Legal action taken against Virgin Media throttling practices - Page 19

XsCode · 13-03-2009, 18:53

Hi all

I specifically joined the forum after reading this thread in it's entirety to wish Mike_A the best of luck with this. I would rather have non-STM'd 10mb than the part-time 20mb that I seem get more and more often now due to my love of HD Podcasts! I see the word "Buffering...." in my sleep!!!

Mike_A · 13-03-2009, 18:55

Quote:

Originally Posted by BenMcr

To the best of my knowledge of the technical workings, they don't!

lol, of course they do - otherwise throttling/STM would not exist! Just wondered about their approach.

Bonglet · 13-03-2009, 18:56

So why are the allot kits sitting in head-ends? just to gather dust?.

broadbandking · 13-03-2009, 18:59

Quote:

Originally Posted by Bonglet

So why are the allot kits sitting in head-ends? just to gather dust?.

They have already said its to collect stats but as all the ALLOT equipment that was installed was way before the PHORM and as afaik is only in the NTL headends not the Telewest that has the equipement however as I have said this could change but until I see proof I can't say yes they do.

BenMcr · 13-03-2009, 19:00

Quote:

Originally Posted by Mike_A

lol, of course they do - otherwise throttling/STM would not exist! Just wondered about their approach.

AFAIK STM is a putting you onto a slower profile than you had, not purposely slowing packets/data

e.g a 1Mbit profile rather than a 2Mbit profile

broadbandking · 13-03-2009, 19:01

Quote:

Originally Posted by Mike_A

lol, of course they do - otherwise throttling/STM would not exist! Just wondered about their approach.

If you are referring to STM using the ALLOT system I don't think they need that for STM, as I am aware its just software that control STM and nothing to do with the ALLOT equipment as I said in a early post the ALLOT system are only in NTL headends I maybe wrong maybe Fatec can confirm this or broadbandings

Bonglet · 13-03-2009, 19:03

Quote:

Originally Posted by broadbandking

They have already said its to collect stats but as all the ALLOT equipment that was installed was way before the PHORM and as afaik is only in the NTL headends not the Telewest that has the equipement however as I have said this could change but until I see proof I can't say yes they do.

How can you see the proof as dpi is undetectable as a passthrough? hope you arent going to do a late night recon of a virgin media head end and take photos or anything lol.

broadbandking · 13-03-2009, 19:05

Quote:

Originally Posted by Bonglet

How can you see the proof as dpi is undetectable as a passthrough? hope you arent going to do a late night recon of a virgin media head end and take photos or anything lol.

Never know I might do lol.

I am going by information that I have been given.

Bonglet · 13-03-2009, 19:06

Can i ask who gave you such information and/or see the details?

broadbandking · 13-03-2009, 19:08

Quote:

Originally Posted by Bonglet

Can i ask who gave you such information and/or see the details?

Check your PM in a minute as I can't give info and you will see why in my pm to you

Mike_A · 13-03-2009, 19:21

Quote:

Originally Posted by XsCode

Hi all

I specifically joined the forum after reading this thread in it's entirety to wish Mike_A the best of luck with this. I would rather have non-STM'd 10mb than the part-time 20mb that I seem get more and more often now due to my love of HD Podcasts! I see the word "Buffering...." in my sleep!!!

Thank you Xs! Try thinking of cute fluffy spring lambs at bed time - you'll probably sleep better!

XsCode · 13-03-2009, 19:23

Well, I seem to be seeing plenty of "sheep" posting in this thread so trying my not be too hard! Whatever happened to the culture of british people complaining??

Mike_A · 13-03-2009, 20:09

Quote:

Originally Posted by BenMcr

AFAIK STM is a putting you onto a slower profile than you had, not purposely slowing packets/data

e.g a 1Mbit profile rather than a 2Mbit profile

I understand that's part of it, but I have been given to believe* also hardware gating logic and software which both intercepts and has the effect of intercepting packets.

Incidentally, do you know about the Canadian anti-throttling case? Runs pretty much on the same lines as here, although different legal procedures.

*The Numbskulls in my head haven't decided yet, they're still debating.

---------- Post added at 21:09 ---------- Previous post was at 20:46 ----------

Quote:

Originally Posted by Bonglet

How can you see the proof as dpi is undetectable as a passthrough? hope you arent going to do a late night recon of a virgin media head end and take photos or anything lol.

On two known servers, client and host separated by distance, send a stream of packets with a loop-back to return them. A reasonable size PDF would suffice as data. If either server log contains 307s then something is picking up packets, redirecting elsewhere and returning to continue the stream. Also, emanate a series of consistent pings from both ends. A long identical delay in a particular node across pings will display interception. If both logs, client/server/client and loop-back streaming, show the same results then you have proof of interception. To corroborate this proof, try the same client/server test with either a VPN or secure HTTPS connection: if the delaying node no longer delays you substantiate HTTP interception. You could go further with such forensic tests and analysis but these simple methods will prove the case.

BenMcr · 13-03-2009, 20:55

Quote:

Originally Posted by Mike_A

I understand that's part of it, but I have been given to believe also hardware gating logic and software which both intercepts and has the effect of intercepting packets.

Hmm. From the way I understood it - no.

Virgin have the ability to send whatever config profile they like to your modem - as the speed uplifts show.

So I always assumed that this is all they did - you hit X threshold they send Y config profile which sets your modem to Z speed. After the time period is up they then send the config profile that matches your original tier.

I'm pretty sure people have posted on here that that is the case - because you can see it in the modem pages

Again - this was just how I think it works.

---------- Post added at 21:45 ---------- Previous post was at 21:38 ----------

If anyone wants go through it and tell us - there is a full 'Subscriber Traffic Management for the Cisco CMTS Routers' manual here http://www.cisco.com/en/US/docs/ios/...sbr_tfmgt.html

But from a quick scan (and if I understand it correctly) it seems to use QoS profiles rather than any DPI

---------- Post added at 21:55 ---------- Previous post was at 21:45 ----------

In fact part of it says:

The STM feature enables service providers to identify and control subscribers who exceed the maximum bandwidth allowed under their registered quality of service (QoS) profiles. STM works as a low CPU alternative to Network-Based Application Recognition (NBAR) and access control lists (ACLs).

Ignitionnet · 13-03-2009, 21:16

Quote:

Originally Posted by BenMcr

So I always assumed that this is all they did - you hit X threshold they send Y config profile which sets your modem to Z speed. After the time period is up they then send the config profile that matches your original tier.

I'm pretty sure people have posted on here that that is the case - because you can see it in the modem pages

You can't see any changes at the modem. It's policed at the CMTS through rate limiting, nothing is changed at the modem at all just the QoS profile at the CMTS. Modems will only take new config files when they are registering, hence the stuff about rebooting to get an uplift / new service tier.

Nothing is sent to the modem it's all network side.

There is nothing 'slowed down' - the traffic follows a simple leaky bucket, traffic in excess of the profiled rate will be dropped in the same manner as normal. It's nothing more or less than a switch from 20/0.768 say to 5/0.192 so no more interception than the standard rate limiting which is at layers 1 to 3, nothing transport to application layer.