Telecom Trends

Respectfully, I think you are mischaracterizing those threads. The issue is not necessarily one of whether Bell is or is not adhering to its tariffs. Nor is it even about whether network neutrality and traffic shaping are good ideas. It is about whether, whatever Bell chooses to implement on the retail side, there is or is not and, separately, whether there should be or should not, the opportunity for wholesalers to purchase some product from Bell that would let them begin offering service before ramping up to build mode, yet make that decision for themselves.

That makes the issue threefold: (1) just what the tariffs cover and don't cover, (2) what the CRTC's goals were in approving those tariffs in their existing forms, and (3) whether the tariffs meet those goals.

For instance, it appears clear -- from the last thread, at least -- that Bell does not make available any layer 2 service, only layer 3 service (encapsulated in IP), which is what renders IP-level traffic management possible.

I have very little doubt that that is perfectly permissible under the relevant tariff items themselves, even without the omnibus item 10 you mention.

The salient question, though, is whether that state of affairs is within the range of outcomes envisioned by the CRTC. The problem is that answering the relevant questions that feed into that requires far more experience with CRTC tariffs than most people have. Still, on the off chance that somebody reads this who is both knowledgeable enough to say and open enough to tell, what I suspect needs answering are questions like these two:

- Is it really true that you can't get a data-link layer DSL+backhaul service from Bell to bring traffic to your switch and router in a, say, Toronto-area CO? that's it got to be IP? And if you could do it, what's the tariff?

- For 5410 and 5420, both of which appear to allow Bell to encapsulate in IP before handing off the traffic: did the CRTC explicitly allow layer-three encapsulation? Or fail to require that the tariff specify the protocol layer because nobody raised it as evidence? Or fail to require it because, althuogh somebody raised it, the CRTC ignored them, or disagreed with them (and on what grounds)?

Those are the unknowns whose resolution would, I think, help move the debate forward in an informed way. Unfortunately, those who know haven't (yet?) said, and those who don't are unlikely to be able to afford the very significant time investment it would take to learn all this stuff.

Or so it appears to the uninformed onlooker, at any rate.
 

That expertise is affordable to an ISP with a business that depends on it. Maybe it's time for them to grow up.
 

Well, I'm a tech, not an ISP. Like many, I am following this not because I am in the business (I do enterprise LANs, and that as a pure employee) -- but because I think the debate is interesting and important.

The reason one side of this debate is heard so much more loudly than the other side is precisely because of the attitude of the previous poster. Those who actually know something and have something to contribute, won't. So the Geists of the world dominate the debate instead, even when they don't really understand what is going on inside the network.

So it goes, I guess. But it's a bit of a shame.
 

To be fair to the 1:25pm poster, who is likely one of my colleagues in consulting, there is some intellectual property about how to design a network that has value.

Suffice it to say that there are a wide number of ways to develop a network, resold, hybrid and facilities based; subject to network management and not.

As an aside - For easier reference, one of the commenting options is to create a 'Name' - not subject to verification. It would be easier for responding than 'the 1:25pm guy'. Can you anonymous folks give that a try?
 

So following the same logic, if Bell determined it's inter-switch trunk capacity was being over utilized, it could decide to implement technology to determine the nature of conversations, and, if they were deemed to be of what they consider to be a frivolous nature, such as say, chatter about the latest American Idol episode, limit the calls to 15 minutes at a time?

That might be be within the strict interpretation of the tariff, but certainly wouldn't be tolerated.

Bell is deploying its throttling technology across the entire DSL aggregation network without regard to the existance of a specific capacity problem in any one particular segment.

I'm also not sure how Bell's use of encapsulating Layer 2 packets within a Layer 3 network makes it OK for them to pick apart the packets to determine the nature of the customer data, and prioritize it - particularly when the product that Bell sells to ISPs is in fact a Layer 2 transport protocol. The ISP is free to carry whatever Layer 3 protocol they desire within the L2TP tunnels. It does not have to be IP. It could be IPX/SPX, DECNET, or LAT traffic. Bell is ripping the L3 packet out of their L2TP tunnel, and making decisions on how to handle it.

Would this be permissable on a Frame Relay network? Would it suddenly become permissable if Bell were to encapsulate Frame Relay on an MPLS pseudowire connection - because now it is carried within a Layer 3 IP network?
 

i think clayton has a great analogy and i think mr.goldberg, you are failing to see the seriousness of this.
And i am pretty sure if bell would continue to expand their trunks and overall throughput instead of "limiting the american idol chatter to 15 minutes at a time".
Why would they do this? not because they want to spend money (as it cost more to expand than to police) but because public outcry would force them to do this (and CRTC guideliens probably wouldnt let them do this either).
Being as this is such new technology, bell believes it can by pass public outcry as "average joe" wont miss what they never thought they had, and CRTC guidelines do not see to have taken into full account of what potential issues would arise.
 

To Mr 4:43 anonymous:

I do understand the seriousness of this issue. That is why I have already written 2 blog posts on the subject.

I'm not crazy about analogies, because there are usually sufficient differences, even subtle distinctions, that are significant enough to make them inappropriate. But let's look at your example of the long distance telephone network - a subject about which I have more than a passing knowledge.

In fact, carriers ar unable to handle all of the traffic that is offered to them. That is why you sometimes get fast busy signals, during busy times of day, busy times of year, or when calling to certain locations. Further, there are reserved channels for operator traffic; corporations can pay for private or reserved capacity; some countries have reserved channels for priority government calling.

Mass calling applications (such as radio station contests) are segregated and traffic is routinely throttled in order to help ensure that the lower priority calls do not interfere with regular traffic. (The 26th anniversary of Larry the Lobster is coming up in 2 weeks!)

This doesn't require listening to the content of the calls - simply understanding the 'class' of the calls. Applying network management to peer-to-peer file transfer doesn't mean the carrier looks at a CBC originated torrent versus pirated software or movies. It is applying management controls to a class of traffic that is generally more tolerant of delays during peak load conditions, in order to allow more delay-sensitive traffic to use the available capacity.
 

In the old days of radio station contest lines, Bell used to have a separate switch within the rate centre set asside for mass calling events. Basically this switch had limited trunking to each of the other switches within the calling area, thereby providing a "choke" point to limit the volume of traffic that leaves each switch. This is how they did it prior to Common Channel Signalling. Mass calling events are no longer handled this way. With newer technology, inter switch trunks are no longer used for blocked call attempts, and the traffic is quenched at the source. If the radio station only has 10 lines, then only 10 DS0 channels would be allocated to active calls. In my home town of Windsor, where AM 800, CKLW ("The Big 8") was known for routinely jamming the old Step By Step switches in the Goyeau exchange, the former 519-796 mass calling exchange has been been re-allocated as just an ordinary general use NPA-NXX. Modern technology solved the congestion problem, and made radio station contests more fair. No longer were you disadvantaged by the limited number of choke trunks coming from your particular switch - your call was handled on a true first come first serve basis.

Bells protocol specific global throttling is similar to the "old" style of handling mass calling events - not very elegant, and not very fair. What right does Bell have to choose a specific protocol to throttle? Why do they throttle globally across their network, even in segments that are not congested? I don't begrudge Bell for wanting to manage traffic on their network, but they should have done it in a protocol agnostic manner. Better yet, they should keep up with the ever increasing demand for bandwidth.

They (the Telcos and Cablecos) did this to themselves by continually ratcheting up the speeds without having the network to back it up. Now they find themselves in trouble, and use a Band-Aid fix to the problem. Even their advertising keeps promising "consistently faster downloads" - and now they limit a particular class of download technology. Why should a HTTP transfer of a movie be preferred over a BitTorrent transfer? Identical content, but one technology is throttled during peak hours, while another is not?

Since 2001, I have been operating a DSL aggregation network in several Ontario cities with co-located DSLAMs. I have to buy my TDM/SONET bandwidth (OC3) from Bell Canada for inter-CO transport at virtually retail prices. Despite this, we do not throttle traffic, nor do we have congestion issues. If I can afford to do this in a cost effective manner with TDM/SONET bandwidth that I have to BUY from Bell, why is it that Bell cannot do it cost effectively when they OWN the network?

I have no sympathy for them in this case. The argument of scale should only work to their benefit. Their margins are significantly higher than mine, and you can probably guess why.
 

Nice info ..! Bookmarked already