 mbind
join:2008-03-30
| reply to An Onymous Re: Illustrating the network and throttling for non-tech people
Sorry, doesn't make too much sense what you write:
(i) if you want upstream control then you bring in a device like a router as soon as possible. This would be in-line with Tobias' comment that they have BRAS in all COs.
(ii) if Bell really has a congested network then you don't throw away traffic at the exit of your network - then it's too late and pretty useless as packets made their way through already. You throw away as soon as possible
(iii) not all P2P is TCP, e.g. newer version use UDP as well
(iv) if Bell can handle traffic or not depends on many factors, including geographical distribution. Not a problem of the pipe to TSI and others.
(v) I don't mint if the ISP "can" traffic shape -Bell would not control the game but depend on the ISP. Very likely ... .
Regards, Marc |
|
  Arbalister
join:2007-11-24 St Catharines, ON
·TekSavvy Solutions..
| reply to TobiasFunke said by TobiasFunke :One thing I'm still not entirely clear on is whether TekSavvy pays Bell for transit over their core network. I wish someone from TSI could shed some light on this. Did you forget we're talking about Bell? We, the common man, send one ping of data across anything they can bill for, they do. 
Yes, TSI pays for transit over the core. Bell's DSL wholesale arrangement is that the indie ISP pays all the costs involved with getting a packet from the enduser premises to a specified point of interconnect. That's 151 front. Once it gets to the TSI "premises" they can do whatever they want with it - which in this case is fire it to the net via either cogent, peer1, or whomever.
Point is...when I buy a DSL circuit from Bell, I'm paying per customer to get a direct link through bell's internal net, from the customer to my servers. And I'm paying a big per month fee for the backhaul. |
|
 TobiasFunke Premium join:2007-02-27 Toronto, ON
1 edit | reply to TobiasFunke The throttling is being done at the BRAS, which is more or less at the subscriber entry point into the Bell core network. This obviously makes it easy to throttle upstream, but it's not all that much harder to throttle downstream - you don't need DPI devices at the peering routers. All that's necessary is to delay TCP ACKs from the DSL user (i.e., at the BRAS in the CO) and the sender - wherever it may be - will correspondingly delay sending more packets until it receives ACKs for what it already transmitted.
If you've ever saturated your upstream data rate, you'll know exactly how effective this can be at limiting downstream data rates. |
|
 recneps
join:2006-06-24 Whitby, ON
| reply to TobiasFunke Not to offend anyone in the general public, but if you want someone to understand the message clearly (like if you submitted this to a newspaper) "dumb it down". 90% of the population (made up number) has no idea how this magical 'series of tubes' we call 'the internet' works. Make it as simple and to the point as possible, and people may understand the issue.
1) As a customer, with TS, you are using the phone line to your house, owned by Bell. TS pays to lease this line. 2) your data then moves over the phone line and bell's DSL equipment, where the signal is transferred to TS through a high-bandwidth link, paid for by TS based on a set bandwidth amount. 3) Bell has recently throttled this link, essentially not providing the service TS has paid for. 4) From there, TS has their own links to the internet, separate from Bell Sympatico customers. |
|
  madavid
@bell.ca | reply to TobiasFunke I don't understand why many of you feel that this is a difficult subject to explain to the average person. The situation can be summed up simply with this:
Bell is stealing from its customers.
Questions? |
|
 TobiasFunke Premium join:2007-02-27 Toronto, ON
1 edit | reply to TobiasFunke I think this is a truer picture of what's going on and it avoids the whole Sympatico issue altogether, focusing instead on what Bell is doing that is contrary to the letter and/or spirit of the GAS tariff (and net neutrality more generally). |
|
 puzz1ed
join:2005-02-20 Markham, ON
1 edit | The best one yet.
What strikes me is that the tunnel must be so insignificant in the grand scheme of things. Sympatico has in the neighbourhood of 2 million high speed subscribers according to the latest annual report. TekSavvy claims about 50,000 all told and CAIP claims to represent 100,000. Bell's core network would have capacity for all the other things it's being used for as well. It'll be interesting to see some hard numbers from Bell to support their claim of congestion. |
|
  Comment
@teksavvy.com
| reply to TobiasFunke said by TobiasFunke :I think this is a truer picture of what's going on and it avoids the whole Sympatico issue altogether, focusing instead on what Bell is doing that is contrary to the letter and/or spirit of the GAS tariff (and net neutrality more generally). Two thoughts:
Bell will DENY that they are inspecting the CONTENT. (They'll interpret "CONTENT" as the clear data.) Better to describe it as "opening your message" (then Bell can try to explain that they're looking at the shape of the letter and the ink that is used, not the words).
You're buying into Bell's "framing" of the situation. They're not "inspecting", they're WIRE-TAPPING WITHOUT A WARRANT. P2P itself is no more "abusive" than HTTP. Better to say "any messages Bell doesn't like". After "protect its core network" put in brackets "(increase profitability)" and put quotes around "protect". |
|
 NiGHTS
join:2008-03-23 | reply to TobiasFunke I liked the old pictures more. Bell Nexxia looked smaller, as it should. |
|
 TobiasFunke Premium join:2007-02-27 Toronto, ON
| reply to TobiasFunke Some minor updates to incorporate some of the comments. |
|
 dmz
join:2006-07-12 canada | Might be worth mentioning that TekSavvy still has to pay a lot of money to put the user ON the internet. By this diagram alone, one might assume that part is free. |
|
  Maynard G Krebs
@teksavvy.com
| reply to TobiasFunke
Great diagrams but I have two small points to make.
According to the GAS tariff, Bell is supposed to supply an unthrottled link from the residential/business user to Teksavvy's 151 Front St. presence (usually numbered as point '4' in your diagrams).
So when you do a before/after diagram I think you need to add the following words on the 'Before' diagram:
'This connection is unthrottled and not inspected as per CRTC Tariff #xxxxx
On the 'After' diagram our need to add text which conveys that fact that Bell is acting illegally, eg. "In violation of CRTC tariff #xxx, Bell is reducing the speed of the connections by up to 90% of the tariffed speed required and illegally reading data between points 1 and 4." |
|
 TobiasFunke Premium join:2007-02-27 Toronto, ON
| Trouble is, the GAS tariff never says the connection must be "unthrottled" or "not inspected". Of course, I think any reasonable interpretation would require that it be read that way, but ultimately it's a question of how the CRTC and the parties will interpret it.
My goal with this diagram is to present the issue without overstating the facts. This is why I backed away from some of the weaker arguments I made earlier that were based on speculation (e.g., that Bell's move is an anti-competitive attempt to shore up Sympatico). It's also why I'm reluctant to downplay Bell Nexxia's role too much -- the fact that Nexxia provides such a large chunk of TSI's infrastructure is the reason we have this mess in the first place.
Obviously, everyone can draw their own conclusions, but I don't want to make any obviously false or simply overzealous claims. |
|
 recneps
join:2006-06-24 Whitby, ON
| said by TobiasFunke :I don't want to make any obviously false or simply overzealous claims. Why not? Bell will..  |
|
  Stewy Premium join:2007-12-12 Kitchener, ON
| Help me understand this.
On point number 3, who is judging "undesirable" content?
On point number 4, if you were to replace all P2P with the exact same amount of HTTP traffic. Would that have the exact same impact on "Bell's core network" as they put it.
In other words is P2P the same to manage as an HTTP packet?
It's as if Bell is saying that yes we can lease you virtual tunnels but you can only use them at %10 capacity in order for our "core network" to run. |
|
  Pegasys66
join:2008-01-16 | reply to TobiasFunke »www.bcenexxia.ca/wholesale/produ···gate.htm
No comment required I think |
|
  R0CKY TSI Rocky Premium,VIP join:2005-05-19 Chatham, ON | LOL... interesting second line under "You will get the following benefits" -- TSI Rocky - TekSavvy Solutions Inc. |
|
  Guspaz Guspaz Premium,MVM join:2001-11-05 Montreal, QC | Whatever else happens with the CRTC, you can at the very least nail them for false advertising :P |
|
  theninjasqua
join:2007-09-26 Oakville, ON
| reply to Pegasys66 Someone should take some screen caps of that for safe keeping before someone from Bell reads this and goes and changes the site. --
-theninjasquad |
|
  Comment
@teksavvy.com
| reply to Stewy said by Stewy :Help me understand this. BELL has arrogated unto itself the determination of "undesirable". But when you speak of "content", it is not the end payload, (the actual bytes of the file or page), but the 'format' of the data within the packet. Worse, it seems that Bell is operating on a 'negative' test - if the 'format' is NOT one they recognize as "valid" (HTTP, FTP etc), then they "throttle" on the assumption that it is "undesirable". This causes all sorts of "collateral damage".
A packet is a packet is a packet. P2P should really be called MP2MP (Multiple Peers to Multiple Peers) in that the various 'formats' support a DISTRIBUTED data base where each peer is sending to many peers (if 'keeping their end of the bargain') at least as much data as they are receiving from many peers who are not necessarily the same. This conflicts with the nature of ADSL (the "A" being "asymmetric" in that the ratio of download capacity to upload capacity is not 1:1). Because the "tracking" requirements of a DISTRIBUTED data base will entail a slight additional overhead in comparison to a CENTRALIZED data base (such as Usenet where one peer is sending to many using NNTP), the total traffic will be a bit higher, but this is a factor of the data base type, not the 'format' used to transmit the data. |
|