Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[roseway]

The main purpose of this is that it adds HEC errors to the MDWS uploads.

Windows version only at present, others to follow tomorrow.
All versions now available.

http://dslstats.me.uk/downloads.html

[Chrysalis]

thanks

[NewtronStar]

I don't know you and BaldEagle1 decided that HEC errors was just a duplicate of CRC errors a long time ago and the reason why it's uploaded was not necessary so why this sudden U Turn ?

[tbailey2]

Well the data for the two is totally different (look at Chrys data) and BE1 has been uploading it since MDWS first came on line....

[Chrysalis]

yeah, usually hec error count is higher than crc, but I would say within 200% range.  But a few nights back it was really high for me in the 100s at least 10x higher than crcs.

hec's based on kitz wiki seems to be the headers, whilst crc is for the data payload.

[burakkucat]

hec's based on kitz wiki seems to be the headers, whilst crc is for the data payload.

Yes, that I what I have always understood.

[ejs]

I not sure if HEC has any defined meaning for VDSL2 connections which use PTM. HEC as defined in ITU-T G.997.1 is only applicable to ATM (used on ADSL).

[NewtronStar]

I would like here what Roseway and BaldEagle1 say about the relevance of HEC errors and how they differ from say your bog standard CRC errors and a HEC error 

[ejs]

A HEC error should be one or more bit errors detected in the ATM cell header. The ATM cell header includes a check byte.

With VDSL2 connections, they use PTM, so there aren't any ATM cells. The PTM layer should have its own different error types, such as CRC-P and CRCP-P.

A normal CRC error applies to a different layer, it's part of the VDSL2 framing layer.

[NewtronStar]

Well there you go HEC counts are meaningless on fTTC 

[Chrysalis]

I dont know if they meaningless, there wouldnt be a counter if they completely redundant.

[roseway]

I would like here what Roseway and BaldEagle1 say about the relevance of HEC errors and how they differ from say your bog standard CRC errors and a HEC error 

I can't answer that question. I just extract the data and process it. Up until recently I think we had all tended to ignore HECs because we didn't see many of them and they didn't feature in discussions about errors. Then Chrys started getting them in large numbers and they went up in significance.

[ejs]

I assume that the HEC (and the other ATM-related counters like OCD and LCD) have been re-used for something else with PTM, but without changing the label given to them. So they're probably counting something, but it's not clear exactly what they're counting.

HEC is not listed in BT SIN 498 under R.VDSL2.17 which lists the stats that a modem must report back to the DSLAM.

[Chrysalis]

thats what I was thinking, maybe they kept the same name to avoid confusion but it might be using another low level counter.

[ejs]

How could using the same, wrong, name for something different possibly avoid confusion?