[*] MBIX gets mentioned in CIRA blog ...

[Colin Stanners]

For the first point, you're only speaking about residential - there's
lots of business ISPs who have their private networks and feeds across
canada. But they're a small part of the market and only some are
interested in IXPs.

I posted a comment in the blog about this point - IXPs won't happen
unless ISPs see a reason to get back the $20k they spend to connect to
them. The only way I see that happening is if the CDNs get on board
and install caching boxes there, thereby heavily reducing the ISPs'
bandwidth costs.


On 6/29/12, John Lange <john at johnlange.ca> wrote:
> IMHO, the blog post misses the mark.
>
> IXPs exist to peer traffic between ISPs. The reason Canada has so few IXPs,
> is because it has so very few "real" ISPs (and Canada has few ISPs because
> our telecom policy protects the existing Canadian duopoly by prohibiting
> foreign investment).
>
> In every region in Canada there are (at best), only 2 ways to get to the
> internet, your Cable co, or your Telco. Everyone else is just a reseller of
> one of those 2 providers.
>
> From the post: "Creating more IXPs is about improving security, speed and
> network resilience, while maximizing the amount of traffic that stays
> within Canada for the benefit of all Canadians."
>
> This is incorrect. The main motivation for setting up an IXP is cost,
> network efficiency is only important to the degree that it helps reduce
> cost.
>
> The main ISPs already peer, so setting up more IXPs does nothing to reduce
> their cost. In fact it increases their cost because they need more
> equipment and management etc.
>
> I'm also pretty dubious about some of the other claims. Security? How does
> an IXP improve security? And I'd be very surprised if any Canada-to-Canada
> inter-ISP routing goes through the U.S so the "within Canada" statement
> doesn't seem a likely motivation either.
>
> Don't get me wrong, I not anti IXP, I just don't see a need or a business
> case for it.
>
> John
>