Traffic managed on downstream not upstream - Page 3

Chrysalis · 07-12-2010, 15:57

this can go some way into explaining why jitter has got worse. Previously one could get upload throttled fairly quickly by downloading too muh, so eg. a torrenter doing heavy both ways would get upload throttled quickly. Now it will just be the download and will take at least a few hours for the upload to get throttled (assuming evading the protocol shaping).

pip08456 · 07-12-2010, 16:06

And how would a torrenter evade the traffic shaping? Most VPN providers either exclude torrenting (due to bandwidth needed) or charge extra to price it out.

For a serious torrenter a seedbox is the better option which won't impact on he network until the HTTP/FTP download from it which isn't covered with the shaping.(is this what you meant?)

Chrysalis · 07-12-2010, 17:24

a few on here have openly stated they use VPN to evade shaping.

I assume on VM I could use port 443 encrypted for torrents to evade the shaping.

broadbandbug · 07-12-2010, 17:52

Quote:

Originally Posted by Ignitionnet

Personally as I'm not a customer of VM I've nothing to complain about.

Anecdotally it was a disaster and remains so. Were you involved in its' deployment? Were you the guy who went on the sick straight after it went live?

---------- Post added at 16:30 ---------- Previous post was at 16:26 ----------

Oddly I remember reading many reports from people on the overlay who were enforced in both directions, maybe they were on BSRs.

Lol.. No not me.. If you remember last time I was around I told you DPI etc was nothing to do with me.. Purely DoCSIS/CMTS.

Not 100% on how CTM on the BSRs invokes the penalty.. Will check and let you know.

pip08456 · 07-12-2010, 20:35

Quote:

Originally Posted by Chrysalis

a few on here have openly stated they use VPN to evade shaping.

I assume on VM I could use port 443 encrypted for torrents to evade the shaping.

Sorry Chrys you are confusing torrent traffic with newsgroups.

Ignitionnet · 07-12-2010, 20:58

Quote:

Originally Posted by broadbandbug

Lol.. No not me.. If you remember last time I was around I told you DPI etc was nothing to do with me.. Purely DoCSIS/CMTS.

Not 100% on how CTM on the BSRs invokes the penalty.. Will check and let you know.

Thanks.

Maybe you can answer one question, why 6.4MHz channels, why not bond 3.2s? Laser load?

pip08456 · 07-12-2010, 21:04

Quote:

Originally Posted by Ignitionnet

Thanks.

Maybe you can answer one question, why 6.4MHz channels, why not bond 3.2s? Laser load?

Are you guys getting all techie on us?

Rik · 07-12-2010, 23:19

Quote:

Originally Posted by pip08456

Are you guys getting all techie on us?

And why not use 12.8Mhz Flux Capacitors instead?

pip08456 · 07-12-2010, 23:44

Ooooooooooooo Flux capacitors!!!!!!!!!!!!!!!1

I'm feeling randy now!!!!!!!!!!!!

Chrysalis · 08-12-2010, 07:14

pip true the VPN's are mostly used for newsgroups but p2p shaping can still be evaded.

adzii_nufc · 08-12-2010, 09:33

To answer the OP: I have a lot of friends on the 20meg service and asked around a little during management times and what not.

It appears that upload speeds are not affected. Wtih two out of three asked reporting 5mb down and 2mb up whilst the other had 6 down and 2 up.

JohnnyEnglish · 08-12-2010, 13:25

O.o

broadbandbug · 08-12-2010, 18:24

Quote:

Originally Posted by Ignitionnet

Thanks.

Maybe you can answer one question, why 6.4MHz channels, why not bond 3.2s? Laser load?

VM havent until very recently had the CMTS hardware to do Upstream Channel Bonding.. Still don't on Motorola.
Will be coming along over the coming months.
So only way to maximise upstream in the short term was 6.4Mhz 16QAM

Ignitionnet · 08-12-2010, 20:00

Quote:

Originally Posted by broadbandbug

VM havent until very recently had the CMTS hardware to do Upstream Channel Bonding.. Still don't on Motorola.
Will be coming along over the coming months.
So only way to maximise upstream in the short term was 6.4Mhz 16QAM

I saw second half of 2011 targetted for upstream bonding on the BSR and not on the 2/8 cards that VM have used thusfar. I guess a move to the RX48 teamed with the TX32 is on the 'to do' list.

From this I take it it is the 10k being used on the 200/20+ trials.

Getting any MC3GX60Vs in to play with?

broadbandbug · 09-12-2010, 18:48

Already using TX32 for downstream, but yes RX48 for upstream channel bonding.

The 200/20 is using CMTS from A.N.Other Vendor ;-)

Got 3G60s going into test.