dslreports logo
site
 
    All Forums Hot Topics Gallery
spc

spacer




how-to block ads


Search Topic:
uniqs
8834
share rss forum feed

Turbocpe
Premium
join:2001-12-22
IA

Far End/Near End Errors & Downstream/Upstream FAQ Wrong?

In the Qwest FAQ it states the following:
quote:
Errors- near end are those recieved at the modem, far end are those reported by the DSLAM.
Would that not mean, according to the above, that Near End is "Downstream" while Far End is "Upstream"?

I have two Qwest circuits and using the GT701 internal signal webpage, and then using telnet into the GT701, I see the following




On this circuit, the "Downstream" in PuTTY (telnet) shows 2590 FEC errors while the signal webpage shows 2590 "Far End" errors.




On this circuit, the "Downstream" in PuTTY (telnet) shows 2243 FEC errors while the signal webpage shows 2243 "Far End" errors.

Does that not contradict the FAQ? Or is this some bug with the GT701 stats? I would think "Far End" would refer to Upstream, since the FAQ states those are reported at the DSLAM.


AthlGrond
Premium,MVM
join:2002-04-25
Aurora, CO
Why aren't near end and upstream numbers aligned as well then?
--
Embrace diversity with enforced conformity.

Turbocpe
Premium
join:2001-12-22
IA

1 edit
I was curious about the myself, but figured it was a whole other topic. But you can see the Far End FEC and Downstream FEC figures match in both cases, so that appears to be a winner. The CRC errors match in all cases.

I'm still curious if the FAQ is wrong, or if the GT701 is wrong.

said by AthlGrond:

Why aren't near end and upstream numbers aligned as well then?


AthlGrond
Premium,MVM
join:2002-04-25
Aurora, CO
Not sure who's right, a brief dig through the interwebs isn't turning up much that could be considered definitive. But if the reporting is from the modem's perspective, it seems like near would be something done at the modem, and far would be something done at the DSLAM.

BTW what makes you think that PuTTY can't be wrong? (Seems like a logical 3rd option.)
--
Embrace diversity with enforced conformity.

dynodb
Premium,VIP
join:2004-04-21
Minneapolis, MN
reply to Turbocpe
No idea- firmware issue?

Far End errors always refer to errors recieved by the "far end" device- in this case the DSLAM; same with any other kind of circuit be it a T1, OC3, etc. Don't know why the modem telnet screen has them labelled that way.

Turbocpe
Premium
join:2001-12-22
IA
reply to AthlGrond
said by AthlGrond:

BTW what makes you think that PuTTY can't be wrong? (Seems like a logical 3rd option.)
It's not an issue involving PuTTY itself. PuTTY is simply the connection to the GT701 via telnet. So what you see in PuTTY is what the GT701 is outputting via telnet.

bobk4000

join:2003-10-17
Saint Paul, MN
kudos:1
Reviews:
·CenturyLink
reply to Turbocpe
I've been confused about the whole Near End / Far End values on my Actiontec, and prior to that my Cisco 678 with it's Local/Remote values. I did find the following in an Actiontec 1524 User Manual years ago:

"Near End CRC Errors (I/F), Far End CRC Errors (I/F) - Displays the number of interleaved (I) and fast path (F) cell errors occurring on the DSL line. Near end errors occur on the DSL network side; far end errors originate from the DSL modem. These numbers can be used to diagnose network problems, such as slow response times."
As I read this, the Near End / Far End values are from the perspective of the DSLAM / Central Office - ie: Near End is what the DSLAM is seeing, and Far End is at the DSL Modem / Customer Premise Equipment.

Not sure if this helps, or adds to the confusion about this...

Bob

Turbocpe
Premium
join:2001-12-22
IA

1 edit
said by bobk4000:

I've been confused about the whole Near End / Far End values on my Actiontec, and prior to that my Cisco 678 with it's Local/Remote values. I did find the following in an Actiontec 1524 User Manual years ago:

"Near End CRC Errors (I/F), Far End CRC Errors (I/F) - Displays the number of interleaved (I) and fast path (F) cell errors occurring on the DSL line. Near end errors occur on the DSL network side; far end errors originate from the DSL modem. These numbers can be used to diagnose network problems, such as slow response times."
As I read this, the Near End / Far End values are from the perspective of the DSLAM / Central Office - ie: Near End is what the DSLAM is seeing, and Far End is at the DSL Modem / Customer Premise Equipment.

Not sure if this helps, or adds to the confusion about this...

Bob
And that does make it confusing because what you describe seems to jive with what I'm seeing. I currently have 3 circuits here, two of them have an upstream SNR of 6-9dB and max attenuation. They do not train at the proper upstream train rate. One is my neighbor's DSL modem and the other is my DSL modem. I notice that both of our circuits shows an alarming amount of errors on the Near End side. The Downstream SNR of both circuits is 14-23dB with a 55-56dB attenuation. So it would seem the downstream looks better than the upstream, according to attenuation and SNR.

So, if the FAQ was right, why would those two circuits, with a "clean" downstream, show such alarming Near End errors if Near End was downstream errors? It would appear to me those are Upstream errors, due to the "poor" upstream train rate and SNR.

The 3rd circuit has a 18dB upstream because it trains at a 640kbps rate (set at the DSLAM) and that circuit has 0 (zero for days) errors on the Near End side.

This is why, to me, it appears the FAQ is backwards. I suppose that isn't very conclusive or scientific basing it on 3 circuits.

roozy

join:2004-09-30
Casper, WY
reply to Turbocpe
far end would correspond to downstream and near end to upstream.

Why?

FE errors are reported by the DSLAM. How does the DSLAM know to report an error? When the modem does not acknowledge receipt of a downstream packet. NE errors are reported by the modem. How does the modem know to report an error? When the DSLAM does not acknowledge receipt of an upstream packet.