said by DanPure :

I've actually questioned this in my own submission to the CRTC. Are these credentialed, licensed & bonded Professional Engineers, or are they improperly (and perhaps illegally) using the title of "engineer"?

One would think lawyers would know better than to get in the titles mess.

said by cooldude9919 :

Real "network traffic engineers" would have supplied mrtg graphs with some sort of description for each link.

said by TK Junk Mail :

The data they are supplying is going to convince politicians and bureaucrats and not other network engineers.

said by NetAdmin :

are pretty vague and the fact that they say that it depends on how you look at it just confirms that they are overly vague and presented that way on purpose. Concrete information should be presented.

said by KrK :

Um if they presented concrete information, they'd lose their argument that they must throttle 3rd party ISP's.

said by NetAdmin :

Yeah, but everybody loves graphs, especially technically clueless politicos and bureaucrats... As well, graphs usually make information easier to read.

said by TK Junk Mail :

The data they are supplying is going to convince politicians and bureaucrats and not other network engineers.

said by Capt Cap :

As you might recall, British Telecom has been screeching about the same thing because of BBC iPlayer.

Say hello to metering!

said by sbrook :

The matter of pirated material is another one altogether and should be dealt with in its own arena ... not this one where the focus is net neutrality.

said by baineschile1 :

heres the biggest problem i see with ISP throttling; yes they shouldnt do it, yes its wrong for them to advertise unlimited access when it is limited, but...most bittorrent throttle was....PIRATED MATERIAL!!! it stinks for the people who use it for legit reasons, but for us to complain that i had packet loss when downloading Iron Man...its tough to complain.

said by NOCMan :

I could not for the life of me think that a corportion would oversubscribe the services they sell to their customers.

Oh wait.. Sprint Frame Relay.

As can be observed in the table above, the total percentage of all four types of congested network links during a given month in the period in question has varied between 2.6% and 5.2%. While these numbers may seem low to the average lay person, they are significant to network traffic engineers such that it is important to consider the number of congested links in the proper context.

Congestion and Circuit Capacity Planning

Like many organizations, we collect router statistics via SNMP every five minutes. While five-minute averages are a useful measure of how a circuit is doing in relationship to its bandwidth limit, they do not tell a lot about the instantaneous (one second, or smaller, time scale) state of a circuit that governs interactive performance. If a circuit is saturated for several ten-second bursts during a single five-minute period the average utilization might appear to be quite reasonable. An interactive user, however, would likely say that the network was slow while his packets were waiting in a router buffer.

said by Bellus_today :

If you look at this data and combine it with some telecom knowledge, you understand BC's major projects that impact xDSL services:
A remote or subtended shelf is linked with 1 or 2 155Mbps OC-3 links. If one of these can have up to 144 ports, you see the issue. Back when speeds where 1.5Mbps, this wasn't an issue, and not all shelves or RT's will have high subscription rates, and those that do will more likely have been upgraded with the sort-of-fix: Dual gig-e fed Stingers with VDSL2 LIM's (3X48 Ports, I9450 chipset). But the issue here is that much of this extra bandwidth will be reserved for IPTV, so DPI/QoS will still be used. At least they will all have the Gige uplinks in a LAG that can gracefully fail over to a single link.

Next the aggregation: this is simple... the upstream Al-Lu equipment that most dslams connect to is ATM based with a pair of relatively slow OC-12 uplinks, and these can feed 48 subtended stingers. This is also being upgraded for the future that has multicast/IGMP and ethernet rather than ATM to support IPTV. The aggregation layer also goes from OC-12/48 to Gige that probably go to a Nortel 8600 edge router, then onto core routers.

**DPI by ellacoya/arbor is now inserted here before or after the BAS. Has slow Gige interfaces considering how much traffic there is, so 10GE upgrades are coming**

Finally, the BAS/SMS servers after which traffic exits onto the 3rd party networks (Gateway Access Service, Teksavvy, etc.). These are servings thousands of clients and aren't exactly crazy fast... Juniper 1410/1440. These will also likely move to E320's which have better throughput, faster interfaces and backplanes and best of all, more redundancy and better failover.

So yes, DPI sucks, especially since the system doesn't apply it only to the clients from congested elements, but to any PPPoE xDSL user. But it is needed to alleviate congestion to a degree, especially until next generation infrastructure is implemented.

said by daveberstein :


Bell Canada cut their capital spending 25% this quarter. I will now review all their postings to found specifics.