e94mli
join:2009-04-07
 ·Acanac
 ·Videotron
| PPPoE session from Montreal terminating in Toronto???
My PPPoE seesion seems to be terminating in Toronto, can that be correct or is it just a fluke?
It adds even more delay on an already high delay connection. I'm now at over 100ms at gaming servers which I used to have about 40ms with Videotron. Getting to my ipv6 tunnel broker, which is almost next door in Tata communications backbone, is also close to 100ms.
I've had some instability issues with the PPPoE session, perhaps that is related to ending up in Toronto. Any sugestions? |
|
Acanac PN
Acanac Support
Premium
join:2008-10-27
Montreal, QC | Please send me your Acanac ID info, I will check what could be done. |
|
Acanac PN
Acanac Support
Premium
join:2008-10-27
Montreal, QC
1 edit | reply to e94mli
Hi,
I checked your file. You are currently on interleaved profile - more stable with error corrections but with higher ping responses.
Generally, IF the profile was set to AL1 (Fast Path) the pings would be lower.
Bell adjusts these DSL profiles to provide stable connections, most of the time adequately, but sometimes there is a room for improvements.
However, after some recent experience, I found many new things regarding this Fast Path stuff.
They are just for your technical information, not officially stated by Acanac Inc by Bell or any other company
o The request switching to fast path (Al1) once created by support people is not always further processed or even answered (by Bell). Simple reason is - according to some people, customer has already stable and working "up to 5M" service. Why to change the things if they work ? This is a ticketing system and we are talking about "Troubles".
o Wholesaler DSL ISPs could forward (like trouble tickets), but not force such changes to be made for standard home accounts unless some particular rules they have established with copper pairs network owner (bell)
o Currently, you are, say on Interleaved (al2) 6M-download setting profile. It may be reduced to 5M at any time, especially when Fast Path request is made
o Al2 to Al1 means Noise margins go lower, occ% go higher, The upstream could start caping. Line is more vulnerable and some disconnects may occur.
o Of course, your Internet would be more "responsive" and "good feeling" if you are on Fast Path. But when it starts disconnecting (some cases I've seen), then whole request needs to go backwards and the customer ends with even lower settings than initial ones.
o There are many so called "IKNS" remotes (or something like this) that give strange stats (always 6db noise margin US) and force Interleave connection with some modems. I don't know why that happens, maybe they are optimized for ADSL2+ (which is Interleave by default) ? Anyway, they are tricky. Unfortunately, only few people know what I am talking about and even fewer could adjust optimal profile settings vs. modem selection in such configurations.
Other ping-related stuff:
Longer line distance to CO = attenuation and ping
Different modems, routers, PC, software could give different ping responses.
PPPoE and internal network routing structure could give you some ping difference but usually is not so big.
In Montreal, I've seen many different average DSL pings.
26ms or 150ms - to same server NY.
A 26ms @ Acanac DSL line tested with different PPPoE logins (Vif, Sympatico, Teksavvy, Acanac) shows almost the same typical pings. Slower line may show some difference, but it is less likely to have 50ms vs 150ms only by changing PPPoE.
You could tell me if you want this Fast Path challenge - you may go down from 6M profile to 5M as a side effect.
Currently, your DSL stats look like this: Al2 6000/512 att 3/9db
I am afraid, if you are on IKNS server it may be difficult to improve your service.
Please, send me another message or just write a request to support@acanac.com with Fast Path note. Then, send me the ticket number, I will follow up.
Thanks |
|
e94mli
join:2009-04-07
·Acanac
·Videotron
| If I understand you correctly it is normal for the session to terminate in Toronto. I guess it doesn't make a huge difference when interleaving is active, at least if it has high interleaving depth, but it makes a difference. A return trip Montreal-Toronto should be about 10 ms and adsl with interleaving adds about 40ms. In the case of accessing something in locally in Montreal the Toronto trip makes a difference since it might add up to 20ms, but it doesn't make up for a 100ms difference as you say.
Going fast path versus interleaving is always a trade off. Also the interleaving depth is something to consider. For a normal Internet access pack errors aren't that critical and it can actually benefit in terms of speed even with a higher error rate since a lower delay might go easier on tcp. For a multicasted tv service packet errors are a disaster and a high interleaving is therefore a must. So what I'm saying is that in my case fast path would be worth the risk. If you don't mind changing it I would want to give it a try.
Except for the somewhat high delay the connections has been working perfectly after the Bell "miss configuration" was sorted out. Thanks for the hard work of keeping us happy.
Mikael |
|
Acanac PN
Acanac Support
Premium
join:2008-10-27
Montreal, QC
| BTW, I saw on one new activation in Montreal - friend of mine who is Acanac user, the technician logged with his PPPoE and the ping to NY was 28ms, Then he did the same test with customer's PPPoE (Acanac) - it showed exactly 28ms, so I dont believe the PPPoE itself makes difference.
Let me create a fast path query for you.
Note - the upstream shows 512 not 800 but you are very close to CO. It could be an IKNS one.
Thanks |
|
Acanac PN
Acanac Support
Premium
join:2008-10-27
Montreal, QC | reply to e94mli
Request made, the ticket number is XXX-73624 |
|
e94mli
join:2009-04-07
·Acanac
·Videotron
| The profile has been changed. Delay has been reduced with about 30ms. First hop ping is now a much nicer 20ms instead of 50ms.
Downlink was reduced, as you suggested, measured is almost down 1Mb/s to 4.5Mb/s. I must say I'm a bit surprised I didn't expect that much of an impact. More surprising is the uplink which increased quite significantly. Measured uplink was before only about 350kb/s and now it is up to 600kb/s. Overall a good change. |
|
Acanac PN
Acanac Support
Premium
join:2008-10-27
Montreal, QC
1 edit | Yes, I see the AL2 to AL1 change
My DSL is similar. Upstream could not sync at 800 but keeps stable 700 regardless the fact that noise margin is displayed too low. In your case, it is 5dB now.
Downstream is decreased from 6 to 5M.
I am glad you feel these changes match better for your needs and you are happy with them.
BTW what model is your modem?
Thanks.
--
We Rock |
|
e94mli
join:2009-04-07 | The modem is the one that came with the service Aztech 605EU.
Sounds like a DSLAM - modem incompatibility that gives the strange readings. |
|
Acanac PN
Acanac Support
Premium
join:2008-10-27
Montreal, QC
| Your modem is good, actually.
I tried many other modems on same scenario(My DSL) and some (old ones) don't work at all or hardly sync at 1-2M
The Bell technician's yellow diagnostics tool didn't sync at all, either.
Otherwise, my and your DSL is considered as good ones 
And they are!
Thanks. |
|
planiwa
join:2009-02-19
Toronto M5S
| reply to e94mli
said by e94mli  :The modem is the one that came with the service Aztech 605EU. Sounds like a DSLAM - modem incompatibility that gives the strange readings. Could you please post your modem's line stats?
In fact, since you sound like you are able to do it, could you please post the more complete stats:
telnet, then:
cat /proc/*/*m_s*
Thanks! |
|
e94mli
join:2009-04-07
·Acanac
·Videotron
| Here you go:
AR7 DSL Modem Statistics:
--------------------------------
[DSL Modem Stats]
US Connection Rate: 736 DS Connection Rate: 5056
DS Line Attenuation: 9 DS Margin: 24
US Line Attenuation: 20 US Margin: 6
US Payload : 876087840 DS Payload: 2706042272
US Superframe Cnt : 34897832 DS Superframe Cnt: 34897832
US Transmit Power : 12 DS Transmit Power: 3
LOS errors: 0 SEF errors: 0
Errored Seconds: 0 Severely Err Secs: 0
Frame mode: 3 Max Frame mode: 0
Trained Path: 1 US Peak Cell Rate: 1735
Trained Mode: 2 Selected Mode: 65535
ATUC Vendor Code: 494B4E53 ATUC Revision: 2
Hybrid Selected: 3 Trellis: 1
Showtime Count: 2 DS Max Attainable Bit Rate: 11296 kbps
BitSwap: 1 US Max Attainable Bit Rate: n/a
Annex: AnxA psd_mask_qualifier: 0x0000
Power Management Status: L0 DS HLINSC: 0
US ACTPSD: -345 DS ACTPSD: -425
Total init. errors: 0 Total init. timeouts: 0
Showtime init. errors: 0 Showtime init. timeouts: 0
Last showtime init. errors: 0 Last showtime init. timeouts: 0
ATUC ghsVid: b5 00 49 4b 4e 53 01 00
T1413Vid: 00 00 T1413Rev: 00 VendorRev: 00
ATUR ghsVid: b5 00 54 53 54 43 00 00
T1413Vid: 00 00 T1413Rev: 00 VendorRev: 00
[Upstream (TX) Interleave path]
CRC: 2 FEC: 3746 NCD: 0
LCD: 0 HEC: 68
[Downstream (RX) Interleave path]
CRC: 0 FEC: 0 NCD: 0
LCD: 0 HEC: 0
[Upstream (TX) Fast path]
CRC: 0 FEC: 0 NCD: 1
LCD: 0 HEC: 0
[Downstream (RX) Fast path]
CRC: 0 FEC: 0 NCD: 0
LCD: 0 HEC: 0
[ATM Stats]
[Upstream/TX]
Good Cell Cnt: 18251830
Idle Cell Cnt: 1011563439
Tx Packets Dropped Count: 0
Tx Bad Packets Count: 24
[Downstream/RX)]
Good Cell Cnt: 145854366
Idle Cell Cnt: 2633561468
Bad Hec Cell Cnt: 0
Overflow Dropped Cell Cnt: 0
Rx Packets Dropped Count: 0
Rx Bad Packets Count: 0
[SAR AAL5 Stats]
Tx PDU's: 8473300
Rx PDU's: 10840400
Tx Total Bytes: 1237212299
Rx Total Bytes: 3087032143
Tx Total Error Counts: 0
Rx Total Error Counts: 0
[OAM Stats]
Near End F5 Loop Back Count: 0
Near End F4 Loop Back Count: 0
Far End F5 Loop Back Count: 0
Far End F4 Loop Back Count: 0
SAR OAM Ping Response Drop Count=15 |
|
planiwa
join:2009-02-19
Toronto M5S
| IKNS identified -- PPPoE session from MTL terminating in To?
said by e94mli : US Connection Rate: 736 DS Connection Rate: 5056 DS Line Attenuation: 9 DS Margin: 24 US Line Attenuation: 20 US Margin: 6 US Transmit Power : 12 DS Transmit Power: 3 ATUC Vendor Code: 494B4E53 ATUC Revision: 2 Showtime Count: 2 DS Max Attainable Bit Rate: 11296 kbps ATUC ghsVid: b5 00 49 4b 4e 53 01 00 [Upstream (TX) Interleave path] CRC: 2 FEC: 3746 NCD: 0 LCD: 0 HEC: 68 YES!!!!!
494B4E53
You are the first to confirm what looks like a way for Aztech modem users to identity IKNS DSLAM devices.
Thank you! |
|
Acanac PN
Acanac Support
Premium
join:2008-10-27
Montreal, QC | clarification
49=I
4B=K
4E=N
53=S
494B4E53=IKNS |
|
e94mli
join:2009-04-07
·Acanac
·Videotron
| And now explain IKNS... I've seen IKNS mentioned in the forum a couple of times but I don't know of any DSLAM called IKNS, although I only know Alcatel and Ericsson so I might be lost. |
|
Acanac PN
Acanac Support
Premium
join:2008-10-27
Montreal, QC
| It should be some abbreviation but the meaning is type (brand) of CO equipment used in DSL . I think, it is Canadian made
For some reason, it behaves different than classic ones and creates some support troubles, as you see.
I am trying to find any official information on that as many people do.
On Bell site I've seen discussions where they claim that all IKNS-reported problems were minor and fixable by other means (something like this).
In two words:
The user is very close to CO but upstream is not synchronizing at full 800 and/or NM is shown very low, say 6dB.
Some strange interleave modem reporting on AL1 profiles and/or higher pings.
Unfortunately, 70% of cases I open recently seem to be in this category.
Best regards.
|
|
