP Profile for your line is - 37.05 Mbps
We must remember the max DS throughput on a 40/10 service is 38000 kbps.
-
For a bit of information
https://community.plus.net/t5/Everything-else/Can-t-run-the-BT-Speedtester-flash-not-supported-in-my-browser/m-p/1447605
These are from that thread
interesting but very odd - it reports this
The current Downstream BRAS rate is: 50.17 Mbps
The current Upstream BRAS rate is: 20 Mbps
but my sync speed is 54.19
this would explain the rather low speed test results
and this is a link to the GEA Test Result
https://community.plus.net/t5/Everything-else/Can-t-run-the-BT-Speedtester-flash-not-supported-in-my-browser/m-p/1447775#M17633 a bit from it
Downstream Speed 51.8 Mbps
Interference Duration Longest Occurrence ; 00:00 to 23:45
Interference Location Customer Premise
Interference Observed In Days 14
Points to note - reported downstream speed is lower than sync and I will need to check the transformers local to the router
EDIT
All local transformers except the router disconnected and no affect
Looking at the reported rtx_rx per minute it isn't continuous as it only really spikes a few times with one major one hitting about 2,700 at about 11.25 am so I suspect that something parked under the overhead wire from the pole to the house
-
type telnet
type set logfile myfile.txt
type open 192.168.1.1
proceed as normal, session is logged in myfile.txt
Many thanks, "set logfile.txt" was the bit I didn't know
-
What was your IP profile? and what's your current sync?
IP profile, 34.84 Mbps
DS synch, 38.26 Mbps
Throughput ratio, 91.06%
-
@tiffy
Can you give us a full set of stats please? Perhaps there is some other indication of a retransmission high profile, not necessarily a value that gets graphed.
Thanks to information from the more knowledgeable forum members and a bit of messing about, hopefully HG612 stats in plain text format.
-
Tiffy have you tried to drop your PPPoE session on the router for 20 or so minutes and then start it again to see if this changes your IP Profile.
-
Tiffy have you tried to drop your PPPoE session on the router for 20 or so minutes and then start it again to see if this changes your IP Profile.
No, my last re-synch was on 15/08 as can be seen on MDWS, I try not to disturb DLM even obaying the 2 X 15 min. rule.
Don't know if I can actually drop the session on my HG612 withour resorting to Telnet commands, could re-boot my Netgear DGND3700v2 ?
To be honest, I don't believe there is an IP profile issue, I am just on retx high profile, the throughput ratio indicates this, I get the impression that most forum members on a G.Inp'ed line seem to be the same ?
-
The HG612 only handles the xDSL link to the DSLAM. Your router handles the PPP session. If the router has no way to drop and re-establish the PPP session then powering it off will have the desired effect. DLM takes no notice of this, only caring about the xDSL circuits stats.
The trouble with the BTw IP profile is that most resyncs happen quick enough that the PPP session doesn't drop. Then Plusnet have their own system on top of that.
-
Years ago you would get a new IP profile and new IP address after a reboot of the Router these days it seems you need to drop/or power off the Router for 10-30 minutes to establish a new PPP session, how things have changed.
-
The HG612 only handles the xDSL link to the DSLAM. Your router handles the PPP session. If the router has no way to drop and re-establish the PPP session then powering it off will have the desired effect. DLM takes no notice of this, only caring about the xDSL circuits stats.
The trouble with the BTw IP profile is that most resyncs happen quick enough that the PPP session doesn't drop. Then Plusnet have their own system on top of that.
j0hn, thanks for the confirmation regarding router re-boot & PPP session.
As I said, I don't think I have an IP profile issue as such, my line is just on retx high profile for whatever reason.
Are you aware of anyone who is actually on retx low profile as proven by the approx. 97% throughput ratio ?
-
Yes, I was always on retransmission low until my line was moved to a 5dB snrm profile. It has remained retransmission high after the drop to 4dB then 3dB.
I'd say that if not the majority then certainly a high percentage of lines with G.INP are on retx low.
It's when lines drop to a lower snrm target that that the vast majority appear to be on retx high.
I don't doubt that your line is on retx high, especially if you have a 5dB target set.
My line is currently retx high, with an IP profile of 92.6%
It's been between 92.4 - 92.6% for the past 10 or so syncs.
I find 2 mins to be long though to establish a new PPP session. I can drop the WAN session with a simple on/off toggle on my Asus router so very easy to make sure my IP profile is correct.
-
I went on to 6dB DS SNRM on my conversation to VDSL 06/06 this year, it stayed there until 27/09 when it dropped to 5dB, presumably from the reduction program.
When G.Inp was activated on 27/07 DS SNRM stayed exactly the same, I gained approx. 3 Mbps DS synch but my data throughput rate dropped (approx. 91% ratio), this has been the case ever since.
Seems that G.Inp retx profile control is another one of the mysteries of BT DLM that we are not privy to such as why did you and I wait so long for G.Inp to be applied when many were activated in a much shorter time scale ?
-
Tiffy have you tried to drop your PPPoE session on the router for 20 or so minutes and then start it again to see if this changes your IP Profile.
Update to an old post:
I chose not to take NS & j0hn's advice to drop the PPP session accepting my approx. 91% data throughput ratio was attributed to a retx high profile.
Yesterday I changed out my Netgear router for my newly acquired ZyXEL 1312A router leaving the Huawei HG612 modem active maintaining line synch, this involved dropping PPP session for quite some time while I figured out the necessary settings for the 1312 running in router only mode.
On running BTWS line test today, noted that my DS IP profile had increased from 34.84 to 36.99 Mbps, my DS synch. was still the same @ 38.256 Mbps, DS data throughput ratio now 96.68%, was 91.06%.
Noted that my SK White Box records had also registered a step increase in DS data speed between the 13:00 and 18:00 hrs. readings yesterday, corresponding to the time of the router change out.
MDWS trends do not show any step changes in any parameter over the router change over period, as modem synch. was maintained this is to expectation.
So, looks like the drop of PPP session as advised did update the DS IP profile, good result.
For future reference, when I pension off the HG612 and connect the 1312A in modem/router service, is there any easy way to drop the PPP session without loosing line synch, I am lucky enough to have a very stable line which very rarely re-synch's by DLM intervention, usually only by local power outage or manual action.
-
Thanks for letting us know that can happen more easily these days during a quick modem resync the old PPPoE session tends to survive even after 10-20 minutes after closing the current PPPoE session that is what I've noticed with my own ISP.
And your still in re-transmission Low from what I can see from here :)
If it's the same on 8924 it will be in the status page on the GUI disconnect in my red tick
-
@NS
Yes, looks exactly the same on the 1312A, initially confused as the cursor doesn't change to "link" format when highlighted but it does work, went as far as the "proceed" warning.
Many thanks for the information.
Will know better in the future not to ignore the advice of old hands such as yourself and j0hn.
So, with a wee boost in DS data speed and the impending, free US boost from 2 to 10 Mbps by Plusnet it's Christmas come early for me !
-
Ah, that's actually extremely valuable information.
Your previous post had poured cold water on my theory that Bearer 0 INP-REIN was a direct indicator of retransmission low or high.
With that, I'm still of the belief that...
Bearer 0 INP_REIN = 0 is retx low and
Bearer 0 INP_REIN = 1 is retx high
-
interesting info newt about how sessions take a while to timeout, it may explain the recent reports of stuck ip profiles.
-
Ah, that's actually extremely valuable information.
Your previous post had poured cold water on my theory that Bearer 0 INP-REIN was a direct indicator of retransmission low or high.
With that, I'm still of the belief that...
Bearer 0 INP_REIN = 0 is retx low and
Bearer 0 INP_REIN = 1 is retx high
Yes j0hn, you could well be correct.
In my case I have never seen anything other than Bearer 0 INP_REIN at 0 but have always assumed I was on retx high since G.Inp application by virtue of the DS data throughput ratio being approx. 91%, however, after my PPP session refresh on Friday and the resulting change in DS IP profile looks like I have in fact always been on retx low, this now being reflected in the new DS throughput ratio figure of 96.68%.
As reported earlier, absolutely nothing else that I can see in any parameter has changed as a result of the PPP session refresh, only DS IP profile.
-
It tries retx low within a couple days if the line is stable enough. Your line perhaps wasn't stable enough.
Since my line went to a lower xdB profile I've been on retx high. Looks like most on a lower xdB profile are on retx high.
I have a theory that:
Bearer 0 INP_REIN = 0 is retx low and
Bearer 0 INP_REIN = 1 is retx high
You may be able to test this theory on my line.
It went to 5dB on 5 October, with an INP of 57 and at the same time switched to Bearer 0 INP_REIN = 1.
I have now fixed a wiring issue (http://forum.kitz.co.uk/index.php/topic,19560.msg357883.html) I had and would expect DLM to shortly reverse the changes it made.
-
@gt94sss2
Yes, that will be interesting indeed, would be good to have a positive marker on retx high/low profile, would certainly help to avoid confusion as per my situation where DS throughput ratio indicated retx high but the DS IP profile appears not to have updated until a PPP session refresh was carried out.
Will watch your MDWS stat's with interest as I'am sure j0hn will eagerly awaiting DLM intervention.
-
Ah, that's actually extremely valuable information.
Your previous post had poured cold water on my theory that Bearer 0 INP-REIN was a direct indicator of retransmission low or high.
With that, I'm still of the belief that...
Bearer 0 INP_REIN = 0 is retx low and
Bearer 0 INP_REIN = 1 is retx high
j0hn, some positive feedback for your theory:
By line had a DLM intervention, early hours of this morning, changes noted:
DS I/L changed from 8 to 4 (first change since G.Inp activation)
B0 INP changed from 47 to 52 (first change since G.Inp activation)
DS SNRM changed from 6.8 to 5.7
B0 INP_REIN changed from 0 to 1 (first change since G.Inp activation)
DS synch was the highest I have ever had @ 39595 Kbps.
For MDWS reference, my G.Inp activation date was 27/07/17.
Unfortunately, I did not notice the DLM re-synch this morning and carried out a planned outage to replace my router/modem DSL cable, would have liked to have seen my DS IP profile after the DLM re-synch, too late now.
After my planned re-synch, my DS IP profile has certainly decreased from yesterday and my DS data throughput ratio is now 92%, was approx. 97%.
Looks like I am now on retx high profile and yes, the B0 INP_REIN state of 1 appears to be a positive indication of this.
I read in another post by one of the "BTOR members" that B0 INP in the 50's could also be an indicator of retx high profile, again, my experience would seem to conform to this thinking.
As to why DLM intervened this morning:
I have an exceptionally stable and noise free line (see MDWS), the only reason I can think of is I put my ZyXEL1312A into full modem/router service on Friday, retiring my HG612, after a 30 min. + S/D. I had a few re-synch attempts until I worked out that the DSL LED being orange means VDSL synch OK (finally read the manual) was expecting green, definately no more than 3 attempts, is DLM punishing me for this ?
Edit 6/11/17:
On closer examination of MDWS noted an occurance of 11 SES's (very unusual to record any on my line) just before the DLM intervention, perhaps a contributory factor ?
-
I had a few re-synch attempts until I worked out that the DSL LED being orange means VDSL synch OK (finally read the manual) was expecting green, definately no more than 3 attempts, is DLM punishing me for this ?
How close together were the re-syncs? I've noticed in the past if I re-sync too often within a short timeframe it'll make DLM intervene the following day.
-
Quite close as the orange DSL LED on the ZyXEL 1312 confused me, was expecting green and had foolishly not consulted the manual at that stage.
Yes, I normally obey the 2 X 15 min. rule on any planned re-synch's as per kitz's excellent knowledge base, just panicked a bit on this occasion.
I have had input from "NeutronStar" on the DLM re-synch, advises that it could be the start of the xdB profile change program as my DS SNRM did decrease as a result of the re-synch ?
-
Had the PN free US speed change applied this morning, DS synch rate currently to expectation for my line.
Did note that my retx profile appears to have returned to low with the associated re-synch, again, the observations confirm "j0hn's" theory on B0 INP_REIN state which reverted to zero, also DS interleaving reverted to 8 and INP to 48, practically the same as pre last re-synch on 7/11/17 when retx high profile was applied.
I did loose just over 3 Mbps on DS synch rate but my DS data throughput rate has just marginally dropped, presumably now at close to 97% TP ratio again.
Can't get my BRAS IP profile as currently neither the BT WS performance tester, "further diagnostics" or the "mouselike" checker will complete.
My DS SNRM has increased by approx 2dB so will likely get some future DLM intervention on that.
So, looks like the B0 INP_REIN state is a clear indicator of retx profile state.
Perhaps even INP being >50 as has been stated by other forum members ?