Kitz Forum

Broadband Related => Broadband Technology => Topic started by: Weaver on April 05, 2021, 11:12:03 PM

Title: SES and ES, and CRCs with L2ReTX
Post by: Weaver on April 05, 2021, 11:12:03 PM
Two questions:

* Firstly, about the definition of SES; Kitz gives a definition:
Quote
As the name indicates this is more worse than an Errored Second (which only needs one event to trigger an ES).
A Severely Errored Second is a one second period which contains 30% or more errored blocks OR several other events such as one or more OOF.

I’m unclear what a ‘block’ means in this context ? Can anyone help me out ?

* Second question: When counters list CRC error counts, is this before the application of any L2ReTX protocol (G.INP or PhyR) ? That is, before L2ReTX can fix the error?
Title: Re: SES and ES, and CRCs with L2ReTX
Post by: burakkucat on April 05, 2021, 11:33:16 PM
* Second question: When counters list CRC error counts, is this before the application of any L2ReTX protocol (G.INP or PhyR) ? That is, before L2ReTX can fix the error?

I believe the count is the number of errors that are left after all attempts at correction have been performed.
Title: Re: SES and ES, and CRCs with L2ReTX
Post by: Weaver on April 06, 2021, 01:20:24 AM
Ah, I know that, thanks !  :)

I don’t know exactly what the definition of SES is though, maybe someone could help me with that.

And I’ve just spotted a typo in that definition on Kitz’ page : "more worse", and elsewhere "additonal" (sic) for "additional".
Title: Re: SES and ES, and CRCs with L2ReTX
Post by: burakkucat on April 06, 2021, 06:02:07 PM
And I’ve just spotted a typo in that definition on Kitz’ page : "more worse", and elsewhere "additonal" (sic) for "additional".

As they are on a main site page I am unable to make the appropriate adjustments.
Title: Re: SES and ES, and CRCs with L2ReTX
Post by: Weaver on September 26, 2021, 02:47:00 AM
> I believe the count is the number of errors that are left after all attempts at correction have been performed.

So, as a check on my understanding, all L2ReTX retransmissions are carried out up to some max count limit or time limit and then the retransmission sequence operation is regarded as ‘fail’. Is that fail counted as 1 CRC error in the stats we’re used to seeing? (Not the number of retransmissions received corrupted.) I ought to read G.INP for some insight.
Title: Re: SES and ES, and CRCs with L2ReTX
Post by: Alex Atkin UK on September 26, 2021, 10:37:58 AM
If only such documents were written in language mere morals can understand. ;)
Title: Re: SES and ES, and CRCs with L2ReTX
Post by: kitz on September 26, 2021, 06:57:51 PM
Quote
I’ve just spotted a typo in that definition on Kitz’ page

Thank you - corrected.

Quote
I’m unclear what a ‘block’ means in this context ?

A block of data in the bitstream.  Best visualised in the diagram on the Interleaving (https://kitz.co.uk/adsl/interleaving.htm) page.

(https://kitz.co.uk/adsl/images/interleave1.gif)

In anticipation of the next question:-
The amount of bits in the block is variable depending upon various factors such as depth of interleaving, type of error protection/correction and even the connection speed.
Theoretically it is possible to work it out based on the configuration parameters set by the CP on the DSLAM and also any configured by DLM   eg these


Code: [Select]
                         VDSL2 framing
                        Bearer 0
MSGc:           -6              150
B:              178             236
M:              1               1
T:              0               5
R:              10              16
S:              0.0950          0.3771
L:              15922           5410
D:              1               1
I:              189             255
N:              189             255
Q:              16              0
V:              2               0
RxQueue:                33              0
TxQueue:                11              0
G.INP Framing:          18              0
G.INP lookback:         11              0
RRC bits:               0               24

A few years ago wombat and I had a go at trying to unravel the relationship between the values.  Wombat spent a fair amount of time on this, but I'm not going down that rabbit hole right now.   I did make some notes in the section Interleaving Parameters (https://kitz.co.uk/adsl/linestats_errors.htm#interleaving_params).  Being that you have BCM PhyR then the G.INP parameters may also be applicable.


PS On there I say "appears to bear some resemblance to sync speed".   
We know that sync speed does affect things like error correction block size based on the set parameters, but arent able to define it to a specific amount.  It's not something that seems to show up in the line stat figures.
Title: Re: SES and ES, and CRCs with L2ReTX
Post by: kitz on September 26, 2021, 07:10:52 PM
> I believe the count is the number of errors that are left after all attempts at correction have been performed.

So, as a check on my understanding, all L2ReTX retransmissions are carried out up to some max count limit or time limit and then the retransmission sequence operation is regarded as ‘fail’. Is that fail counted as 1 CRC error in the stats we’re used to seeing? (Not the number of retransmissions received corrupted.) I ought to read G.INP for some insight.

Yes.   I seem to think we covered this topic elsewhere recently.  Basically retx attempts to do its thing, but if that fails then its passed over to the more traditional forms of error correction and error protection if that is set.   Its possible to use retx and Interleaving/FEC.

 Re-tx has separate counters.

Code: [Select]
                        Retransmit Counters
rtx_tx:         500357          0
rtx_c:          3154            0
rtx_uc:         399             0

                        G.INP Counters
LEFTRS:         1               0
minEFTR:        59987           0
errFreeBits:    3872056703              0

See ~ G.INP Retransmission Parameters & Counters (https://kitz.co.uk/adsl/linestats_errors.htm#retx_params) for explanation.   LEFTRS is the retx equivalent of an Err Second
Title: Re: SES and ES, and CRCs with L2ReTX
Post by: kitz on September 26, 2021, 07:53:52 PM
I don’t know exactly what the definition of SES is though, maybe someone could help me with that.

From SES - Severely Errored Seconds (https://kitz.co.uk/adsl/linestats_errors.htm#SES)
Quote from: kitz
As the name indicates this is worse than an Errored Second (which only needs one event to trigger an ES).

A Severely Errored Second is a one second period which contains 30% or more errored blocks OR several other events such as one or more OOF.

For starters Im going to ignore the part after the OR as events such as OOF should be self explanitory with the aid of the lookup  and going to concentrate on the first part of the paragraph before the OR :

Bearing in mind that an ES is triggered when a single coding violation (eg a CRC) occurs in a 1 second period.... then a SES is triggered if >30% of the blocks contain a coding violation during a one second period.     
Here we cant say a definitive "x errors" because it all depends upon the the datastream and block sizes which can vary as per convo above... so instead its a percentage of the total blocks transmitted during a one second period.

Hope that clarifies :)
Title: Re: SES and ES, and CRCs with L2ReTX
Post by: Weaver on September 27, 2021, 01:19:08 PM
Thank you so much!
Title: Re: SES and ES, and CRCs with L2ReTX
Post by: ejs on September 30, 2021, 09:05:17 PM
A SES is defined as 18 or more CRC errors in one second.
Title: Re: SES and ES, and CRCs with L2ReTX
Post by: Weaver on October 01, 2021, 05:42:00 AM
Wandering off track for a moment, what determines the values of the ADSL B and M parameters?