Kitz ADSL Broadband Information
adsl spacer  
Support this site
Home Broadband ISPs Tech Routers Wiki Forum
 
     
   Compare ISP   Rate your ISP
   Glossary   Glossary
 
Please login or register.

Login with username, password and session length
Advanced search  

News:

Author Topic: Useful figures derived from stats  (Read 1215 times)

Weaver

  • Senior Kitizen
  • ******
  • Posts: 11459
  • Retd s/w dev; A&A; 4x7km ADSL2 lines; Firebrick
Useful figures derived from stats
« on: March 27, 2019, 11:37:32 AM »

Sincerest apol for the intentionally vague and ill-formed meanderings in the following.  :) ???

My question: what figures might one usefully calculate based on the stats that can be had from your modem ? I’m thinking about useful calculated numbers that one might invent that would give various kinds of litmus tests about the health-state and performance of a line, in particular

(a) derived values that are functions of more than just one of the well-known stats, and
(b) harvested/logged time-series data based on one or several stats.

Some suggested derived values, in no particular order, random unconnected brain-dump:

(i) The first thing I have been wondering about is the significance if any of the values
    p1D = crcD / fecD and p1U = crcU / fecU for downstream and for upstream respectively, where each has been counted over some time window.

(ii) Also the ratio f1 = p1U / p1D .

(iii) then there is something derived from the G.992.3 framing parameters -
    ρ = ( R / ( M * ( B + 1 ) - 1 ) ) which I am hoping is a measure of level of FEC-based error protection but without allowing for the beneficial effects of interleaving, L2 retx, and any other mechanisms.

(iv) taking the values of ρ in (iii) for upstream and for downstream, I would be interested to know about the value for the ratio of the upstream vs downstream values

(v) distance between max and min SNRM values that you reach during a day, for both upstream and downstream.

I would be interested in:
  • more suggestions for candidates,
  • (kindly) comments on the worth/meaning/interpretation of the above,
  • kitizens’ own values for the above. These can be obtained by either looking at the web admin UI of your modem, or from one of the useful tools for querying stats, or telnet plus the usual incantation - xdslctl info --show would be the magic runes for my ZyXEL modem, for example.

I am going to work out these numbers for all my own lines and see firstly if there is a lot of variation between the lines. If there is, in some cases it will mean that the measure is invalidated, but in others it can highlight a valid line characteristic where I have a line that is in some way, by some criterion, considered sickly.

I also want to look at upstream vs downstream to try and work out if anything worthwhile can be said.
« Last Edit: March 27, 2019, 11:51:13 AM by Weaver »
Logged

DiggerOfHoles

  • Member
  • **
  • Posts: 64
Re: Useful figures derived from stats
« Reply #1 on: May 10, 2019, 06:22:08 PM »

(iii) then there is something derived from the G.992.3 framing parameters -
    ρ = ( R / ( M * ( B + 1 ) - 1 ) ) which I am hoping is a measure of level of FEC-based error protection but without allowing for the beneficial effects of interleaving, L2 retx, and any other mechanisms.

https://www.itu.int/rec/T-REC-G.Sup22-198410-I

Is this where you got undefined variables R,M,B?
Logged